ACL Netherlands BV & Ors v Lynch & Anor [2022] EWHC 1178 (Ch) (17 May 2022)

HARDWARE CLAIMS

The Claimants’ ‘Hardware Case’ in outline

A flagship of the Claimants’ case is their claim that although Autonomy was presented to the market (and in due course to HP) as deriving substantially all its revenue from the sale of high margin proprietary software, in fact from Q3 2009 onwards (and throughout the Relevant Period) a substantial (and certainly material) part of its revenue was derived from undisclosed sales of third-party hardware. In internal Autonomy emails, these hardware sales were often euphemistically referred to as “low margin” sales; but in fact they were almost invariably at a loss.^{^[126]}

The Claimants divided Autonomy’s hardware sales into three categories: (i) ‘pure hardware’ sales, (ii) ‘ Appliance sales’ and (iii) ‘other hardware sales’ . These categories, and the difficulties or obfuscation which they caused, are explained further, together with other issues of terminology in relation to the Claimants’ hardware claims, in paragraphs 648 to 659 below.

Of the three categories of hardware sales, by far the largest was “pure hardware sales” . Put shortly, ‘pure hardware’ sales were sales of hardware where the hardware sold had no Autonomy software in it and the sale did not include any Autonomy software as part of the overall sale. The revenue generated by ‘pure hardware’ sales was substantial: some $200 million over the course of the Relevant Period (constituting some 11% of total reported revenues), including $105 million in 2010 alone.

According to the Claimants, the sales of ‘pure hardware’ lacked any legitimate commercial rationale. Their only real purpose was to assist Autonomy to meet (and if possible, “beat and raise” ) market expectations for overall revenue in the quarter or other longer period in question. The Claimants dismissed as a pretence the Defendants’ justification of the ‘pure hardware’ sales and all the hardware sales as being to assist software sales and promote the development of Autonomy’s core software business.

The Claimants contended further that to achieve the real purpose of what became a routine activity or programme (“the hardware reselling strategy”), it was necessary that the extent of the sales and their contribution to Autonomy’s revenues should never be disclosed or visible to the market. The purpose being to present substantially all Autonomy’s revenues as derived from sales of its own software, it would undermine that purpose if the market was told, or was able to discover, how much of Autonomy’s revenues were in fact derived from loss-making hardware sales.

The necessity to hide the hardware sales in turn necessitated the concoction of a narrative which would persuade what the Claimants described as a “very reluctant” Deloitte and the Audit Committee to accept that revenue from hardware sales should be aggregated with revenue from software sales without differentiation, and that no disclosure of the separate source of part of the revenue was required in Autonomy’s accounts or published information.

In the result, although hardware sales were accounted for in that the revenue and costs of sales were brought into account, Autonomy’s published information mentioned neither the ‘pure hardware’ sales nor any other hardware sale, except the sale of a small quantity of ‘appliances’ at margins represented to be “not dissimilar” to software sales. The hardware reselling strategy was thus never mentioned to analysts and the market. The Claimants contended that the Defendants personally “went out of their way to prevent the market knowing that this was happening”.

In the Claimants’ words in their written closing submissions, Autonomy’s “ hardware reselling strategy” in which both Defendants were “centrally involved” and “actively concealed” :

“…was little more than a scheme whereby Autonomy in effect ‘bought and paid for’ recognisable revenue to inflate reported revenue in order to mislead the market.”

The Claimants further contended that to mitigate and disguise the inevitably adverse effect which low margin/loss making hardware sales would have on Autonomy’s gross margins ^{^[127]} and gross profits, two further refinements of the scheme came to be adopted. Both further demonstrated that hardware sales had nothing to do with driving software growth (which, put broadly, was the Defendants’ justification for it). The two refinements were:

(1) if expected shortfalls in software sales did not eventuate, so that revenue from high-margin software transactions unexpectedly transpired to be sufficient to meet forecasts, revenue from hardware sales would be deferred to a subsequent quarter for later use; and

(2) instead of accounting for all costs as Costs of Goods Sold (“COGS”), as much of such costs as ingenuity would allow and Deloitte could be persuaded to approve would be allocated as Sales and Marketing expenses, thereby diminishing the adverse effect of loss-making sales on Autonomy’s gross profit margin and thus its apparent performance.

The Claimants’ primary case is that the financial position and performance, and thus the long-term value, of Autonomy was thereby fundamentally distorted and overstated.

The gist of the FSMA claim against the Defendants is that they instigated all this, or at least had full knowledge of it, knew that this was wrong and repeatedly misled Autonomy’s auditors, the Audit Committee, the market more generally, and during the pre-acquisition discussions and due diligence process, HP in particular. The Claimants contended that the contemporaneous documentary evidence makes all this plain.

The Claimants’ case on ‘pure hardware’ was summarised in paragraph 54A of their RRAPoC as follows:

“There was no legitimate business reason for Autonomy Inc to enter into any pure hardware transactions. In fact, the pure hardware transactions were not entered into genuinely in furtherance of, or pursuant to, Autonomy Inc’s business, but rather were entered into for the improper purpose of allowing Autonomy falsely to portray itself as a high-margin software company whose revenues were growing rapidly and meeting market expectations, when in reality a substantial portion of these revenues and the apparent rate of revenue growth were the result of undisclosed (and indeed, as pleaded further below, actively concealed) pure hardware sales.”

The Claimants seek very large damages or compensation on the basis, in broad summary, that the fact that a material part of Autonomy’s revenues was derived from loss-making or (at best) low margin hardware sales which were not disclosed meant that Autonomy was a very different company, and very much less valuable, than it was presented to be.

In addition to their FSMA claims, the Claimants also pursued direct claims against the Defendants on the basis that the Defendants breached their duties to Autonomy Inc and/or ASL by causing those companies to conduct hardware transactions at a loss.

Thus, the Claimants’ causes of action in relation to their Hardware Case are:

(1) Claims pursuant to FSMA, on the basis that in consequence of the failure to disclose the hardware sales Autonomy’s published information was untrue or misleading, rendering Autonomy liable to Bidco (which liability Autonomy admitted and now seeks to recover against the Defendants);

(2) Claims in deceit and under the Misrepresentation Act 1967 directly against both Defendants on the basis that they fraudulently made material misrepresentations as to the financial position, performance and prospects of Autonomy and thereby induced HP, and thus Bidco, to proceed with the acquisition of their shares in Autonomy at the offer price.

(3) Claims for breach of duty against both Defendants on the basis that at their instigation Autonomy Inc entered into what they termed “pure hardware” transactions (as to which, see below) “not for any legitimate business purpose, but for the improper purpose of generating the revenues that would enable Autonomy to portray itself, falsely, as a high margin software company”.

Claims within category (3) above seek recovery of transactional losses. However, the losses from the hardware transactions themselves were relatively small (in the region of $20 million) and transactional losses are not the real focus of the Claimants’ complaint. It is important to keep those much smaller claims separate in order to preserve focus on the far larger claims based on misrepresentation and non-disclosure: and I consider the claims for transactional losses in another part of the judgment.

Defendants’ case in summary and matters primarily in dispute in the FSMA Claim

The Defendants did not deny that Autonomy transacted substantial hardware sales in volume and value terms between Q2 2009 and Q2 2011. The Defendants’ case was that reselling hardware at a discount to its customers was a carefully weighed strategy, the principal purpose of which was to protect, support and facilitate software sales and Autonomy’s software business.

As presented by the Defendants, the strategy had two primary prongs: one pointing towards Autonomy’s hardware suppliers and the other pointing towards its software customers:

(1) As regards hardware suppliers, its objects were to enhance Autonomy’s relationships with those suppliers and increase the prospect of them (a) optimising their hardware products for use with IDOL software, (b) co-operating with Autonomy to develop a joint “appliance” if the market adopted what appeared to be a move towards appliances as the means of delivery of software and (c) generally, encouraging their hardware customers to buy Autonomy/IDOL software. Autonomy also hoped and expected thereby to increase Autonomy’s market presence and the market penetration of Autonomy’s software offerings.

(2) As regards Autonomy’s customers, its objects were (a) to enhance its ability to offer one-stop shopping and bulk purchasing and thereby (b) to consolidate its position as a favoured IT supplier and achieve or retain “key supplier status” at a time when some of its customers were limiting the number of their IT customers in order to make procurement cheaper and more efficient. Further, many of its customers had procurement departments, whose performance was measured on the discounts they could obtain. Although the discounts offered by Autonomy caused most of its hardware reselling to be at a loss, the Defendants contended that the losses were small, especially in comparison to the very high margins on increased software sales encouraged by the strategy.

Their principal answers to the question posed by Mr Rabinowitz, put summarily, were that:

(1) The sales of hardware were never an end in themselves, and Autonomy never undertook a separate business as a hardware seller or reseller: its hardware reselling strategy was part of its single segment software business and, in effect, a loss leader and/or a form of marketing to protect and promote that business;

(2) The hardware sales or re-sales were disclosed to and reviewed in great detail by Deloitte, who approved their accounting treatment;

(3) Every deferral of revenue was justified, and the allocation of costs of hardware sales in part to Sales and Marketing was discussed and agreed with Deloitte and approved by both Deloitte and the Audit Committee;

(4) It was for Deloitte to advise, and they did advise, what disclosure was required by accounting standards, and Autonomy never acted contrary to that advice; and that

(5) It was a management judgment whether further disclosure should be made voluntarily, but on rational and substantial grounds, management determined not to provide further disclosure, and their decision was supported by both Deloitte and the Audit Committee who signed off on Autonomy’s published information throughout the Relevant Period.

The Defendants relied principally on the evidence of the Claimants’ own witnesses, and in particular, that of Mr Sullivan and Mr Egan, which (they submitted) confirmed that there was a valid commercial rationale for the hardware sales (including the “pure hardware” sales) as a means of driving software sales. The Defendants submitted that (since it is their own) the Claimants are bound by that evidence, which was also given support by another witness called by the Claimants, Mr Welham.

The Defendants acknowledged that the sales of hardware did provide “flexibility” in terms of supplementing revenue and could be and were used to enable revenue to be supplemented where quarterly revenue slipped. However, that was a collateral advantage rather than the purpose, let alone the primary purpose, of such sales.

In asserting that “None of these claims bears scrutiny” it was an important element of the Defendants’ case that Deloitte reviewed all the published information now contended by the Claimants to have been fraudulently presented, and (as Mr Welham accepted) would not have allowed any statement or presentation to be made that they considered misleading.

According to the Defendants, Deloitte were aware of Autonomy’s hardware sales, right down to a granular deal by deal basis, their amounts, their loss-making nature (to the extent that they were so) and the identities of the counterparties. They knew that some of the hardware was sold without adding any software to it. Deloitte considered and discussed with Autonomy’s finance department and the Audit Committee not only the inclusion of the revenues as revenue, but also the disclosure, categorisation and explanation of Autonomy’s various types of revenue.

Dr Lynch also stressed that he was not an accountant; he was entitled to, and did, think that in preparing and scrutinising the accounts the finance department (made up of experienced accountants), Deloitte and the Audit Committee had properly done their job. Thus, if more disclosure was required, he expected them to say so; and if any of those people had thought the published information was misleading, he expected them to say so, explain why, and offer an alternative which would ensure its fairness and accuracy.

In the round, Dr Lynch depicted the aggregate gross losses from the hardware reselling strategy as a small price to pay for what in effect was, as he saw it, a successful marketing strategy at a time of threatening market changes which was of considerable benefit to Autonomy.

The Principal Issues

The Claimants identified and summarised the principal issues to be determined in relation to their case impugning the hardware transactions as follows:

(1) What was the Defendants’ purpose in causing Autonomy to resell “pure hardware” ?

(2) Was Autonomy’s published information untrue or misleading by reason of the ‘pure hardware’ sales, and did the Defendants appreciate this?

(3) Did Autonomy’s treatment of the costs of the ‘pure hardware’ sales render Autonomy’s published information untrue or misleading, and did the Defendants appreciate this?

(4) Should Autonomy, at the least, have made clear in its published information what its accounting policy was with respect to hardware?

(5) Did Autonomy wrongly recognise revenue in Q2 2009 on a specific ($6 million) hardware transaction with Morgan Stanley?

That summary of the issues, which as to (2) to (5) follows the structure of the first report of the Claimants’ expert witness, Mr Peter Holgate, provides a useful overall structure for the analysis and determination of the hardware case. However, these headline issues need to be broken down further to understand the various strands of the Claimants’ hardware case as it was ultimately put forward, and the Defendants’ answer to it (especially their contention that they relied on Deloitte’s approval).

First, it is necessary to identify three different contexts in which the Claimants’ case as to the purpose of the hardware reselling strategy (Issue (1) in paragraph 638 above), which was the central area of disputed fact, is relevant:

(1) Its principal context is the Claimants’ contention that the overall purpose asserted by the Defendants was no more than a pretext falsely to justify describing and accounting for revenue from the hardware sales as if it were part of, or at least not to be distinguished from, revenue from software sales. That is a self-sufficient part of the Claimants’ hardware case, in that they contended that if they established it, Autonomy’s published information was plainly false and misleading, and can be determined to have been so without reference to the expert evidence. Mr Shelley, co-head of one of Autonomy’s Corporate Brokers (UBS), had to concede in cross-examination that if the market had found out that Autonomy was selling hundreds of millions of dollars of hardware to inflate its apparent software revenues then that would have been fraud and the market would not have reacted well.

(2) A second context in which the identification of the purpose of the programme is relevant is the Defendants’ defence that Deloitte approved all the published information relied on by the Claimants. The Claimants contended in that context that the defence is invalidated if it is established that the Defendants misrepresented to Deloitte the true purpose of the hardware reselling strategy.

(3) The third context in which a determination of the purpose of the hardware reselling strategy is relevant is in assessing the credibility of the Defendants, and in particular of their justification for their determination that the programme should not be disclosed to the market.

Secondly, it is necessary to distinguish between the two different limbs of the Claimants’ case that Autonomy’s published information was false and misleading:

(1) One limb relates to positive statements or representations in Autonomy’s published information: this aspect (“the positive misrepresentations case”) focuses especially on the way Autonomy chose to present its business and sources of revenue in the ‘front-end’ or ‘narrative’ part of its accounts.

(2) The other limb relates to alleged omissions from Autonomy’s published information: this aspect (“the expert accountancy case”) focuses especially on the Claimants’ expert evidence that under IFRS rules hardware transactions should have been, but were not, separately disclosed.

The Claimants emphasised that their positive misrepresentations case is not dependent on expert evidence, since no expert evidence is needed for the proposition that a company’s published information must not contain false or misleading statements. The issue is one of interpretation (and more particularly what meaning would have been given in the market to what was positively stated in Autonomy’s published information about the nature of Autonomy’s business and its sources of revenue). By contrast, the Claimants’ expert accountancy case is reliant on the view of their expert, Mr Holgate , that the IFRS rules and statements required disclosure, and that this was too plain to be a matter of fine judgement.

Thirdly, it is important to understand the relationship between the Claimants’ “purpose case” (see (1) in paragraph 638 above) and the other aspects of their hardware case (see (2) to (5) in paragraph 638 above). I have already explained that if the Claimants succeed in demonstrating that the avowed purpose was a pretence that also concludes Issue (2). The question now addressed is as to the effect if the Claimants do not succeed in that regard. As to that, in oral closing argument, Mr Rabinowitz emphasised that neither the Claimants’ positive misrepresentations case nor their expert accountancy case is dependent on proof of pretence or even simply some improper purpose. As to the positive misrepresentations case, which became by the time of closing arguments, the Claimants’ favoured route to success, Mr Rabinowitz submitted that:

“it really doesn’t matter why Autonomy was selling this hardware, and whether, as we say, it was simply a revenue-pumping exercise or whether, as the Defendants say, it was part of some marketing strategy. Because either way this could not have justified misrepresenting the totality of its sources of revenue, which is what it did.”

Fourthly, however, and the above submission seems to me to offer an example, the Claimants tended to glide over the need to establish in relation to Issues (2) and (3) (see paragraph 638 above) not only that Autonomy’s published information was untrue or misleading, but also that the Defendants appreciated this. As I have already signalled, the Defendants relied on Deloitte’s approval of their narrative description and accounting treatment of the hardware reselling strategy (and, in particular, its revenue) as both justification and a litmus test of the acceptability of describing and treating hardware sales all as part of the IDOL Product business. If, to borrow the way Mr Rabinowitz put it in his oral closing submissions, “the real question that arises”, is whether the Defendants had any legitimate basis for the way they presented their business and sources of revenue (which did not reveal the hardware reselling strategy), the Defendants’ answer was both the purpose they asserted and the acceptance of that presentation by Deloitte. That is why identification of the ‘purpose’ of the programme is relevant also to the Claimants’ positive misrepresentations case.

I would add three further glosses to the Claimants’ adumbration of the issues as set out in paragraph 638 above. The first is that the Defendants raised two preliminary issues as to the permissible extent of the Claimants’ case in respect of the purpose of the hardware reselling transactions. In particular, the Defendants contested whether it was open to the Claimants (a) given the way they had pleaded and pursued the case prior to their closing submissions, to rely on an alternative claim that even if revenue pumping was not the only, it was the primary or preponderant, purpose of the hardware sales; and (b) given the evidence of their own witnesses, to contend that marketing and assisting software sales was not one of the substantive purposes of the hardware sales. I address those two issues as preliminary points in the course of my detailed discussion of the Claimant’s ‘purpose’ case in paragraph 660 et seq below.

The second gloss is that the Defendants also raised a further issue, which is whether HP knew about Autonomy’s hardware sales pre-acquisition and continued them post-acquisition. I address that issue in paragraph 1802 et seq below.

The third gloss is that there were also subsidiary, but important, disputes as to terminology, to which, to assist better understanding of the principal issues, I turn immediately.

Issues of terminology

Before turning to the substantive issues in more detail, it is convenient to identify and explain certain phrases deployed by the Claimants to define and confine their hardware reselling claims.

Although ultimately the Claimants’ FSMA claims extended to all hardware sold on the basis that (because the margin on hardware sales was so much lower) they all should have been disclosed, initially their challenge to the legitimacy of the transactions was confined to sales of “pure hardware” (though, on their definition over 99% of the hardware revenue was attributable to “pure hardware” and “other hardware” was a small residual category); and they reserved particular criticism of “pure hardware sales” on the basis that there was no good reason, as they saw it, why a software business should want to sell “pure hardware”.

The expression “pure hardware sales” is not a term of art: it is a term of the Claimants’ own devising. It is distinguished in their pleading from “appliance sales” and “other hardware sales” . Their case as pleaded is that “pure hardware sales” were any “ sales of third-party computer hardware (with or without third party software) without modification by Autonomy and unaccompanied by any Autonomy software” . That last phrase engendered a further point of dispute, as to whether the sale of hardware with a view to later provision or use of Autonomy software was or was not a “pure hardware sale”.

The Claimants defined “other hardware sales” , which they described as “insignificant compared with pure hardware sales” as “sales of hardware (that were not appliance sales) that were in connection with sales of Autonomy software” (for example, because a sale of software appears on the same order). Thus “other hardware sales” were, in the Claimants lexicon, a residual category, being neither “pure hardware sales” nor “Appliance” sales. Obviously, it is not a term of art either and it is a rather loose categorisation. The Defendants contended that all of Autonomy’s hardware sales were in some sense “in connection with” sales of software: and for their own part, they characterised such sales as sales of an “Appliance” or an “Appliance-type sale” .

The content to be given to the expression “appliance” is also of considerable importance and was a contested issue. Again, the phrase “appliance sales” is not a term of art with a fixed meaning. The Claimants used “ appliance sales” to mean “[sales of] hardware on which Autonomy software had been pre-installed and which was offered where the customer had an urgent need to deploy IDOL”. But that was not how the Defendants understood the term, nor how (they say) it was generally used at Autonomy. Mr Miles described it as a self-coined term of some elasticity. Dr Lynch emphasised, for example, that:

“…the term “appliance” was used, in different contexts, to refer to a variety of things, including hardware pre-packaged with software, hardware generally and hardware to be used with software. There was no singular definition of the term. If, for logistical or budgetary reasons, a customer bought hardware, such as servers, separately from the archiving software that went with it, customers, Autonomy staff and the auditors would still call these “appliances”. I should add that the term “appliance” was an industry term and, like many technical terms, its meaning has evolved over time.”

Dr Lynch elaborated on this in the course of his cross-examination:

“…the fundamental essence of an appliance is that you have a standard block, usually of hardware but not always, with a standard block of software and the important point is that you put the two together and they do one thing. So a normal computer like your laptop, you can put lots of different programs on and it will do lots of things and you have to install those programs. In the enterprise world, when you sell a piece of software, someone then goes and gets some piece of hardware from many different possibilities, they then configure that hardware, they get the software, they then have to configure the software to match the hardware and that whole process, if you’re a company, can take three or four months. If you’re a government it seems to take years, but it’s a big process.

What an appliance is is something that is designed to do one thing, so you don’t need to worry about all the other possibilities and to do that there’s a standard block of hardware and a standard block of software and you put the two together and it does the standard task and it’s much quicker and more reliable to get working. So “turnkey” is the phrase that’s used in the industry for that. And that is the defining characteristic of an appliance. So, for example - I just used the example of a games appliance that people have in their homes, that’s called an appliance because it’s a standard thing that’s designed to run games, it doesn’t do anything else, you can put games in it and it will run games and that is an appliance.

The term has been around for a while, it has changed its meaning over time, but the fundamental point is it’s not a generic computing device, it’s a standard block of hardware, standard block of software, standard function.

…

I do not believe that it is deterministic of whether something is an appliance as to whether the hardware and software are ordered together…

Q. For it to be an appliance, would it have to be standard Autonomy software?

A. Well, to be an Autonomy appliance, it would have to be one, yes.

Q. All right, so when Autonomy talks of Autonomy selling appliances, what you are saying you sold or what you are saying you were talking about was the sale of hardware in order for Autonomy software to be loaded on it?

A. Correct.”

A further important point of disputed terminology is the meaning (both (a) as intended and (b) as received)^{^[128]} of the phrase “pure software model/company” as used by Autonomy in its published information (and especially in its 2009 Annual Report and 2010 Annual Report)^{^[129]} . In the 2009 Annual Report, Autonomy’s financial model was stated to be “one of the very rare examples of a pure software model”. In the 2010 Annual Report it was described as “a rare ‘pure software’ model” and the Business Overview section elaborated as follows:

The Claimants also claimed that the same phrase had been used by the Defendants in pre-acquisition discussions, and that this was intended to deceive (see the Chapter in this judgment on “Deceit and Misrepresentation” ) .

The Defendants contended that they intended the description to convey by the phrase that a distinguishing feature of Autonomy was that, unlike most software companies, it did not undertake or derive substantial revenues from any servicing work. They did not intend thereby to convey that Autonomy did not undertake or derive revenue from hardware sales, still less sales of hardware with no software content and/or entirely separately from any software sale. However, it is part of the Claimants’ case in relation to hardware that by representing itself to be a “ pure software” company, Autonomy not only sought to distinguish itself from a company that derived a significant proportion of revenue from services. Its description also, when combined with express references to appliance sales and the provision of categories of revenue which made up the total revenue of Autonomy in the Relevant Period, would either individually or cumulatively, lead a reasonable reader to conclude that:

(1) Autonomy was not engaging in any (or any material) sales of hardware apart from appliance sales, or in the very least was not engaging in any (or any material) pure hardware sales, and/or

(2) any revenue from hardware sales and/or pure hardware sales (apart from appliances) was lower than the revenue from appliance sales, and/or

(3) any revenue from hardware sales (apart from appliance sales) was lower than revenue from Services.

I have concluded that, at least in the way the statement was deployed after the commencement of volume hardware sales, the purpose of the statement was to convey a special selling point, the success and self-sufficiency of Autonomy’s software business without the need for other revenue streams. A reader of the statement in its context would in such circumstances not have expected any substantial revenues from hardware sales to be as much as or more than its expressly disclosed activities, such as servicing. In the end, therefore, I have concluded in favour of the Claimants on the point.

There was also a dispute as to the meaning to be ascribed to the phrase “organic growth” . The Claimants pleaded and submitted that the inclusion within the metric of hardware sales improperly inflated it and gave a different and false picture. The Defendants disagreed: as Dr Lynch said, to exclude hardware from the calculations would be to produce a different metric, “organic growth ex hardware”.

These disputes are not merely semantic. They go to the root of the issue as to the interpretation to be given to certain statements in Autonomy’s published information, and whether HP was misled, whether by Autonomy’s published information or by what was said prior to the conclusion of the transaction. I shall return to each disputed phrase in context when dealing with the issues identified by the Claimants in paragraph 638 above. I turn now to the first of these.

Issue (1): what was the Defendants purpose in causing Autonomy to resell “pure hardware”?

The Defendants’ case as to the purpose of Autonomy’s hardware reselling strategy was very largely based on (a) the evidence of Dr Lynch as to its initial rationale, (b) the evidence of the key employees charged with its implementation (and in particular, Mr Sullivan, Mr Egan and Mr Scott), and (c) Deloitte’s acceptance of the rationale and the conclusion that hardware sales were incidental to the software business so as to require no separate disclosure.

The Claimants’ case that the reselling of “pure hardware” was simply “a device deployed by the Defendants to assist in meeting Autonomy’s revenue targets, inflating both overall revenue growth and ‘organic growth’” was based largely on the documentary evidence, and especially on what that evidence revealed as to the chronology and the way that hardware sales were in fact conducted. Their exegesis of the documents and chronology was very detailed, as indeed was the Defendants’ response.

Three notable features

Before turning to that considerable detail, three general features of or issues relating to the Claimants’ contentions on the hardware purpose case may be noted:

(1) First, as pleaded, the Claimants sought to impugn all ‘pure hardware’ sales on the same basis of being in every case for the improper purpose of generating reportable revenue and presenting that revenue as if it were referable to software sales: the Claimants asserted a “systematic policy” in operation between Q2 2009 and Q2 2011, and did not make any distinction according to the context of the particular sale or the business of the purchaser. In his oral reply, Mr Rabinowitz accepted that there really was not the material to enable an informed assessment on a transaction-by-transaction basis. Thus, the Hardware Case concerned the propriety of a strategy of which the hallmarks were (almost invariably) loss-making sales of hardware, the revenue from which was recognised and then, without differentiation or disclosure, included as part of the overall revenues of Autonomy’s software business.

(2) Secondly, the case (as pleaded) asserts the improper purpose identified as the only purpose of the strategy. As Mr Miles put it in his oral closing submissions, “It is not a case that involves, as it were, the weighing of purposes”. However, over the course of the Trial the Claimants increasingly sought to treat it as being sufficient for their case if revenue generation to enhance the appearance of software sales growth was shown to be a substantial purpose, even if not the exclusive purpose. In their written closing submissions, the Claimants relied on “the primary - if not the only - purpose” which seemed to allow for the possibility of some subsidiary commercial rationale: but that was not pursued or pleaded by amendment, whether in the alternative or at all. Mr Miles submitted in his closing submissions that, having pleaded, and in their opening presented, their hardware case on the basis that the binary question was whether (their case) it was a device to pump revenue without revealing its source or (the Defendants’ case) a rational strategy based on a business judgment that it would protect and promote Autonomy’s software business, it is not open to the Claimants in closing to run a case based on primary or preponderant purpose. That gave rise to a pleading issue which it is convenient and necessary to deal with as one of two preliminary matters (see paragraphs 663 to 675 below).

(3) Thirdly, and obviously related to (2) above, the witnesses called by the Claimants on the issue as to the purpose of the policy or strategy (and, in particular, Mr Egan and Mr Sullivan) had all accepted in earlier evidence they had given in the US criminal proceedings, and they again accepted in these proceedings, that part of the purpose was to drive software sales and realise other benefits as explained below. It was submitted for the Defendants that the Claimants were bound by that evidence, and that this was fatal to their “binary” case. The Claimants did not accept this, their main submission being that there had “been some cherry-picking of Mr Sullivan’s evidence at Mr Hussain’s criminal trial” , and more generally, the evidence as a whole was more nuanced than the Defendants suggested, and in any event could not be read out of the context of the documentary evidence. That dispute gives rise to the second issue which it is also convenient and necessary to address as a preliminary matter.

A pleading issue: Is it open to the Claimants to impugn the sales on the basis that ‘revenue pumping’ was the predominant , even if not the only purpose of them?

The Claimants’ Particulars of Claim at paragraph 54A plead the Claimants’ factual case for impugning the hardware sales as follows:

“54A. There was no legitimate business reason for Autonomy Inc to enter into pure hardware transactions. In fact, the pure hardware transactions were not entered into genuinely in furtherance of, or pursuant to, Autonomy Inc’s business, but rather were entered into for the improper purpose of allowing Autonomy falsely to portray itself as a high-margin software company whose revenues were growing rapidly and meeting market expectations, when in reality a substantial portion of these revenues and the apparent rate of revenue growth were the result of undisclosed (and, indeed, as pleaded further below, actively concealed) pure hardware sales.”

That pleading is, as to establish the FSMA claim it had to be, an allegation of dishonesty: the purpose pleaded is to achieve an objective ( “to portray itself as a high margin software company” which undertook no substantial hardware sales) by deception. The Claimants’ case as pleaded made no express allowance for any other purpose than to generate revenue from one source (hardware) and present it as derived from, and an expense of, another source (software sales). The allegation is that what was presented to be a commercial purpose was in fact a dishonest pretext contrived in order to present a false description of the true nature of its business.

Paragraph 57A then pleads a claim for the direct loss sustained on the pure hardware claims on the basis that:

“The Defendants each knew (since, as particularised at paragraphs 133 and 135 below, it was they who instigated the practice of Autonomy Inc entering into pure hardware transactions) that the practice of entering into pure hardware transactions was not genuinely in furtherance of Autonomy Inc’s business but was for the improper purpose pleaded at paragraph 54A above. The losses suffered by Autonomy Inc on such transactions are therefore losses occasioned by the Defendants’ breaches of duty”.

The direct claim also is based on dishonesty, even though in a claim for breach of duty it would ordinarily be sufficient to plead and establish that an improper purpose was in fact the predominant purpose.

Until closing submissions, the Claimants had never put their case on the basis that there might be legitimate business reasons for the hardware reselling, but they were secondary to some other improper purpose. Mr Miles, with a forensic side swipe that the change of tack was “doubtless because of the difficulties they have with their own evidence” , submitted that that would be a new, distinct and unpleaded case, and as such it was not open to the Claimants. His position was that having pleaded and opened the case on the basis that the entire purpose of the strategy was to generate revenue, even if at a loss, with a view to presenting that revenue as derived from Autonomy’s software business without disclosure of its true extent and source, it was too late now for the Claimants to base their case on the weighting of preponderant purpose in respect of individual transactions.

Mr Miles stressed that this was not “an arid pleading point” : had the Defendants understood there to be a case based on the primacy amongst mixed motives the case would have taken a different course. He submitted that, for example:

(1) He would have wished to cross-examine Mr Welham and explore what difference it would have made to the accounting treatment;

(2) The Defendants “might very well…have taken the view…that…steps should be taken to ensure that [Mr Sullivan] gave evidence by video” as a condition of his witness statement being admitted into evidence as hearsay (Mr Miles made clear he could put it no higher);

(3) They would certainly have wished to cross-examine Mr Holgate, and to explore with him how accountants would deal with the idea of mixed motives, and what if any guidance there is in accounting literature;

(4) They would have wished to assess the mixed purposes of each transaction, to weigh its preponderant purpose: and that was not possible, as Mr Rabinowitz had accepted; and

(5) Furthermore, even if such an enquiry had been possible, they would also have wished to question the experts as to the feasibility of such an enquiry in the course of an audit.

Mr Rabinowitz sought to answer all this with the submission that the possibility of mixed purposes, necessitating a determination of the primary purpose, had “always been in play”. He submitted that it was not a new case at all. He especially relied on (a) passages in Mr Hussain’s Defence relying on the primary, if not the sole, purpose being to drive high-margin software business, (b) passages in witness statements by the Claimants’ witnesses referring to “primary” or “principal” reasons, (c) passages in the cross-examination of the Claimants’ witnesses, and in particular Mr Egan, exploring whether driving software sales was or was not a principal or predominant reason, and (d) references in the experts’ respective reports identifying the key area of judgement as being (to quote Mr MacGregor):

“whether management considered that the sale of hardware was the separate sale of a different product to the sale of Autonomy’s core IDOL Product, or that it was incidental to sales of the core product, with its principal purpose being to facilitate further software sales.”

I must confess that during the hearing I tended to follow and accept the argument that the question was whether the Claimants’ pleading accommodated a weighing of purposes; and if that was the right question, I would agree with Mr Miles that it does not.

On reflection, however, I think that is a mischaracterisation. The true gist of the Claimants’ pleading is that the transaction was dishonest. The real question is not whether the hardware sales could be said to have been intended to have had, and in fact had, some potentially beneficial effects such as to suggest that that was their purpose, or part of it; it is whether a dishonest purpose underlay the transaction. In other words, the real answer lies in the plea of dishonesty.

In my judgment, if a substantial part of the reason for a transaction is found to have been dishonest, the fact (even if established) that it may have been perceived that it would also have beneficial effects cannot save it. The dishonest purpose need not be exclusive or even predominant (though it is unlikely not to be so).

In this case, dishonesty, not predominance of improper purpose, has always been the Claimants’ case. According to their pleading, what caused Autonomy to enter into the hardware sales was not only improper (which might include a purpose beyond the powers of the directors) but dishonest: the false portrayal of Autonomy as a high-margin software company deriving all its revenues from software sales when in reality a substantial proportion of its revenues and its apparent revenue growth were the result of undisclosed, and indeed actively concealed, ‘pure hardware’ sales. If that is proved, that is sufficient. No weighing of purpose or other effect is required.

It follows that the pleading point raised is based on a false premise, perhaps encouraged by some of the language used, the focus on whether avowed purposes were real, and the dispute that developed as to whether revenue recognition or relationship development to facilitate software sales most weighed with the witnesses and in particular, Mr Sullivan and Mr Egan. The Claimants’ hardware case has never been about competing purposes: it has always been that the objectives and purposes deployed by the Defendants to justify the hardware reselling strategy were pretexts: the real purpose of the programme was to generate revenue which could dishonestly be presented in the accounts and Autonomy’s published information as part of the revenue from Autonomy’s main software business. In my judgment, that case, which requires proof of pretext and dishonest concealment but not the weighing of purposes, is sufficiently pleaded.

Two other matters connected to this need also to be borne in mind:

(1) First, an important part of the defence (of both Defendants) was their reliance on Deloitte’s approval of Autonomy’s accounts and published information even though Deloitte were well aware that the hardware sales generated revenue which was included in those accounts without differentiation from Autonomy’s revenues from its software business. However, any such reliance would be negated if what Deloitte were told was the purpose of the hardware reselling (protecting and promoting software sales) was an anticipated incidental benefit of what was in truth a device to preserve the appearance of revenue growth by covering from hardware revenue a shortfall in software revenue. If the Claimants establish that those were incidental benefits to disguise the true purpose, and that the true purpose was never disclosed, then Deloitte will have been shown to have been misled. Put another way, although the Claimants’ case did not allow for a weighing of competing commercial purposes, it did extend to demonstrating that the purpose asserted, though fulfilled, was (or became at some point in the implementation of the programme) no more than an incidental or ancillary benefit and not the real purpose.

(2) Secondly, in the context of the expert accountancy case, Mr Holgate’s evidence was that the purpose of the hardware sales did not ultimately matter, it being his opinion that any proper application of the accountancy rules required separate disclosure of the hardware reselling strategy in any event. By contrast, on Mr MacGregor’s evidence as I understood it, only showing that revenue pumping was the sole real purpose would require such disclosure. On that basis, as to the expert evidence, a case based on primary purpose would not have been necessary for the one expert or sufficient for the other. (I shall address further the expert evidence in this regard in the context of the Claimants’ alternative case.)

An issue as to the evidence given by Mr Egan and Mr Sullivan in the US criminal trial

As to the proof of the Claimants’ case that the real purpose of the hardware reselling strategy (what they described as “this…bizarre and expensive programme” ) was to generate revenue from covert hardware sales to make good shortfalls in sales of software, and thus continue the depiction of Autonomy as a high margin software company which was (usually) meeting or beating market expectation, the Claimants placed primary reliance on the contemporary documentation.

The Claimants observed that the Defendants tended to shy away from this documentary evidence, and especially the correspondence between Mr Sullivan, Mr Hussain and Dr Lynch, almost all of which, the Claimants pointed out, appeared to be directed towards the issue of revenue generation and what hardware deals would be needed to plug gaps in revenue.

There is force in this, as will be clear later; but similarly, there is something in the Defendants’ opposing contention that the Claimants have themselves sought to shy away from the evidence of their own witnesses on the issue as to the purpose of the hardware strategy, which is barely mentioned in the witness statements of Mr Egan and Mr Sullivan, but which is more substantially dealt with in their evidence in the US criminal proceedings. In those US criminal proceedings, evidence from many of the witnesses for the Claimants had already been examined on substantially the same issues.

Both sides referred to and prayed in aid parts of the transcripts of that witness evidence (which in the case of Mr Sullivan, for example, was entered into evidence under a hearsay notice since he did not appear at the trial here). Before discussing its substance, the question I address now relates to the status of that evidence.

At a pre-trial hearing, I directed the parties to side-line what parts of the transcripts of evidence given in the US criminal trial by witnesses who would not be called in this trial they respectively intended to refer me to or rely on. Evidence side-lined by one party was taken at trial to be the evidence of that party. During the trial, evidence was taken to be that of the party that had indicated that they wanted to refer to it, as if it was that party’s evidence. However, in every case, the witness concerned was called in the US criminal proceedings in support of the case against Mr Hussain; and in assessing their evidence I have always taken that into account, and that each was in reality the Claimants’ witness.

However, and I did not understand any party to contend to the contrary, the side-lining did not preclude any party or the Court from referring to parts of the transcripts of evidence as necessary to obtain a proper understanding of the evidence as a whole. The whole of the relevant transcripts were before the court in agreed bundles (and see CPR PD32/27.2) and were admissible as to their contents, and although the Defendants in correspondence expressly stated that they did not thereby accept their treatment as evidence of the facts stated in them, it was nevertheless open to them, and in any event to the Court, to have regard to the whole in testing the true meaning, reliability and veracity of the side-lined parts.

In addition, as it seems to me, and as was submitted by the Defendants, the party which has sought to rely on the witness as a witness of truth, cannot invite the Court to believe the parts identified by that party as helpful to its case and yet disbelieve other parts which go the other way. The whole is the evidence of that party’s witness, for good and ill. As Brooke LJ said in McPhilemy v Times Newspapers Ltd (No 2) [2000] 1 W.L.R. 1732 , at 1740:

“I know of no principle of the law of evidence by which a party may put in evidence a written statement of a witness knowing that his evidence conflicts to a substantial degree with the case he is seeking to place before the jury, on the basis that he will say straight away in the witness’s absence that the jury should disbelieve as untrue a substantial part of that evidence.” ^{^[130]}

In any event, and as the Defendants also pointed out, the passages in the evidence of Mr Sullivan in the US criminal proceedings on which the Claimants have chosen to rely include evidence as to the existence and legitimacy of the hardware sales strategy: and there can be no doubt that that evidence must be taken to be part of the Claimants’ own case. However, that does not, of course, preclude them from reliance on other evidence, such as that of Mr Camden on whose evidence in the US criminal proceedings, about three transactions with Zones Inc, the Claimants focused in considerable detail (and see paragraphs 1159 to 1182 below).

The position is less unusual in the case of Mr Egan. He did attend, albeit via video link. I did not take it to be disputed that the whole of his evidence, including that in the US criminal proceedings, needs to be weighed.

The relationship between purpose and concealment

The Defendants’ depiction of the purpose of the hardware sales as being to protect and promote Autonomy’s software business is the basis on which they sought to justify both the loss-making sales themselves (Autonomy not having previously engaged in material sales of hardware) and their treatment as part and parcel of Autonomy’s software business in Autonomy’s accounts (both in the narrative ‘front end’ part and in the ‘back end’ accounts proper).

For the purposes of analysis, two related questions need to be addressed. The first is what in truth was the real purpose of the hardware sales. The second is whether having regard to that purpose it was rational and permissible to treat and account for the hardware sales and the revenues from them as an integral part of Autonomy’s software business and all its revenues, requiring no separate identification or explanation.

That, on the Claimants’ case, is the link to concealment. The Claimants’ case is that the true purpose of the hardware reselling strategy is revealed by the Defendants’ determination that the sales and the revenues they generated should never separately be identified: even if the ‘purpose’ case and the ‘disclosure’ case have for the purpose of analysis, to be separately examined, they are inseparable. Their true driving purpose was never, and could never, be disclosed: because that true purpose was to generate revenue in order covertly to make good shortfalls in software business revenues by including hardware revenue without differentiation or disclosure in Autonomy’s accounts and reports as if it had been generated by Autonomy’s ‘pure software’ business.

Thus, in his oral closing, Mr Rabinowitz melded, and in a sense reversed, the two related questions in submitting that ultimately what has to be determined is whether the Defendants have been able to identify any legitimate basis for having withheld from the market information which would have had a “devastating impact on the picture that was painted of Autonomy in the published information”.

Structure of this chapter

The Claimants’ hardware case, as one of their lead cases (with their case on VARs), engendered a great deal of paper and gave rise to a plethora of issues and a considerable volume of evidence. That has resulted in this part of the judgment being of considerable length and factual density. There was a marked contrast in the way the two sides approached the issues.

The Defendants based their main defence to the claim on (a) Dr Lynch’s explanation of the rationality of the strategy (b) an analysis of the witness evidence, and especially the evidence in cross-examination of the Claimants’ two principal witnesses on the Hardware Claims, Mr Sullivan and Mr Egan, and (c) Deloitte’s approval of the accounting treatment which the Claimants now criticised.

They also sought to rely on evidence which they suggested revealed that HP knew about the hardware reselling before the acquisition; and certainly did so after it; but until, in the context of serious investor criticism of the acquisition, blaming the Defendants became a commercial imperative, neither HP nor its accountancy advisers (Ernst & Young for these purposes) were concerned.

The Claimants did not accept the Defendants’ depiction of the evidence of Mr Sullivan and Mr Egan; but their main approach was to focus on the documentary evidence of the way the hardware reselling strategy was actually implemented, not only initially in Q3 2009, but over the course of the Relevant Period. In that context, they relied, not only on the documents that revealed the structure and nature of the transactions and contemporaneous exchanges within Autonomy as to their objectives, but also the lack of documentation to support the Defendants’ suggestion that the programme was in the nature of a marketing exercise and was succeeding in its objective of driving software sales.

Contending that the programme was simply a device for “in effect, paying for, or buying, revenue that could be recognised” without disclosing its real source and nature, the Claimants rejected as dishonest the reasons given for the hardware sales and for the inclusion of the revenues they generated as if they were part of, or at least not to be distinguished from, revenue from software sales.

In light of these differences in approach I have felt it necessary to deal in turn with each side’s case in the round. That has further increased the length of this part of this j udgment. But it was a centre-piece of the claims and invites and (to my mind) requires, detailed treatment.

In broad outline:

(1) I discuss first the dispute between and competing cases of the parties as to the origins of the hardware reselling strategy.

(2) I address next (somewhat unusually) the Defendants’ case that its purpose was to protect and promote Autonomy’s software business and that it was both properly accounted for and sufficiently disclosed. The Defendants’ case focused on the approval of Deloitte and the evidence of certain key witnesses, especially Mr Sullivan and Mr Egan (both of whom were, of course, the Claimants’ witnesses), as well as Dr Lynch himself.

(3) I then turn to the Claimants’ case that the purpose asserted by the Defendants was or swiftly became a pretext for what in reality was a relentless drive to generate sales from hardware to cover shortfalls in sales in software licences, which had to be and was dishonestly disguised to maintain the appearance of continuing to maintain or increase revenues in line with market forecasts. The Claimants’ case focused on the documentary evidence, a detailed assessment of the sequence of the hardware sales, of the correlation with the need to make good shortfalls in revenue from Autonomy’s software business, and of what Deloitte were actually told from time to time.

(4) Lastly, I summarise my conclusions, and their relevance to other aspects of the overall case.

(A) The first hardware sales and the origins of the hardware reselling strategy

In seeking to demonstrate its true purpose, it was important to both the Defendants and the Claimants to explain and establish the origins of the hardware reselling strategy.

The Defendants’ case centred in this respect on elaborating upon the risks to the software business which they perceived in mid-2009 and to which they contended the hardware reselling strategy was a strategic response, and upon the benefits to the software business which the programme was conceived and intended to secure, including closer associations with hardware suppliers and the ability to present itself as a one-stop shop to its customers.

It was common ground that Autonomy’s first substantial sales of hardware commenced in the early summer of 2009. Until then, Autonomy had purchased substantial amounts of hardware for deployment in its datacentres (principally for hosting and archiving services), but its sales of hardware were minimal. Reflecting their dispute as to the purpose of the programme, the parties disagreed as to what prompted the initial hardware sales.

The Claimants placed the origin of these sales, and what became a programme of hardware reselling, in the sudden and urgent need to find revenue in order to meet forecasts after (a) what Mr Hussain described in an email to Mr Egan and Mr Mooney (cc Dr Menell) dated 10 June 2009 as a “major disaster” in the form of the sudden collapse of two OEM deals (one with Adobe and the other with Microsoft) that he had confidently expected to complete by the end of Q2 2009, and (b) more generally, Autonomy’s uncharacteristic failure in Q2 2009 to achieve analysts’ consensus expectations of revenue in the quarter.

On the day before the email from Mr Hussain referred to in paragraph 699 above, which was subject-headed “OEM disaster”, Mr Hussain had notified Dr Lynch by email that the Adobe transaction was “not happening” , but he had cushioned that by stating that “MS frank cook actively considering hardware - so hope to replace”. In that email, Mr Hussain noted that “50 IDOL reps in the US…are delivering very little” ; and he put forward “MS hardware to have a total $10m” as first amongst other things to be done to ensure recovery.

On the terms of a one-off letter agreement reached on 30 June 2009 between Autonomy Inc and Morgan Stanley, Morgan Stanley agreed to purchase from Autonomy on or before 30 June 2010, $20m worth of Hitachi hardware (or other third-party hardware agreed between the parties) for a discounted price of $13.5m. At around the same time, Autonomy Inc entered into a “One-Time Reseller Authorization Agreement” with Hitachi, under which the latter authorised Autonomy Inc to resell Hitachi hardware to Morgan Stanley.

The overall effect of these arrangements was that Morgan Stanley bought the hardware at a discount from Autonomy as reseller, and Autonomy funded the discount by buying from Hitachi at full price and selling on at the discounted price, thereby incurring a financial loss.

The Claimants relied on these arrangements as demonstrating that Autonomy had resorted to the loss-making hardware sale to Morgan Stanley as an expedient to cover an unexpected shortfall in software revenue, and not to advance its software business. They also relied on an internal Autonomy email chain relating to the execution of the agreement with Hitachi which was copied to both Defendants to show that both were aware of these arrangements.

They submitted that this loss-making agreement with Morgan Stanley became the start of and template for “an extended and more ambitious programme of reselling pure hardware at a loss” in which Autonomy’s only involvement was its interposition as a seller into an existing deal for the supply of hardware by a hardware supplier to its customer, which might or might not be a customer of Autonomy. The benefit to the hardware supplier was that Autonomy paid it the full price whilst the end customer would purchase from Autonomy at a discounted price with Autonomy subsidising the difference or ‘delta’ : a ‘win’ for the hardware supplier, and a ‘win’ for the end customer: the only ‘loss’ was to Autonomy.

On the Claimants’ case, Autonomy’s only part in these sales to Morgan Stanley, and all that the hardware sales ever involved in terms of Autonomy’s role, was Autonomy’s interposition as a seller into an existing deal for the supply of hardware by a hardware manufacturer to its customer; and the quid pro quo was simply that Autonomy charged the customer substantially less for the hardware than the price that Autonomy paid to the hardware manufacturer (by itself taking the loss).

According to the Claimants, once its potential had been established, the hardware reselling strategy developed quickly to become one of the principal “levers” (another was the VAR sales) deployed by Autonomy at the instance of the Defendants in the context of concerns about shortfalls in software sales. Hardware (like VAR) sales enabled Autonomy to supplement or accelerate and manage revenue in each quarter to enable Autonomy to meet, and if possible, to “beat and raise” , analysts’ consensus expectations notwithstanding, and to cover, software sales shortfalls.

The Claimants contended that the context of the instigation of the hardware reselling strategy, and the reason why it quickly became and remained such an important and addictive strategy, was the constant pressure of the market perception of Autonomy and its stock as being, to quote Mr Marc Geall (“Mr Geall”), who was Autonomy’s head of Corporate Strategy and Investor Relations until May 2010, “an expensive growth stock that needs to ‘beat and raise’”. The problem with such a reputation is, of course, the difficulty of consistent fulfilment of ever-increasing expectations without regard to market conditions.

The Claimants illustrated this (and the volatility of market perception and in consequence Autonomy’s share price) by reference to analysts’ reaction to two announcements relating to Autonomy’s revenue forecasts or forward guidance about revenue expectations:

(1) An announcement by Autonomy on 7 July 2009, ahead of reporting its final figures, that it expected to report “revenues in line with current consensus estimates” [my emphasis]: so accustomed was the market to Autonomy exceeding analysts’ consensus expectations that merely meeting those expectations was a disappointment, and Autonomy’s share price immediately fell 8%.

(2) Further forward guidance on 16 July 2009 (when the Q2 2009 figures were announced) estimating for Q3 2009 $180 million for revenue and EPS at 19 cents per share, compared to an earlier market consensus forecast for the quarter of $198-199 million and EPS at around 26 to 27 cents per share. Autonomy’s share price fell again, so that by August 2009 its share price had fallen 18% since the pre-announcement of the Q2 2009 results.

The Claimants also referred me to the analysis and prescription put forward by Mr Geall in the light of (a) a continuing aggressive fall in Autonomy’s share price in mid-2009 and (b) a UBS note written by its house analyst, Mr Briest, dated 4 August 2009 suggesting that part of the reason for the decline was “disappointment at the level of Q3 guidance and the seeming loss of Autonomy’s ‘beat-and-raise’ status”. Mr Geall observed that:

“…lack of beat and raise in Q2 and the weaker Q3 outlook which means that there is a lot to do in Q4 is going to continue to get some legs. Obviously delivering $190 - 200m in Q3 will help off-set this fear”.

According to the Claimants, these circumstances show that the importance to Autonomy of shoring up its ‘beat and raise’ status, and the fact that in Q3 2009 and following quarters this was not possible to achieve on the basis of software sales revenues, drove the Defendants to the hardware reselling strategy. What the Defendants advanced as the purpose of the expedient was a pretext contrived:

(1) to justify sales of third party hardware by a software company which touted itself as selling only its own proprietary software;

(2) to justify the fact that such sales were not only so out of character but also at a loss;

(3) to justify the description and treatment of revenue from sales of hardware as all being sufficiently part of Autonomy’s software business to be included in software revenues without differentiation or explanation, and thus to obviate disclosure and prevent discovery.

The Defendants, however, submitted that it is the Claimants’ case which is contrived. Their case is that the hardware reselling strategy was not a response to the failure of any particular transaction, nor to concerns about diminishing or fluctuating results and the need to satisfy and exceed analysts’ expectations of revenue or other performance: it was a commercial response in the interests of the software business in changing, and possibly threatening, market conditions, and a successful and cost-effective marketing strategy.

As to contrivance, the Defendants dismissed the Claimants’ notion that the programme was directed to the sale of “pure hardware” as misplaced and the Claimants concept of “pure hardware” sales, as if there were some category of sales which by nature were unconnected to Autonomy’s software business as self-coined. Dr Lynch insisted consistently that Autonomy never did sell “pure hardware” : it was never interested or engaged in the sale of hardware as a separate line of business. For the Defendants, a mischaracterisation of the programme as almost a separate enterprise, and a “bizarre and expensive programme” to boost reportable revenue to cover any shortfall in software sales, rather than an extension of the way it needed to conduct its software business, lay at the heart of the Claimants’ mistaken approach.

An example, according to the Defendants, of the Claimants’ misunderstanding and mis-presentation of the nature of the hardware sales was the transaction with Morgan Stanley which the Claimants relied on as marking the beginning of and establishing the pattern for the hardware reselling strategy. The Defendants contended that the Morgan Stanley transaction should not be regarded as a sale of “pure hardware” even on the Claimants’ definition of the term. They pointed out that:

(1) It was concluded on the same day as, and should be considered in the context of, two software agreements with Morgan Stanley, which was a longstanding software customer of Autonomy.

(2) In the contractual documentation the link between the software and hardware transactions was explicit; and it also made express provision for Morgan Stanley to deploy the equipment purchased not only in any of its own but also in Autonomy’s (or any third-party service provider’s) facilities, connoting that Morgan Stanley wanted the flexibility to use the hardware with Autonomy’s software in Autonomy data centres.

(3) The hardware transaction should be regarded, in effect, as a “strategic package” or “appliance” sale, in the broader sense which the Defendants supported: a sale of hardware for use with Autonomy software.

(4) Mr Egan, who was involved in the transaction as the relationship manager with Morgan Stanley, accepted in evidence that the document made that link clear and obvious.

(5) Furthermore, in an email chain in May 2010 involving Mr Scott, Mr Hussain and Mr Egan, which is not said to be pretextual by the Claimants, Mr Egan referred to the sales as part of “a large solution sold to them which included HW, restructuring the safe, and the SW licences to cover ILM” and Mr Scott responded on 10 May 2010, stating inter alia:

“…I don’t know what TJs take is on this or whether he feels $2m is suitable but we obviously had to make major concessions across the board to get the Morgan business, including giving Morgan a very aggressive discount on their hardware purchase commitment in June 2009, the same time at which we set the $2m option for Morgan to purchase the ILM license”.

The Defendants contended that, contrary to the Claimants’ case that the Morgan Stanley hardware transactions had nothing to do with protecting and promoting software sales, the discount funded by Autonomy in respect of the hardware sale to Morgan Stanley was the price that Autonomy felt it had to pay to be sure of securing the high-margin software business and (as Dr Lynch put it) “snuggling up” to both hardware suppliers and the ultimate customer.

Thus, the dispute as to the origins of the hardware reselling strategy reveals the main issue between the parties as to the purpose and objectives of the hardware reselling strategy.

(B) The Defendants ’ case as to the purpose of the hardware reselling strategy

I elaborate the Defendants’ case under the following headings:

(1) The Loudham Hall meeting in July 2009 and Dr Lynch’s evidence as to the development and rationale of the hardware reselling strategy.

(2) The Defendants’ depiction of the evidence of Mr Egan and Mr Sullivan as to their understanding of the purpose of the programme as initially conceived and thereafter implemented, and of Mr Welham’s evidence (which is said by the Defendants to be corroborative).

(3) Autonomy’s relationship with EMC: the development of their relationship in Q3 2009 and the dispute as to the reasons for EMC bringing an end to its participation in the hardware reselling programme at the end of the same quarter.

(4) The replacement of EMC and the continuation and expansion of the hardware reselling strategy with Dell (after EMC withdrew) in every quarter from Q4 2009 to Q2 2011 and (on a more limited basis) with Hitachi.

(5) The further evidence relied on by the Defendants as showing the use and success of the hardware reselling strategy in protecting and promoting Autonomy’s core software business.

(6) The Defendants’ case that the Claimants’ assertion of secrecy is unsustainable: and that the hardware reselling strategy was openly recorded in Autonomy’s ledgers and discussed within Autonomy, fully explained to Deloitte and approved by them, and had become well known in the market;

(7) The Defendants’ honest belief in the rationale of the purpose and the support given to that and the propriety of the way the programme was presented and accounted for by the approval of Deloitte and the Audit Committee with full knowledge of the facts.

(1) Dr Lynch’s evidence as to the rationale of the hardware reselling strategy

Just as the transaction with Morgan Stanley was not, in the Defendants’ presentation, an expedient to pump up revenue, but a strategy to promote a software sale, so too the Defendants presented the hardware reselling strategy as a carefully weighed strategy to safeguard its software business, which was presented and launched, and thereafter openly discussed, within Autonomy.

There were two main limbs of the Defendants’ justification of what they often referred to as “strategic hardware sales” , though the ultimate objective was the same: to drive software sales. One limb was essentially defensive: it was to meet the risk posed by two market changes which Dr Lynch considered threatened Autonomy’s software business. The other was more proactive: to use hardware sales as a marketing tool.

At what the Defendants depicted as the formal presentation and official launch of the strategy in July 2009, at a management meeting which took place between 8 to 10 July 2009 at Dr Lynch’s country house in Suffolk, Loudham Hall (“the Loudham Hall meeting”), the emphasis was on the marketing strategy.

I was not shown any agenda for or contemporaneous record of the Loudham Hall meeting.^{^[131]} Dr Lynch told me that it was unlikely that there was any: “We didn’t take notes at the meetings”. But he described it in his witness statement, as follows:

“In July 2009, at a management offsite meeting in Loudham, Messrs Hussain, Egan, Scott, Mike Sullivan and I discussed Autonomy’s hardware strategy. I believe that around 20 people were present at the management offsite meeting, so there would have been several other people present during this discussion. Although I cannot recall precisely what we discussed, the gist was that we should sell hardware to the company’s biggest customers in an effort to drive software sales. The commercial strategy was not to sell hardware at a loss in order to inflate revenue, it was to sell hardware to strategic customers, even if we made a loss on the hardware sales, for the dual purpose of servicing our strategic customers at a time of industry consolidation and to improve our position with hardware manufacturers with which we could develop appliances. We discussed the level of sales we would aim for in pursuit of these objectives. I cannot remember if we decided on $10M as HP claims, or some other figure. We agreed that Mr Sullivan would take the lead in this. In that context, HP claims that Mr Hussain or I joked that we would buy Mr Sullivan a Porsche if he succeeded in executing the strategy. He was not given a Porsche; his remuneration was based on a sales package, as I explain below.”

That description of the presentation at the Loudham Hall meeting focused on what the Defendants often referred to as “strategic hardware sales” as a marketing strategy and does not mention the other more defensive rationale for the programme which was asserted by the Defendants. In his witness statement Dr Lynch explained those threats he considered were posed by “a few key market changes”. In more detail:

(1) Dr Lynch perceived that in the wake of the financial crisis in 2008-2009 customers had started consolidating their vendor lists and limiting what suppliers they would use because they thought that it would make procurement cheaper. Dr Lynch described this when cross-examined as “an industry trend at the time” . He was concerned that some of Autonomy’s key customers, in particular financial institutions, were reducing the number of approved suppliers or would be likely to do so, and that if Autonomy was removed from the list, Autonomy would likely be forced to sell via a third-party provider still on the list. This would make Autonomy reliant on another company (which could well also be a software competitor) for the sale of software and lead to an erosion in its margins because the third-party supplier would demand a margin on sale (typically 30%). He added:

“We held internal discussions as to what we should sell to increase overall volumes of sales to these customers. We concluded that we did not want to sell third-party software owned by our competitors and we did not want to sell services because that was against the ethos of the company and carried significant overhead. Hardware was a good option as all our customers were also hardware customers^{^[132]} , and it supported our software business.”

(2) Dr Lynch’s objective was “strategic package sales” or “ one-stop shopping” under which Autonomy would supply all customers’ IT needs (hardware as well as software) as a means of tying such customers into Autonomy and reducing the risk of customers looking to mainstream hardware suppliers which might seek to introduce other software companies. In cross-examination, Dr Lynch elaborated that:

“So we were very, very keen to avoid that happening and we spoke to the banks and one of the things they said to us is: there’s going to be an element in this of how much you are selling us. And it was actually a suggestion that came from the banks that hardware was something they would be happy to do this way.

That was the customer-facing part of the strategy and it turned out to be very successful and unlike companies that were selling more than us, we did not get pushed into being a supplier through one of the majors and that saved us 30%. So the amount of business we’re talking about here is about $150 million, so that saved us $50 million profit immediately, so it was a very good outcome.”

As to the second threat Dr Lynch identified, which was the perception that the software industry appeared to be moving towards appliances as the means of packaging and selling its offering:

(1) Dr Lynch’s assessment was that the way software was delivered was changing. At that time, the Cloud (which in fact ultimately became, to Autonomy’s advantage, the predominant delivery mode) was in its early development: appliance sales appeared at the time to be the most likely platform and mode of delivery of software offerings.

(2) The threat was that companies focusing exclusively (or almost exclusively^{^[133]} ) on software, and which could not offer any packaged solution, would find it difficult to compete with large companies that had the capability to provide both. Dr Lynch gave as an example, Google’s development and production of an appliance that overlapped with some of Autonomy’s products; he was concerned that some analysts, such as Mr Whit Andrews of Gartner (an influential industry analyst), had identified Google Search Appliances as one of the main competitive threats to Autonomy.

(3) Dr Lynch described the perceived likely industry move to appliance sales as “an existential threat to our software business” , unless Autonomy could establish a relationship with one or more hardware suppliers to develop an appliance or packaged solution which it would be to their joint benefit to market in a combined product or (as it was often referred to) “solution” . For that purpose, he considered, Autonomy needed to establish strong “strategic relationships” with hardware suppliers (what he described as a “partnership that we could leverage and rely on if the market did aggressively move (as feared) to appliances” ) so as to be able to procure appropriate basic hardware ( “without unnecessary bells and whistles” ) at a competitive price. That hardware could efficiently and economically be configured with Autonomy software and then either sold and delivered as a package either with Autonomy software embedded or already configured for Autonomy software to be supplied later. (Either package was, as Dr Lynch used the term, capable of being an “appliance”.)

(4) Dr Lynch’s evidence was that his perception was that:

“Autonomy’s hardware sales helped us overcome that threat: they enabled us to partner with hardware suppliers to purchase appropriate hardware at a competitive price, develop an appliance and meet the perceived market demand.”

When the “existential threat” subsided, as in the event it did with the increasing dominance of the Cloud as the means of both the delivery and the archiving of software, the assistance of a hardware supplier to develop an appliance became much less of a concern. Dr Lynch accepted that the weight given to the various objectives of the hardware reselling strategy shifted from about Q2 2010 onwards, and became more centred on what he called “strategic package sales” and the “customer-facing aspect” : or, in other words, positioning Autonomy to make software sales by becoming the customers’ preferred supplier across the spectrum of its IT needs.

Dr Lynch always emphasised, however, that it was never intended that Autonomy should become, and he insisted that it never did become, a generic reseller of hardware. In his witness statement his evidence was that:

“…we did not seek to sell hardware into companies who were not in the market for our software and we did not support the hardware we sold with related services . We sold hardware to our actual or prospective software customers, for a variety of reasons connected to our core business, which I will explain…”

That is consistent with what he said close to the time, for instance, in an email dated 23 October 2009 to Mr Barry Ariko, a member of Autonomy’s Audit Committee:

“ Personally despite the benefits you cite I feel we would not want to become resellers of unrelated hardware which is not about furthering our software sales. ”^{^[134]}

It is also consistent with Mr Sullivan’s evidence that Dr Lynch explained from the outset that the aim of the hardware sales was to get a bigger share of the financial market and in particular, big banks’ software business. (As Mr Sullivan told the US court, those banks and financial institutions were his largest customers: that marked him out as the logical person to drive forward the programme.)

Dr Lynch also referred in his witness statement to a number of “additional benefits to the hardware strategy” , some anticipated, and some which emerged, including:

(1) Autonomy could leverage its buying power to negotiate discounts on hardware needed for its own data storage centres.

(2) Many of Autonomy’s customers had procurement departments, the performance of which tended to be measured on the discounts they obtained. Autonomy found that, if it sold hardware at a loss, it could more easily maintain its margins on its software. If Autonomy failed to maintain its margins on its software with one customer, the company faced erosion of its margins with all of its other customers. Commercially, it was better for Autonomy to discount certain hardware sales to large customers with strong negotiating positions in order to maintain (typically very much more substantial) margins on its sales of software.

Dr Lynch did not dispute or seek to disguise the fact that a further material but collateral advantage was that the hardware strategy gave “a degree of flexibility” and “a little bit of a degree of freedom” which enabled Autonomy “on occasion” to “plug gaps in relation to other revenues” . Dr Lynch confirmed further that he was sure that the freedom and flexibility had been used. However, he was adamant that this was not, and never became, the driving force or reason for the strategy.

He also contended that hardware sales would be a poor tool for inflating revenue, and that if that had been Autonomy’s goal, it would have been easier and more efficient to sell third party software. He explained in cross-examination:

“…If the goal was to produce recognisable revenue, we would have just resold third party software, which would have been a far more efficient way of doing things. It would have not had the delivery issues. So one of the big problems with this as a strategy for trying to get revenue of course is that hardware has to be physically delivered. So you've got 30% of your business in the quarter is done in the last week and generally, although there can be other arrangements, that's too late to turn on a hardware tap. So if the goal of the exercise was to –- solely to get revenue, you would resell third party software, which you could do on exactly the same basis but obviously it wouldn't meet our strategic goals.”

On that point, Ms Harris also gave evidence about the relative complexity and timing problems inherent in hardware sales. Whereas with sales of software, delivery was effected by sending an email key, and revenue could be recognised straight away provided the other revenue recognition criteria were met (e.g. collectability), the process was much less simple for hardware, with no guarantee that the revenue would be recognisable in the quarter: again, it was suggested that this meant that hardware sales would not have been an effective tool to enable Autonomy to “plug gaps” in order to meet its targets.

In the round, Dr Lynch portrayed the hardware reselling strategy as “the right decision and a good insurance policy for the company”. He was also adamant that at a net cost of no more than $20 million the hardware reselling strategy was good value for money in marketing terms. He submitted that when looking at the commercial sense of the strategy, it is not the size of the revenues that matters, but the cost to Autonomy. When cross-examined he expressed this as follows:

“So if we just look at this in context of what’s actually going on here. Autonomy spends $500 million over this period on marketing. This strategy costs us about $20 million and we get back through the fact that the linked discounts come - meant that we don’t have to discount software to get the procurement discount, so it costs us less than $20 million. We then get EMC, for example, bringing us into their extended deals where they replace EMC with us and that goes into multiple customers. Dell bring us into 40 customers. They all make appliances for us. We displace our main competitor FAST out of Hitachi so we actually become the software for Hitachi. And we protect losing the profit margin on about $160 million of software business. So what’s going on here is the execution of what actually is probably the best $20 million I ever spent on marketing in terms of the return that we got for it.”

He summarised what he saw as its advantages and successes in his witness statement, as follows:

(a) “Autonomy maintained its strategic supplier status with its major customers.

(b) Autonomy demonstrated that it was not threatened by the move to appliances.

(c) Autonomy protected its software margins.

(d) Autonomy developed its appliances with Dell, and sold them to customers.

(e) Autonomy and Hitachi were developing an appliance as well.

(f) Autonomy and EMC agreed to develop an appliance^{^[135]} .

(g) Autonomy and EMC entered into discussions for the purchase of Documentum^{^[136]} for $2.2bn (which did not ultimately proceed).

(h) EMC changed the code on its hardware to enable it to run Autonomy software, leading to sales.

(i) Autonomy obtained discounts worth several million dollars on its own hardware purchases.”

Dr Lynch could not and did not shy away from the fact that the hardware sales were, on a stand-alone basis, almost invariably loss-making; but his case was that when the benefits (especially increased high-margin software sales, but also including discounts on hardware purchases for its own considerable needs in its archiving and hosting offerings) were weighed in the balance, the costs were outweighed by the advantages.

In the round, the Defendants stressed especially the importance of putting the programme in proper perspective: when looking at the commercial sense of the strategy, it was not the size of the revenues that matters, but the cost to Autonomy; and even the gross cost was small in comparison to the very high margins that could be made on software sales, as Mr Welham agreed.

There is no evidence of the hardware reselling strategy having been discussed by the board: the strategy was presented as having been discussed at the Loudham Hall meeting and developed thereafter.

(2) Witness evidence which the Defendants contend supports their case

As indicated previously, the Defendants reinforced their case as explained by Dr Lynch principally by reference to the evidence of the Claimants’ own witnesses. They relied in particular on the evidence of Mr Sullivan and Mr Egan, which they contended supported their case as to the purpose of the hardware reselling strategy in all its aspects, and on the evidence of Mr Welham.

Both Mr Egan and Mr Sullivan were directly involved in the hardware reselling strategy, which was almost exclusively centred in the US. Both had given evidence in the US criminal proceedings. Both appear from that evidence to have considered that there was a valid commercial rationale for the hardware sales, including sales of what the Claimants would characterise as “ pure hardware ”. The omission from their respective witness statements of any substantial recognition or explanation of the outlook they expressed at the time has tended rather to emphasise the Defendants point that the Claimants’ case now is not supported by their own witnesses.

Mr Sullivan

In his witness statement, Mr Sullivan did not express any view as to the purpose of the strategy of hardware sales, which he considered began to be implemented in Q3 2009 after the meeting at Loudham Hall and his return to the US after that. He accepted that it was not unusual for software companies to sell hardware. However, the emphasis in his witness statement was on the fact that Autonomy’s hardware sales were at a loss; and that although he was familiar with the process of reselling hardware, had engaged in it when at SteelPoint, and had a well-established network of large institutional customers to which hardware could be sold, he stressed that in his experience there the sales had been for a profit and never at a loss.

Mr Sullivan explained that he was in regular contact with Mr Hussain, who regularly set revenue targets for hardware resales and frequently requested updates on the status of hardware sales (which, he added, were often referred to as “low margin” business). He acknowledged that he was offered and received bonuses if he achieved the targets. He identified EMC as the initial usual hardware provider; they were replaced by Dell after Q3 2009.

His evidence was that after the HP acquisition Dr Lynch and Mr Hussain agreed that reselling hardware should cease except for committed deals, though he was asked to do some new deals after that; and that the programme finally came to an end in April 2012. He seemed to imply that termination post-acquisition signified some suspicion about propriety.

The Defendants depicted Mr Sullivan’s evidence in the US criminal proceedings as giving a fuller and fairer picture of his view of the hardware reselling strategy. They relied on his evidence in that context as demonstrating that in his perception:

(1) There was nothing wrong or illegal in a software company selling hardware at a loss as part of a marketing programme: he considered that it was something that happened frequently in the software and hardware industries. He explained his own thought process as follows:

“Absolutely. I mean companies sell printers at a loss to sell ink. Companies sell razors to sell blades. You know, we were a software company. We gave away professional services for free to get software deals. We gave away training for free to get software deals. I mean, in the end what matters is you make more than you don’t make when you add it all up and group it together. That’s the way I looked at it.”

(2) Autonomy and its then preferred hardware provider, EMC, had mutual customers (in particular, large financial institutions) and it was with a view to mutual benefit that “we specifically targeted, you know, our mutual customer list to provide this benefit too, yeah.”

(3) Autonomy wanted the hardware manufacturers to drive sales of Autonomy products, and there were opportunities for mutual benefits if they did. For example, if a manufacturer’s introduction led to Autonomy selling Digital Safe to a customer, that in turn would increase Autonomy’s or the customer’s need for hardware, which would result in further hardware purchases from the manufacturer.

(4) After mid-2009, when he started selling the low-margin hardware, he only sold it to big software customers of Autonomy and such customers included Citigroup, Bank of New York, and Bloomberg. He explained that a lot of software was sold to a company called Insight which was Citi’s purchasing agent for hardware so that when they were selling to Insight they were selling to Citi. He also confirmed that a company called SHI played the same function for Bank of America as Insight for Citi and so if they were selling to SHI they were selling to Bank of America. All of those financial institutions were software customers of Autonomy before the hardware program started.

(5) He agreed that the programme or strategy was “a way to develop good relationships and to induce further business…an investment in the relationship…”

(6) It was never suggested to him that he should, and he never did, try to conceal hardware sales, saying that he “never thought there was a reason to. In fact, I tried to make it clear”. He expressly told both EMC and Dell that the sales were part of a marketing programme, and never thought there was anything at all wrong in it.

(7) He saw nothing wrong in either the payment of bonuses, nor in the setting of targets measured by sales under the programme: he said he did not know “why you would do it any other way”.

(8) It was impressed on him by Mr Chamberlain and the legal department that to that end, it was essential for Autonomy (a) to take title and own the hardware before reselling it (as he put it “We had to hold the receivables risk…” ) and (b) to ship the product before the end of the quarter for which revenue was to be recognised: if shipment was delayed, then revenue recognition would have to be deferred (which he said “happened regularly” ).

The Defendants relied also on Mr Sullivan’s evidence when asked in the US criminal proceedings in “direct” examination what in September 2010 he understood the purpose of the “Dell resells” to be:

“Well, clearly there was an element of wanting more revenue. That was one. But there was also the overarching purpose…The overarching purpose, again, as I asked multiple times, was to generate good will and marketing benefit…”

However, Mr Sullivan’s evidence in the US criminal proceedings also signalled that he was not himself verifying the rationale of the programme: he was explaining what he had been told by the Defendants and that that was the basis on which he had proceeded. Thus:

(1) When asked whether the explanation of the strategy in focusing on big financial institutions and selling them discounted hardware which he was given by Mr Hussain and Dr Lynch, which was that this would help to develop relationships with them with a view to getting a bigger share of their software business was plausible, he side-stepped, and simply answered that the big financial institutions were his “largest customers for the portion of the business that I ran.”

(2) When asked whether he knew that the marketing aspects of discounted hardware sales were important to, for example, Mr Hussain and Dr Lynch he said he did “because of their words and that’s what they told me.”

(3) He was aware that it was important for Autonomy that the sales of hardware should be papered and structured to result in recognised revenue: he accepted that revenue recognition was part of the rationale.

It is clear from Mr Sullivan’s evidence in the US criminal trial that he was personally unsure what was the purpose of the sales of hardware at a discount on the price which Autonomy had paid the supplier and thus at a loss to it. The fact that he repeatedly asked Mr Hussain and Dr Lynch for an explanation suggests that he remained unsure and needed reassurance from them. His evidence was replete with the variety of answers he was given, though all revolved around the central theme of fostering goodwill with suppliers and software customers, and mining that to develop and extend relationships with them both.

Of course it is dangerous to place too much reliance on nuance having not had the opportunity to watch and listen to him being cross-examined, but I also formed the impression from the transcripts of his evidence in the US criminal proceedings that he was ultimately unsure quite what the story was that he had been told and tended to grasp at different formulations according to whether he was being examined or cross-examined, being anxious to satisfy both and insulate himself.

Further, although Mr Sullivan well understood and accepted the notion of a promotional deal and using discounting for marketing purposes (see, for example, paragraph 801(1) below), when confronted with particular transactions, he often had to accept that the reality was that the real driver was revenue generation albeit at a loss. Thus:

(1) As elaborated below (see paragraph 773 et seq ) in considering the nature of the relationship between Autonomy and its first major hardware supplier, EMC, Mr Sullivan confirmed that though he “certainly characterised the deal as, you know, a marketing deal” and had told them that he “had marketing funds to use to spend on some of our mutual customers” the reality was that:

“You know, really, the deal that I had with EMC was just to help us sell some hardware. It really didn’t go much beyond that.”

(2) In relation to Autonomy’s dealings with a Dell customer called Zones Inc, which Dell introduced to Autonomy (and which I address in more detail later, see paragraph 1159 et seq ) Autonomy was contractually precluded from revealing its involvement. Mr Sullivan duly confirmed that he never mentioned Autonomy’s involvement in any transactions with Zones, or in other transactions with Dell customers such as Oracle where like prohibitions had been imposed.

(3) Mr Sullivan had to accept, in relation to an email from Mr Hussain dated 25 March 2011 in which Mr Hussain had written that making further hardware sales was “the last part of the low margin jigsaw” , that this was being done to try to hit a revenue number, rather than for any marketing purposes.

More generally, Mr Sullivan’s role, as the person chosen to develop business with hardware suppliers (first, with EMC and subsequently with Dell), was defined exclusively by reference to revenue generated from hardware sales. He reported regularly to Mr Hussain as to (a) what “low margin” sales could realistically be achieved and (b) the value, purely in revenue terms, of hardware sales secured from time to time. In other words, the accent was almost exclusively on revenue, without a thought to the pursuit or accomplishment of any other benefit. In that connection, Mr Sullivan stated in his evidence in the US criminal proceedings that he did not know whether there was any effort at all at head office in England to try to track or relate “low-margin hardware deals with software sales”.

My overall impression was that Mr Sullivan had been carefully assisted as to the line to be followed for his own protection if and when questioned about the purpose of the hardware sales. He tried to repeat his lines (with some glitches in his recital of them), that he had been assured and had then persuaded himself that there was nothing improper in what was being done. However, although I should stress that the Claimants did not accuse Mr Sullivan of dishonesty, he knew that it was all about revenue, with any other benefits being uncertain and peripheral, and unproductive in terms of his take-home pay.

Indeed, in a passage from his evidence which was not quoted in the Defendants’ respective closing submissions, even though it was part of the same passage quoted at the end of paragraph 742 above which was especially relied on by Dr Lynch, Mr Sullivan appears to have accepted that where the deals were arranged at the very end of the quarter, their real purpose was simply to generate revenue. Thus, when giving his answer that “the overarching purpose” was to generate “good will and marketing benefit” he expressly differentiated what he referred to as “the last second request”. He was asked to explain what he meant by this:

“Q. What do you mean by the last second request?

A. Well, in this case this is two days before the end of the quarter

Q. What significance do you attach to that?

A. To get more revenue for the quarter.”

Mr Egan

As mentioned above, Mr Egan (like Mr Sullivan) said very little in his witness statement about the rationale of the hardware sales, and he tended to minimise his own involvement. However, Mr Egan (unlike Mr Sullivan) did attend to be cross-examined via video-link in these proceedings; and when cross-examined he confirmed that he did not think there was anything wrong with Autonomy’s strategy of selling hardware, even though it typically involved a loss on the resale.

Mr Egan also confirmed that the evidence he had given in the US criminal proceedings was correct, and could be taken still to be his evidence. The gist of that evidence was captured in the following exchange:

“Q. Did you believe that the sales to these banks of hardware had the extended benefit of enhancing the relationship with the banks?

A. I did believe that.

Q. Who were big customers for software?

A. Some, yes; some targets.

Q. Did you consider it a relationship builder to sell the hardware to these targets or customers for software?

A. I did.

Q. And did you see it as a benefit to future software purchases?

A. Absolutely.

Q. Did you believe that hardware and software were leveraged together with these big customers?

A. Often, yes.

Q. And did you believe it increased Autonomy's standing with the big financial institutions that bought their software?

A. The hardware sales?

Q. Yes.

A. Absolutely.”

Mr Egan’s evidence thus seemed directly to corroborate Dr Lynch’s explanation of the purpose of the hardware reselling strategy, and (perhaps more importantly) that the purpose was pursued and seemed to him to be realised in its implementation. Again, the absence of any acknowledgement of this in his witness statement tended to support the Defendants’ submission that the Claimants had provided no reason to suggest that the rationale was contrived or implausible.

The Claimants pointed out, however, that Mr Egan was never once asked to identify the basis of his stated belief as to the purpose of the hardware sales. Apart from the so-called “Linkage Analysis” which I refer to later, no objective evidence such as to inform or validate the statements of belief was ever provided. They contended that it is clear from other parts of his evidence that his understanding was based on discussions with Dr Lynch and Mr Hussain, and cited in particular the following evidence in his witness statement:

“I understood from discussion with Mr Hussain and Dr Lynch that a reason for making these sales was to engender goodwill with customers that were also customers for Autonomy software; and that was always a stated reason for these money-losing sales.”

Further, Mr Egan, who (with Mr Sullivan) was present at the Loudham Hall meeting in July 2009, but did not address it in his witness statement, implied that the notion of hardware sales driving software sales might have informed the strategy initially, but soon faded when the revenue-generating potential of hardware sales was recognised. He stated in his witness statement that:

“…it was also clear to me over time that a principal reason for these money-losing sales was to generate revenue that could be recognised in the quarter in which the sale was made so that Autonomy could achieve its revenue target and meet market expectations. This benefit was discussed as a primary motivation for doing these deals amongst myself and others, including Mr Hussain and Dr Lynch.”

Mr Egan agreed with Mr Miles in cross-examination that this was carefully phrased, and that the use of “a” rather than “the” before “principal reason” was deliberate and what he was “comfortable with”. That was not least perhaps because in his evidence in the US criminal proceedings Mr Egan (when cross- examined in that context) had accepted that he had regarded hardware sales, and in particular sales to large banks who were big customers for software, to be “a relationship builder” so that “hardware and software were leveraged together with these big customers” . He was careful to maintain the position that even if, “later in the process”, such reasons “became secondary in a couple of cases to the absolute desire for revenue” , they always, in his view, subsisted.

Most importantly, at least from the Defendants’ perspective, Mr Egan’s evidence was that he considered that it was genuinely always part of the purpose of hardware reselling to turn Autonomy into a favoured seller to large customers and particularly banks and other financial institutions, and that engendering goodwill with customers that were also customers for Autonomy software was not only a stated but “always a very big reason” for the money-losing hardware sales.

Mr Egan made the point also that “…you could have an argument in any one transaction about whether something was secondary or primary to the other…” He accepted that the evidence he had given in his witness statement that revenue generation had “been discussed as a primary motivation for doing these deals amongst myself and others, including Mr Hussain and Dr Lynch” was not based on any actual or specific conversation. He accepted as correct Mr Miles’s suggestion to him that the notion of a primary motivation for a deal:

“…sounds rather like the kind of thing a lawyer would say rather than what business people would say at the time…”

That is probably realistic. Different transactions may have differently weighted purposes and objectives, and business decisions are usually multi-faceted; the weighing of which is the primary or predominant purpose is beloved by lawyers, looking at the matter after the event; but commercial men tend to take a more holistic approach and weighting as between various commercially welcome effects is not the reality of commercial life. Mr Egan sought in this way to reconcile his evidence in the US criminal proceedings, his witness statement, his conscience and what increasingly also struck me as a desire so far as possible, not further to implicate Dr Lynch, or indeed himself ^{^[137]} .

This difficult line was easier to follow where Mr Egan was involved in, or did know of a prior or contemporaneous hardware sale; and it is clear that in those instances, Mr Egan did try to leverage the hardware and software sales together.

The Defendants made much in this regard of an email exchange between Mr Egan and Mr Hussain on 23 December 2009 which had as its subject heading “A few additional points for your prep”. This exchange (which Mr Egan had not mentioned in his witness statement) appeared to suggest that when ‘prepping’ Mr Hussain how best to present to Mr Christian Lucas of Morgan Stanley a large proposed deal with Morgan Stanley which also involved a ‘loss leader’ sale to Morgan Stanley of Dell hardware at a discounted price, Mr Egan had presented what was described as the “Dell/reseller” deal as a loss leader which promoted good relations with Dell and led to faster adoption of Autonomy software and “grease[d] the process of lucrative OEM bundling of Autonomy software on Dell kit” especially when Dell could see “the level of usage of our software in a prestigious account like MS…”

The Defendants also relied on a passage of evidence from Mr Egan’s re-examination in which Mr Rabinowitz had sought to elicit from him that the use of hardware sales to leverage software sales in the Morgan Stanley deal was a single exception, and that Mr Egan had not been involved in any other hardware sales which were linked to a sale of software. Mr Egan, contrary perhaps to the re-examiner’s expectation, replied that he was sure he was, causing Mr Rabinowitz quickly to break away from that line of questioning.

However, the surprise of the initial answer may have somewhat obscured Mr Egan’s addition, which explained its limited scope:

“To be clear, I answered that question based on knowing that I’ve sold what I believe I referred to the other day as more occasional hardware in conjunction with software versus very large hardware deals.”

Earlier, in the course of his cross-examination, Mr Egan had referred to having

“sold some incidental hardware prior to the period that we’re talking about and I don’t think anyone has questioned, or that’s not under scrutiny.”

What in the round I took from Mr Egan’s evidence was that:

(1) Although he knew of money-losing hardware sales, and of their general justification as “to engender goodwill with customers that were also customers for Autonomy software”, he was “more tasked with selling software” and it was Mr Sullivan who “was tasked by Dr Lynch and Mr Hussain with managing most of the hardware reselling.” In other words, the tasks were usually separately undertaken, and Mr Egan was only occasionally involved in hardware sales.

(2) “As a general matter, [he] did not know which companies had purchased hardware from Autonomy or how much third-party hardware was being sold”; and he was not, therefore, in a position to and did not “use the fact of a prior or contemporaneous hardware sale to promote the software sale [he] was trying to make”.

(3) Even in some cases where there were long running negotiations for software sales he was simply not told or aware of the fact of a prior or contemporaneous hardware sale which might have assisted him in concluding a software sale. The example given in his witness statement of a software licensing and data hosting transaction eventually concluded with BofA was especially instructive because not only did he not know of a prior large hardware sale to BoA which might well have assisted him in his negotiations but also he “never heard Dr Lynch or Mr Hussain…refer to these hardware sales during our extensive negotiations with Bank of America” either).

(4) As to what he described in cross-examination as the “small universe” of large value end of quarter hardware sales, his evidence to me was that in his discussions with Dr Lynch about this “I feel as though we had an understanding that they were needed from a revenue perspective…”.

(5) It was, inferentially, for that reason that he was not told and usually did not know any of the details of such sales, or (in particular) which companies had purchased hardware from Autonomy or what the sales involved.

I accept that evidence. If Mr Egan, who was so often responsible for software sales, (a) in important instances where there were ongoing negotiations for software sales did not know and could not and did not use the fact of prior or contemporaneous hardware sales as a negotiating point , and (b) often did not know relevant details of end of quarter hardware sales and in any event understood them to be directed simply to revenue generation, that tends substantially to undermine the notion that the objective of such hardware sales was to benefit software sales and that “hardware and software were leveraged together with these big customers”.

This fed through into the main difficulty for the Defendants. This was that although Mr Egan sought to emphasise what he called (see paragraph 751 above) the “extended benefit” (of enhancing relationships with the banks), the overall impression persisted in his evidence that the sine qua non and driving force of the hardware sales was revenue generation to meet what he knew was the Defendants’ desire and need to satisfy revenue forecasts.

That impression was supported further by other parts of his witness statement which illuminated how little part any ancillary benefits for software sales actually played in the planning and practical implementation of Autonomy’s hardware sales. To exploit an advantage in terms of software sales that might be gained from a sale of hardware, the software salesman had to know of the hardware sale; but in his witness statement, he stated:

“As a general matter, I did not know which companies had purchased hardware from Autonomy or how much third-party hardware was being sold. When I was involved in selling software licences to end-users, I usually did not know whether hardware had been sold to that end-user and when I was not involved in, or aware of, a hardware sale I did not use the fact of a prior or contemporaneous hardware sale to promote the software sale I was trying to make.” ^{^[138]}

It may also be noted that Mr Egan stated in re-examination that: (i) he was unaware that each quarter from Q3 2009 to Q2 2011 the Defendants set hardware revenue targets for Mr Sullivan; (ii) he was generally not aware of the dealings that took place between the Defendants and Mr Sullivan in relation to the generation of revenue through hardware sales; and (iii) he did not know that bonus payments were offered and paid to Mr Sullivan linked to the amount that he generated from hardware sales. All were factors tending further to undermine any belief that the hardware reselling strategy was intended to generate and increase software sales.

My overall assessment is that Mr Egan’s evidence was more a recital of reasons he had been given as to the purpose of the hardware reselling strategy dressed in the language of belief, and that the real import of his evidence was that his own experience in practice was that the hardware reselling strategy had very little relevance to, and no measurable benefit for, the software business, and that it was the need to show revenue which was its true explanation.

Dr Lynch

Dr Lynch’s evidence did not attempt to show any direct or causative link between hardware reselling and increased software sales. His justification was stated in general terms with little or no evidence-based examples:

(1) His overall justification was to depict the programme as similar in intended effect to an advertising campaign, but (at a net cost of about $20 million) cheaper.

(2) He offered the further justification that the discounts which Autonomy offered its software customers on hardware sales could be used to maintain prices for software products, because it meant “that Autonomy did not have to discount software sales in the same amount”.

He contended also that what he called the customer-facing part of the strategy turned out to be very successful, and that the effect of promoting good relationships with hardware suppliers was that:

“ unlike companies that were selling more than us, we did not get pushed into being a supplier through one of the majors and that saved us 30%. So the amount of business we’re talking about here is about $150 million, so that saved us [$]50 million profit immediately, so it was a very good outcome .”

As will be seen later, these issues were reflected both in a ‘Strategic Deals Memorandum’ , which referred to the “ continuing rationalization of IT suppliers to major financial institution customers ”; and also in Autonomy’s Quarterly Results for Q3 2009, which noted that “ During the quarter we saw some of our large customers promote Autonomy to strategic supplier status. ”

However, none of this could explain:

(1) The frantic use of hardware reselling when gaps emerged in expected revenue from software sales;

(2) The determined efforts not to disclose the hardware reselling strategy, to the point of accounting for the costs in such a way as to minimise the risk of revelation as well as reducing the adverse effect of loss-making sales on the ‘bottom line’;

(3) The repeated efforts to paint a misleading picture of Autonomy’s efforts to establish a partnership with the first of its main hardware suppliers, EMC (to which I now turn).

(3) Autonomy’s relationship with EMC and the dispute as to why it abruptly ended

Mr Sullivan’s evidence on Autonomy and his own relationship with EMC was:

(1) Whilst at Zantaz (which had purchased servers and storage from EMC for use in its hosting business) and then after the acquisition of Zantaz by Autonomy, Mr Sullivan had established a good working relationship with EMC’s director of Global Sales and Marketing Operations, Mr Bill Scannell (“Mr Scannell”).

(2) The potential for some collaboration with EMC preceded the Loudham Hall meeting. Autonomy was, even before the adoption of the hardware reselling strategy, one of the largest buyers of data storage in the world. EMC was “a very large hardware provider, really the leader in storage systems…a top-notch company” . There was advantage to both in securing and increasing each other’s business.

(3) At the same time, Autonomy and EMC were in discussions about building appliances. Also in June 2009, following a difficult negotiation, Autonomy had convinced EMC to create a “ custom configuration for us that will be useful for the Digital Safe ” during this period, and as also Mr Goodfellow accepted in cross-examination.

(4) On 18 June 2009, Mr Sullivan wrote to Mr Scannell of EMC to say that Autonomy would like “ a more strategic relationship with EMC, ” in view of the large investment Autonomy was making in purchases from EMC, in particular in relation to Digital Safe and Introspect. In his evidence, Mr Sullivan explained that what he meant was getting some of EMC’s hardware customers to buy Autonomy software:

“ Q. And what you wanted them to do in this more strategic relationship, you were buying a lot of hardware from them, you wanted them to drive some revenue to you; right?

A. Absolutely.

Q. And what did that mean? Did that mean get some of their hardware customers to buy Autonomy software?

A. Yes. ”

(5) To encourage EMC further, Mr Sullivan had mentioned to Mr Scannell the possibility of an alternative tie-up of some kind between Autonomy and HDS (Hitachi). This was intended to and did excite some rivalry and interest from Mr Scannell. The exchange and its upshot seem to me to have some bearing on the dispute as to the nature of the subsequent arrangements between EMC and Autonomy:

(a) Mr Scannell’s email to Mr Sullivan stated in relevant part:

“This is what I took from our discussion last night. Please clarify if I missed anything or misstated anything.

HDS is looking to provide Autonomy a slick deal where they would provide you:

- a close configuration to what you have with EMC infrastructure already

- they will become a reseller of Autonomy hosting services

- Autonomy are now moving Digital Safe to HDS drives and away from the current vendor

You would like to see EMC try to take a stab at the same type of deal where we would drive Autonomy Hosting Service deals and that would allow us to perhaps do a larger overall deal with you (grow from $2.9m to $5m)???”

(b) Mr Sullivan’s reply stated in relevant part:

“Correct. We would commit to the current deal (subject to some details) plus make an additional commitment for more hardware related to DS and/or Introspect (to be determined). Since we would be making a very large investment in EMC^{^[139]}, we would like a more strategic relationship with EMC so that you could also drive revenue for us. We would also be looking for a commitment this QTR.”

(6) The upshot as reported by Mr Sullivan in an email dated 26 June 2009 to Mr Hussain and Dr Menell was a good deal on the purchase price for EMC hardware which Autonomy was proposing to lease through Bank of America for its own use^{^[140]} , and also a commitment by EMC to refine its hardware for Digital Safe needs, as emails sent at the time demonstrated.^{^[141]}

(7) On 1 July 2009, Mr Sullivan emailed Dr Lynch and Ms Eagan, noting that:

“EMC would like to setup a day where we could get our respective product people together to discuss our respective offerings and see if there are more opportunities for us to work together. For example, they do not have a supervision product or a legal hold product. Not sure what is realistic given that we compete in other areas, but I think it is worth having the conversations.”

It was natural therefore that it was to EMC that, on his return to the US after the Loudham Hall meeting, Mr Sullivan immediately turned to take forward the hardware reselling strategy that had been agreed upon at Loudham Hall. Mr Sullivan elaborated on the relationship established between Autonomy and EMC, which was (as appears later) a point of dispute, as follows:

(1) After the Loudham Hall meeting, Autonomy worked for a relationship with EMC in the nature of a partnership: there was advantage for both in that Autonomy was on its own behalf a large purchaser of hardware from EMC for its data storage business (which Mr Sullivan explained was one of the largest in the world) at the time (before the end of Q3 2009), and EMC was a large and successful provider of hardware, and an OEM willing to embed IDOL into its hardware and thus increase market penetration of IDOL. There was a strong incentive for Autonomy to secure its supply chain for hardware and EMC’s partnership for software by purchasing EMC hardware using its negotiating power to secure material discounts^{^[142]} .

(2) With the EMC hardware sales, Autonomy focused on mutual customers for hardware sales. As Mr Sullivan said, “ we specifically targeted … our mutual customer list to provide this benefit. ” The way he saw the programme was set out in an email sent by Mr Sullivan to EMC on 15 September 2009:

“The purpose of the program is to strengthen our relationship with enterprise customers in the NYC region by offering them EMC information management products at attractive pricing. We will focus primarily on existing Autonomy Enterprise customers in the financial sector, but other companies may be added as we mutually agree.”

(3) Mr Sullivan said that the goal was to get some of EMC’s hardware customers to buy Autonomy software and to encourage EMC to push Autonomy products: “We would have loved to have them, for example, push our Digital Safe project.”

(4) Mr Sullivan also explained that he opened and pursued discussions with EMC in the latter half of 2009 about “building an appliance, so something like potentially shipping a Digital Safe or actually e-Discovery on [sic] bundling it on their hardware and shipping it…”

(5) Discussions about an appliance continued after that and on 5 October 2009 Mr Sullivan wrote to Mr Scannell to say: “ Per our previous conversations, I would like to get together to discuss the continuation of the programme to include development of an appliance bundle, mutual cross referrals, account introductions, reselling and other opportunities. ”

(6) Mr Sullivan explained that this referred to discussions he had been having with EMC about building an appliance to bundle Digital Safe or e-Discovery on EMC software and discussions from November 2009 onwards between technical teams evaluating EMC hardware that might be used in an appliance. These discussions continued through 2010.

Contractual arrangements between Autonomy and EMC

The Claimants could not dispute the fact of those discussions; and indeed Mr Goodfellow had to accept in cross-examination (after at least three obviously evasive answers) that EMC hardware was being refined for use with Digital Safe towards the end of 2009. Instead, the Claimants contended that in the event, EMC never made any enforceable commitment beyond those set out in a “Purchase & Fulfillment Agreement” executed on 10 September 2009, which they submitted was definitive of the relationship between Autonomy and EMC, and involved nothing more than “a straightforward purchase by Autonomy of EMC hardware for resale (at a loss to Autonomy) to EMC’s customers.” The Claimants submitted further that this was illustrative of the hardware reselling arrangements as a whole.

The Claimants highlighted the following substantive provisions of that agreement:

(1) Under clause 2, EMC would discuss with the end Customer its (the end customer’s) needs and EMC and the Customer would agree the specification, which Autonomy could not change or amend. EMC would then issue a price quote to Autonomy based on that specification.

(2) Under clause 3, Autonomy would issue a purchase order to EMC based on the quote, stating the name and address of the Customer, the specific products and/or services to be purchased by the Agent for resale to Customer and “total fees payable to EMC by Fulfillment Agent.” EMC would then arrange for the hardware to be shipped directly by EMC to the customer.

(3) Clause 9 (headed “ relationship of parties ”) made clear that Autonomy was “ not a reseller (except as described herein), referral partner, joint venturer or any other sales agent or representative of EMC ”. Instead, the parties’ relationship was “ that of non-exclusive independent contractors ”.

(4) Clause 13 provided that the agreement constituted the entire agreement between EMC and Autonomy in relation to the hardware resales.^{^[143]}

However, that summary gave no content to the first clause which envisaged that “the Customer” would be someone with whom Autonomy had an existing relationship, and who might be interested in the enhancement of the hardware in question by the provision of Autonomy software and other “enhancements”. Its first clause, headed “Background”, recited as follows:

“A ‘Customer’ is a third party buyer located in the USA (excluding any United States federal government entity, agency or department) which desires to obtain one or more products or services offered by EMC, and with whom Agent may: (i) hold one or more relationships; and/or (ii) through its resources, other relationships, product offerings, or otherwise, be capable of enhancing Customer’s purchase of such EMC products or services. The term ‘products’ may include both hardware and software. This Agreement states the terms and conditions under which EMC agrees to (a) accept purchase orders and collect payment from Agent for purchases made hereunder, and (b) permit Agent to sell EMC products or services to Customers.”

That first clause seemed to me to support the Defendants’ case that Autonomy selected its own customers, or at least, supplied to companies which, even if customers of EMC, were also customers of its as well. On that basis, the arrangements did enable it to further the purpose it had identified by supplying its customers with a packaged solution, whilst strengthening its relationship with the hardware supplier. That case was further supported by the fact that although selected by EMC, most (perhaps all, see below) of the “Customers” (principally banks and other financial institutions, such as BofA, JPMC, Deutsche Bank and Societe Generale), were also large customers of Autonomy ^{^[144]} , with large software and archiving needs and an existing relationship with Autonomy (such as the hosting of their archives in Digital Safe at an Autonomy/Zantaz datacentre).

However, as often transpired, the contractual wording did not in fact confine the relationship as it developed. An example given by the Claimants to explain their perception of the operation in practice of the hardware reselling arrangements with EMC in Q3 2009 suggested that the chosen “Customer” was not always an existing customer of Autonomy with software needs. The example given was as follows:

(1) On 22 September 2009, the specification having previously been agreed between Bloomberg and EMC, Autonomy agreed to sell EMC hardware to Bloomberg for $8.663 million .

(2) On 24 September 2009, Autonomy issued a purchase order to EMC for that hardware in the sum of $10.089 million. The shipping instruction was for hardware to be shipped directly by EMC to Bloomberg’s offices in New York and New Jersey.

(3) The upshot was that Bloomberg received the hardware at a discount of almost 15%, but otherwise its position was identical to that which would have applied had it purchased the hardware directly from EMC.^{^[145]}

(4) The Claimants emphasised that Dr Lynch accepted when cross-examined, that Autonomy was effectively paying money to EMC to allow Autonomy to sell EMC hardware at a fairly substantial loss .

As the Claimants pointed out (albeit in a different context), the documentary evidence suggests that Bloomberg made no software licence purchases during the period Q1 2009 to Q4 2010. It did not purchase software from Autonomy until 2011. Although Dr Lynch suggested in cross-examination that it had made a large software purchase just before the hardware deal, it appears from the documentary evidence that it was not a purchaser of Autonomy software in 2009 or in 2010, and (except obviously for the hardware purchase) it does not appear to have been an Autonomy customer at all in that period.

The “devil in the detail” of the evidence thus suggests that Autonomy was content to take a loss to generate income without either any existing relationship with the customer chosen by EMC, or any reason to suppose that it would service a software need or generate a software sale in the immediate future.

Apparent success of the trading relationship with EMC

The Claimants further supported their case by reference to the immediate availability of as many hardware transactions as Mr Sullivan could handle. At the Loudham Hall meeting, Mr Sullivan had said that it would not be difficult to resell $5 million to $10 million of hardware, and Dr Lynch had joked that if he could do that, he would buy him a Porsche. As it was, within the week (on 15 July 2009) Mr Sullivan reported to Dr Lynch and Mr Hussain by email confirming that he was confident that he could:

“deliver all the revenue we need per our discussions at Loudham. I have verbal commitments for up to $20MM and could probably get twice that if you want it…My biggest concern is that 4 or 5 Porsches are not practical for me. Perhaps we need to discuss a small plane or yacht?”

The Claimants relied on that jocular confidence and the exclusive reference to “ the revenue we need” as confirming their contention that all that was involved was agreeing with an established supplier a means of purchasing from the hardware manufacturer and then reselling on hardware to end-users at a discounted price, the discount in effect being funded by Autonomy. Both manufacturer/supplier and end-user gained, and it cannot have been difficult to interest them, if all that was sought by Autonomy was revenue without any further tie or commitment. The Claimants relied on that, and also on the almost immediate expansion of the programme well beyond what was envisioned at the Loudham Hall meeting, apparently without any re-assessment, as further indications that the programme had nothing substantively to do with business building and everything to do with revenue generation.

EMC decide not to proceed with the programme

Nevertheless, what appeared to be a developing relationship with EMC on hardware reselling rather abruptly came to an end, just after the end of Q3 2009 in December 2009. The reasons for this too were disputed :

(1) The Claimants contended that EMC’s sudden refusal to remain involved with the programme (a) suggested some sense on its part that there was something amiss and, in any event, (b) confirmed that there was nothing beyond the reselling arrangements to bind them, and that Autonomy’s suggestion that EMC had committed to a strategic partnership and the development of a joint appliance was baseless. They cited an exchange of emails between Mr Sullivan and Mr Scannell at the beginning of December 2009, in which, in response to a request from Mr Sullivan asking “can you help us out at all in Q4?”, Mr Scannell had emailed Mr Sullivan on 9 December 2009 that

“At this point we unfortunately are uncomfortable with these deals”.

(2) The Defendants, on the other hand, dismissed any suggestion that EMC had sensed anything untoward. They ascribed the change to EMC having bought one of Autonomy’s competitors, a company called Kazeon. Employees of EMC working in the division that had bought Kazeon were lobbying against the Autonomy hardware reselling deals. Mr Sullivan told the US court:

“I think that the dynamic that I understood within EMC is that the software business within EMC would rather have EMC bringing them into the deals rather than - rather than us.”

(3) The Defendants suggested that this too was telling in that it appeared to reflect a concern that the hardware sales strategy did assist in driving software sales. (They suggested further that the fact that it was possible for a hardware manufacturer suddenly to cease supplying hardware without any prior notice confirmed the importance of having strong relationships with them.)

(4) The Defendants relied also on the fact that not long after, in March 2010, EMC indicated they would be prepared to consider exploring a different business relationship with Autonomy, suggesting that the reason for the break was not based in any suspicion of impropriety.

The evidence is incomplete, and of itself provides no certain explanation for the sudden decision of EMC to withdraw. However, Dr Lynch plainly regarded it as important to his case to demonstrate that there was nothing sinister in EMC’s withdrawal, and that it did not indicate that what Deloitte and the Audit Committee had been told in Q3 2009 - that Autonomy and EMC were in a strategic partnership and were developing an appliance - was baseless. He developed a complex explanation which the Claimants denounced as a demonstrable lie.

The explanation Dr Lynch developed was that people within one of EMC’s divisions, called Documentum, were lobbying against the pure hardware deals. He continued:

“ But then unbeknown to Mr Scannell or Mr Sullivan at this point, a conversation broke out with the senior people at EMC which was about buying a division of EMC which was called Documentum for about $2.2 billion and because of that everything got very sensitive about how it was handled from that point onwards ”.

The Claimants submitted that in fact, it is clear on the documents that the discussions with EMC about Autonomy potentially purchasing Documentum did not commence until much later , in April 2010. They relied on the following:

(1) An internal Autonomy memorandum produced on 28 May 2010, which set out the background to a possible deal (under the name “ Project Dynamo ”) . The memorandum stated as follows: “ Morgan Stanley have made the senior level introductions between the two companies and MRL met with Harry You (Executive Vice President, Office of the Chairman, reporting to Joe Tucci Chairman and CEO) in April 2010 ” . Nowhere does the memorandum suggest, nor is it consistent with, discussions having taken place any earlier, and certainly not as early as August 2009 as Dr Lynch suggested in his evidence.

(2) Other documents further demonstrate the inconsistency. On 5 March 2010, Dr Lynch sought an update from Mr Sullivan about relations with EMC. Mr Sullivan responded:

“As for EMC, the tone on “working with us” is very positive in general. They still are not willing to do reselling deals but have said they are happy to explore other types of partnerships. Billy [Scannell] is always willing to consider creative ideas. Billy says the software group (Documentum / Kazeon etc.) is the only group where we might see resistance…

…

They are very interested in keeping us as a customer and we have discussed working on the appliances with them although it does not seem a great fit vs alternatives.”

(3) The Claimants submitted that the document indicated that (a) Documentum’s anticipated “ resistance ” was to EMC exploring any type of collaboration with Autonomy in March 2010; and (b) no appliance was being developed with EMC by March 2010, nor was there any plan to do so in the future, given that Autonomy did not view EMC as “ a great fit vs alternatives ”. Even by August 2010, the possibility of EMC creating an archiving appliance utilising Autonomy software had not proceeded beyond tentative discussions between the parties: see the “ Discussion Paper re Strategic Relationship ” prepared by Autonomy for EMC in August 2010, which identified the creation of an appliance as no more than a possibility under “ Expansion of Existing Relationship ”. The “ Existing Relationships ” section of the paper made no reference to appliances.

(4) The discussions with EMC about Autonomy purchasing Documentum did not, in the event, lead anywhere. Dr Lynch’s evidence in cross-examination was that “ the deal [with EMC regarding Documentum] was all ready to go by the summer ”. Again, that was not supported by the contemporaneous documents; these suggested that the discussions remained tentative by October and November 2010. Ultimately, as Dr Lynch accepted, no deal took place.

Even if (and I do not need to decide this) the analysis may not reveal a lie, I would accept that it supports the Claimants’ suggestions that:

(1) The relationship with EMC never amounted to anything like a partnership: the essence was that EMC was content to sell hardware at a full price which it could offer at a discount subsidised by Autonomy until (as seems to me more likely than not) concerns were expressed higher up the management chain to the effect that it seemed too good to be true;

(2) Any suggestion that the EMC relationship demonstrated or offered the potential of a strong nexus with software sales and development was exaggerated;

(3) The nature of the EMC relationship was not in consequence properly represented to Deloitte.

(4) Autonomy’s relationships with Dell and Hitachi

EMC’s withdrawal necessitated urgent action to find a replacement hardware provider. As I elaborate in paragraph 912 et seq below, there is no doubt that by now, Autonomy had become addicted to hardware reselling to cover any software revenue shortfall headaches. Mr Sullivan found Dell. Dell had already supplied “thousands of servers” for Autonomy’s data centres and hosted offerings which (Mr Sullivan elaborated) included “the world’s largest archiving and discovery offerings, which we were at the time” : The parties’ very different depictions of the relationship between Autonomy and Dell provides one of the clearest illustrations of the clash between, on the one hand, what the Defendants extracted from the evidence of Mr Sullivan and Mr Egan, and, on the other hand, what the Claimants submitted the actual sequence of events and the documentation revealed. In this part of my judgment, I focus on what the Defendants made of the witnesses’ evidence; but it must be borne in mind that a very different story emerges when I turn later to the Claimants’ analysis.

The Defendants relied on Mr Sullivan’s evidence in the US criminal proceedings as providing substantial support for the Defendants’ case that as with EMC, there were sound reasons for extending Autonomy’s co-operation with Dell:

(1) Dell was a natural choice: as well as being large in its own right, Dell was the largest EMC reseller in the world as well.

(2) Further, at the time, Dell was seeking to build its own intelligent storage product, and was evaluating IDOL technology for that purpose, in a project known as Project BlueJay.

(3) Project Bluejay (which Mr Sullivan sometimes referred to as ‘Bluebird’) was Dell’s response to an industry move to build ‘intelligent storage’, which Mr Sullivan explained in his evidence in the US criminal proceedings meant “systems that indexed everything and made it searchable and provided analytics and things like that”. Dell was evaluating technology with a view to developing its own product for the market, including Autonomy’s IDOL.

(4) Discussions on this had begun in February 2009, but moved into “ high-gear ” in November 2009 (about the time that Autonomy started reselling Dell hardware). In this context, Mr Sullivan explained that Autonomy was heavily engaged working with Dell’s R&D team.

(5) In an email to Mr Hussain on 6 November 2009, Mr Sullivan outlined “Potential components of Autonomy/Dell deal” . Among other things, they included Autonomy launching archiving and e-Discovery appliances with Dell, Dell acting as an OEM for Autonomy’s IDOL search platform, Dell reselling Autonomy appliances and software, and Autonomy using Dell equipment and storage for its archiving and discovery offerings. As Mr Sullivan explained,

“…we really wanted to win that because that would mean every time they sold this, we would get a little piece of it. It was an OEM project…That would have been a very large software deal if it closed…”.

Although Project Bluejay did not ultimately result in an OEM deal, the Defendants relied on it as exemplifying the sort of collaboration which Autonomy’s reselling of Dell hardware was intended to support. They relied in this regard also on the following exchanges:

(1) An email dated 19 November 2009 from Mr Craig Irons of Dell to his colleague Mr Bob Barris, introducing him to Mr Sullivan. His email explained that Mr Barris was “ Dell’s Vice President of Sales responsible for the Financial Services vertical in New York ”, and it emphasised the benefits both sides could obtain from collaboration:

“As you and I have discussed Dell Product Group is currently engaged in strategic partnership discussions with Autonomy (Zantaz’ parent company). As those discussion continue I thought it might make sense for you and Mike to know each other as much of Autonomy/Zantaz’s enterprise search, archiving, and ediscovery market leadership resides within the financial services community. As such there may be great synergies available to both you and Mike if you were to begin collaborating within the finserv vertical.”

(2) In an email sent at the time, Mr Egan described some of the benefits of the reselling arrangements with Dell as follows:

“Regarding the Dell/reseller deal:

- This is a purely commercial deal done on a lost [sic., loss] leader basis by Autonomy.

- We gain greatly by seeing faster adoption of our software and we view Morgan Stanley as one of the fastest innovators in the space.

- Dell appreciates the relationship and it greases the process of lucrative OEM bundling of Autonomy software on Dell kit. This is strategic and we pride ourselves at getting our OEMs to move fast and adopt IDOL as a standard. When they have front row seats to the level of usage of our software in a prestigious account like MS they move faster on OEM decisions.”

(3) Emails from March 2010 demonstrated that discussions continued between Dell and Autonomy (run on Autonomy’s side by Mr Sullivan and Mr Matt DeLuca, not Dr Lynch) regarding a strategic partnership between the two companies. Mr Mollo of Dell stated that “ The intent is to get a strategic partnership arranged in several key areas short and long term ” and referred to the intended relationship with Dell as a “ strategic and important partnership .”

Autonomy’s collaboration with Dell continued until HP acquired Autonomy. An email of 17 August 2011 summarising “ Dell recent activities ” dealt with discussions about Autonomy as a “ GTM [go to market] partner ” for Dell with respect to eDiscovery and Cloud, noting that the Dell Software Partner Management Team was “ engaged at highest levels ”, and that this was “key because insistence from Dell HQ we will be promoted by Dell sales will be very helpful in our relationship building in the field.”

While the principal focus was on EMC and later Dell, Mr Sullivan explained in his evidence in the US criminal proceedings that Autonomy also worked on a strategic relationship with Hitachi/HDS:

“Q. And you were working on a similar strategic relationship with Hitachi at that time, weren’t you?

A. Yes. A little different but, yes.”

By way of example, in 2010, Autonomy partnered with HDS on a proof of concept for Morgan Stanley. The proof of concept involved running HDS’s software - Hitachi Content Archive Platform - using IDOL, rather than Microsoft’s FAST. The parties sought to showcase this product to Morgan Stanley, as a replacement for Morgan Stanley’s existing third-party product:

“The purpose is to show Morgan Stanley an archive platform that has the strengths of Hitachi storage coupled to the advanced features of the IDOL index as well as the ability to plug this solution into Morgan’s existing IDOL implementation. The instance of IDOL running on HCAP can be easily “connected” into other IDOL servers, providing an immediate link between the HCAP data and the existing Autonomy solution-set. This is a far cleaner solution than inserting an archive platform that runs a different index application (i.e. MSFT FAST).”

Throughout 2011, Autonomy continued to work with HDS, and its parent, Hitachi Data Systems, on promoting each other’s products to customers. Autonomy secured an OEM agreement with Hitachi Data Systems just prior to the HP acquisition. The deal envisaged Hitachi OEM-ing Autonomy technology into its own enterprise search and storage solutions and reselling the solutions to customers.

In summary, the Defendants depicted Autonomy’s relationships with EMC, Dell and Hitachi as having been intended to enhance Autonomy’s relationships with key suppliers, enable it to provide its own customers with a one-stop and complete solution, and protect it against the threat posed to it, given its focus on software, if appliances did become the preferred means of providing software. They urged that the programme should be regarded as a commercially rational way of protecting and promoting Autonomy’s software business.

(5) The Defendants’ further evidence as to the use and success of the programme

As to what the Defendants claimed was the networking effect of the hardware sales and the potential that offered of leveraging into far more valuable software deals, Mr Hussain’s closing submissions focused especially on Autonomy’s dealings with UBS.

Mr Hussain provided the following chronology in support of this:

(1) After unsuccessful attempts by Autonomy in December 2009 to sell a Legal Hold software product to UBS, Mr Egan approached Mr Dominic Lester (UBS’s head of EMEA Technology) the next quarter with a proposal to resell to UBS $10 million of Dell hardware, fishing for an introduction to UBS senior IT people (with whom Autonomy had no existing relationship).

(2) Although the initial proposal came too late (UBS had already contracted directly with Dell) Mr Lester was sufficiently interested to arrange a meeting between Mr Hussain and Mr Sullivan for Autonomy and senior managers in UBS’s procurement team (Mr Jaffrey and Mr Dille) which took place in February 2010.

(3) In Q2 2010, Autonomy as reseller sold $6.3 million of Dell hardware to UBS via another reseller, Metro (which is one of the ‘pure hardware’ transactions alleged to have been fraudulent). Until then Autonomy had not sold UBS either any hardware or any significant software.

(4) In May 2010, Dr Lynch met with Ms Michele Trogni (the UBS Group CIO) “to discuss [Autonomy’s] vendor relationship with UBS” , following which two telephone conferences were arranged: (a) the first was “to introduce [Autonomy] to Senior Management within our Compliance Group and for [Autonomy] to provide an overview of their compliance solutions and views on industry best practices” and was to be attended by the senior technology officers from UBS’s London and Zurich offices, including the Global Head of Compliance and the Group Chief CIO ; and (b) the second, which took place on 10 September 2010, was for Autonomy to present its offerings for archiving and e-Discovery, and was attended by even more senior UBS personnel, including the Global Heads of IT (Contracting and Shared Services), of Legal and Compliance IT and of Strategic Planning.

(5) In Q1 2011, Autonomy and UBS began negotiations in which Mr Jaffrey of UBS was closely involved for the software deal that Autonomy had (until now unsuccessfully) been trying to interest UBS in since Q4 2009; and on 20 July 2011, UBS entered into a master agreement with Autonomy and made a supply order pursuant to that agreement for a package of software including ACA, DS Mail Supervisor and Legal Hold. The licence fees totalled $13,585,860.

Mr Hussain submitted in his written closing submissions that:

“…the chronology shows that the hardware sales, and proposals for hardware sales, in 2010 were the means by which Autonomy was introduced to and built relationships with the bank’s key technical and procurement officers. These were the individuals who would decide what software UBS would buy, and it was the hardware sales that put Autonomy on their radar. As a result of selling hardware at minimal cost, Autonomy was able to close a high margin US$13 million sale of its flagship software.”

Other evidence put forward by Mr Hussain of hardware sales being used to facilitate software sales or oil the wheels of a desired or difficult relationship included:

(1) An email from Mr Sullivan to Messrs Hussain, Menell and Egan dated 16 September 2009 sent after receiving a forwarded serious complaint from Citi about deficiencies in Autonomy’s provision of support for e-Discovery software supplied to Citi’s subsidiary, Primerica. The complainant had stated in an email to Mr Hussain and others that the deficiencies had caused him to question whether he was “going down the correct path” in proposing to buy software from Autonomy and that he could not recommend that course to “Tony DiSanto CTI Head North America and the guy I have to convince we should be buying Autonomy products”. Mr Sullivan wrote that “Toni DiSanto is also at least in the loop on our EMC deal with Citi. Pete - I will call to share some thoughts and some history…” It was submitted on behalf of Mr Hussain that this showed that (a) earlier hardware sales to Citi had the attention of the bank’s key procurement officer and (b) Mr Sullivan considered, or at least hoped, that this would influence the general relationship for the better, and translate into software deals then under discussion^{^[146]}.

(2) An internal Autonomy email dated 5 August 2009 (cc Mr Hussain) from which it appears that Autonomy sold discounted hardware to SJ Berwin Solicitors to facilitate the migration to Autonomy from Vivismo, a competitor.

(3) Three discounted hardware sales to Citi in March 2010, August 2010 and December 2010: (a) the first to encourage Citi to proceed with a proposal in relation to email management; (b) the second, a discounted sale of storage cells to persuade Citi to persist with Digital Safe after “relationship problems”; and (c) the third, a discounted sale of storage cells to facilitate negotiations for the data migration to Digital Safe from Iron Mountain.

(4) An email dated 29 June 2011 from Mr Egan to Mr Mooney outlined the strategic advantages of Autonomy agreeing to act as a reseller for SGI in return for SGI embedding Autonomy software^{^[147]}.

(5) An email dated 20 December 2010 from Mr Egan to Mr Dan Manners of Deutsche Bank from which it appears that Autonomy offered Deutsche Bank discounted hardware to encourage the bank to sign off on a deal for the IDOL-isation of its Digital Safe.

Mr Hussain contended that these were merely documented examples of the type of dealings between Autonomy and its customers and suppliers. He suggested that “based on Mr Egan’s and Mr Sullivan’s evidence, [they] happened all the time in face-to-face meetings and by telephone…” He supported this with the following:

(1) Mr Sullivan informed HP after the Acquisition that his modus operandi was to negotiate the initial discount with the hardware supplier and that, on at least some occasions, this would be followed by direct negotiations with the customer.

(2) Mr Hussain offered two examples from the evidence to support this. Thus, in March 2010, Mr Sullivan met representatives of Bank of New York Mellon to discuss the possibility of Autonomy supplying it with Dell hardware. On 10 June 2011, Mr Sullivan informed Mr Hussain that he was engaging directly with JPMC’s procurement team to offer them direct incentives in relation to hardware sales.

Such or similar conversations were relied on also by Dr Lynch, whose evidence in his supplemental witness statement was that in seeking to encourage BofA to enter into a software sale before the end of Q4 2010 he had had “several private conversations” with a senior director of BofA, Mr Simon Mackenzie-Smith (then head of BofA in London) when he:

“sought Mr MacKenzie-Smith’s assistance in encouraging the bank’s procurement team to prioritise the deal, [and] emphasised that Autonomy was a big supplier of software and hardware to the bank.”

However, this was not accepted by the Claimants. They urged me to reject this on the footing that it was unsupported by documentation or by any witness evidence from Mr Mackenzie-Smith, had not been referred to in any communication in the period and had not been mentioned by Dr Lynch in his first witness statement. For reasons I have set out later in this Chapter (see paragraphs 1143 to 1155) when dealing specifically with the BofA transaction in question, I have concluded that whatever discussions there were, and even if Dr Lynch did in passing mention the previous discounted hardware sales, the mention of them had no immediate or discernible effect. The evidence of Mr Reagan Smith, who was negotiating the deal for BofA, was to the effect that he had never heard of any such discounted sales, and accordingly they had no influence on the negotiations at all.

Another point made in Dr Lynch’s written closing submissions in seeking to show that Autonomy was discriminating in its hardware sales was that Autonomy was not prepared to take hardware sales where they did not fit with the strategy explained above. There was some factual foundation for this. For example, the evidence appeared to show that Autonomy was not interested in the revenue that could have been obtained from providing services to EMC customers - even pass-through services where the obligation to provide services would be passed on by Autonomy. It refused to provide services despite some pressure from EMC: as Mr Hussain wrote at the time, “ I don’t want us fronting any of the services on behalf of EMC. If it means the deals disappear then I think we have to accept that ”; and he persisted in this position even after Mr Sullivan gave his view that a refusal to provide services “ will be a disaster. All the deals will go away and we will ruin our relationship. ”

Dr Lynch submitted that had the aim of the strategy been to drive revenue, Autonomy would not have taken this position ; and that if it had been pursuing revenue as an end in itself, it would not have confined its hardware sales to its actual or potential software customers: it would have sold to any company that was prepared to buy discounted hardware. The Defendants relied in this regard on an analysis prepared for Deloitte in January 2011 which had shown that 75% of the hardware sales in the period Q3 2009 to Q4 2010 were made to software customers of Autonomy who entered into large software deals in 2009 and 2010.

Dr Lynch accepted that the sales gave Autonomy a small secondary source of revenue; but he was adamant that there is nothing improper in that. As mentioned previously (see paragraphs 633 and 728 above) he accepted that the hardware sales gave Mr Hussain a degree of flexibility that he would not otherwise have in meeting targets and plugging any gaps in revenue: but he presented that as a perfectly legitimate benefit. Autonomy’s 2010 Annual Report made clear that one of the risks which the company faced was “ fluctuations in results due to quarterly reporting, and variability in results due [to] late-in-the-quarter purchasing cycles common in the software industry ” and that to mitigate that risk it engaged in “ Close management of sales pipelines on a quarterly basis to improve visibility in results expectations. ”

Dr Lynch added to this the consideration that not only were hardware sales logistically complex (as explained by Ms Harris, see paragraph 730 above), but also it would have been counterproductive for Autonomy to use loss making hardware sales as a tool in a scheme to inflate its revenues and to portray itself as a “ a high-margin software company whose revenues were growing rapidly ”, as the Claimants allege. The strategy depressed rather than inflated Autonomy’s margins, and other metrics that were important to investors and analysts. In short, whilst headline revenue was increased, profit margins were eroded, as also were critical valuation criteria such as earnings momentum, organic revenue growth and cash generation/flow. The point relied on by Dr Lynch was illustrated by calculations, which were put to Dr Lynch in re-examination. These showed that Autonomy’s organic revenue growth would have been significantly higher if the hardware transactions were stripped out, both for Q2 2011 and H1 2011 (in each case, compared against the equivalent period the previous year).

Thus, the Core Business Organic Growth Rate Calculation in Autonomy’s interim results up to 30 June 2011 gave growth rates of 15% (Q2 2011) and 17% (H1 2011). Adjusting these figures to remove the “pure hardware” sales listed by the Claimants in Schedule 1 of the RRAPoC, those figures are increased to 28% (Q2 2011) and 21% (H1 2011 ). Of course, according to the periods compared, different results going the other way may be obtained: and the Claimants cross-examined Dr Lynch on such different comparisons. But the Defendants contended that the point endured that, in the key period before the acquisition, organic growth increased if hardware was stripped out; and that if hardware sales were a ruse, as the Claimants claimed, artificially to inflate the apparent value of Autonomy, then according to those valuation criteria, it backfired.

All relevant witnesses had to accept (though some were reluctant) that stripping out hardware would have increased profit margin and cash generation/flow, as did HP’s investigating accountants Ernst & Young in 2012. Thus:

(1) Even one of Autonomy’s then harshest critics, Mr Morland, accepted (eventually) when cross-examined that higher growth rates would have given a higher valuation (though he also stressed that the decrease in sales would have exerted downward pressure which he considered to be greater).

(2) Mr Apotheker, somewhat obstinately and ultimately unsuccessfully, cavilled against this conclusion in cross-examination; and in particular, he initially sought to maintain and defend a point that he had made in his witness statement that but for the hardware sales Autonomy would not only have been a smaller company in revenue terms (which is obviously true) but also that it would have had “significantly lower profit margins” (which almost equally obviously is incorrect in circumstances where, as was common ground, Autonomy’s accounts did include all losses and costs of hardware sales, and the issue is whether the separate source should have been disclosed).

(3) Mr Giles made a similar point: “ Without hardware sales - which were loss-making - the Autonomy business may have lower revenues, but it would have generated more profits and more cash .”

(4) Mr Pearson explained that “ because Autonomy’s hardware sales were made at a loss, the sales actually depressed Autonomy’s cash flow, profit and EPS. Given EPS was the key valuation driver, these sales would more likely have the effect of reducing Autonomy’s value, not increasing it. ” In his view, a missed earnings target was more likely to have an adverse impact than a missed revenue target, and loss-making hardware sales would only make it more difficult for Autonomy to hit its earnings target.

(5) Ernst & Young reached a similar conclusion in December 2012, in its paper addressing “ our considerations of the allegations on the original valuation of Autonomy at Q4’11 ”:

“Gross treatment of hardware pass-through

The Autonomy standalone hardware sold in FY 10 and FY11 (prior to acquisition (January - September)) was roughly $105 million and $85 million respectively. Having evaluated ASC 605-45-15 - Principal Agent Considerations Management concluded that the majority of indicators support net accounting treatment.

While this does impact the presentation of the income statement we do not believe these items have an impact on the “run-rate” cash flows of the business. Since Autonomy sold the hardware at a loss, these transaction reduced cash flows. Alternatively, HP may have considered different growth rate assumptions Therefore, it is unclear as the impact on HP’s used DCF valuation model that was used to estimate a fair value of the Autonomy business.”

Pausing there briefly, the mathematical logic that loss-making sales will reduce gross margin seems clear. However, I shall return later to discuss whether that correlation was a disincentive to using hardware sales as a means of enabling Autonomy to generate revenue or an incentive to the Defendants devising ways in which to account for hardware sales so as to attenuate the adverse effect (in particular, as will be seen, by inventively accounting for the costs).

Returning to the Defendants’ presentation, and lastly under this heading, and in relation to the Claimants’ point that the Defendants’ presentation of the rationale of the hardware reselling strategy lacked any documentary support, I should deal with a dispute about a document which was uploaded onto the trial Magnum system only after the conclusion of the evidence. The document comprises an email exchange on 2 September 2009 between Dr Lynch and Messrs Egan, Hussain and Dr Menell, with the subject heading “key customer strategy” .

The Defendants relied on this email exchange to underpin their case as to the challenges which Dr Lynch explained the hardware reselling strategy was meant to address, and to contradict the Claimants’ point that the explanation had no documentary support. But what Mr Miles described in his oral closing submissions as “an important document in the case” the Claimants sought to dismiss as “pretextual” and in any event, in its singularity or egregiousness, simply demonstrative of the dearth of documentary evidence supporting the Defendants’ case.

The exchange is thus of some relevance (a) for what it says, (b) as to whether it was indeed contrived or pretextual, but also (c) as another litmus test of the honesty of the Defendants and the reliability of the record they have put forward.

The first in the thread was an email from Dr Lynch to Mr Egan, Mr Hussain and Dr Menell which the Claimants depicted as pretextual:

“As we have achieved scale in a set of major accounts such as Bloomberg, Citi, jpmc etc and are dealing at CIO level let me make it clear I am happy to supply these firms with the full turnkey package in order to maintain key supplier status....sw hw services etc so we cement our vital key supplier status. This is so strategic and valuable in the long term that even if a sub component of this strategy is at a loss that is fine...overall keeping the strategic nature of the relationship will pay dividends.”

Mr Egan responded as follow:

“Agreed, as evidenced by the continued orders from JPMC etc for extended areas of software that in part come from our status as large scale and strategic supplier.”

The Defendants relied on this exchange as encapsulating and demonstrating the fundamental aspects of the rationale for the programme, that is (a) consolidating relationships with hardware suppliers and (b) consolidating and developing relationships with key customers such as Bloomberg, Citi, JPMC and the like, as the means of (c) protecting and promoting Autonomy’s software business.

The Claimants sought to interpret the reference to a “full turnkey package” as a reference to appliances; but Mr Miles in his closing submissions rejected this and insisted that it was a reference to “the full series of suppliers: software, hardware, services, etc.”

Of course, that interpretation was Mr Miles’ gloss; and in that regard and more generally, the Claimants invited me to discount the document in any event on the ground that it had been put forward too late for any examination of witnesses about it. But Mr Miles pointed out that the document was in the Claimants’ own disclosure, and although lately uploaded onto the Magnum trial bundle, it had always been part of the wider ‘corpus’ of disclosed and available documents. He submitted that, on that basis, if they considered the emails in question to be, or wished to explore whether they were, pretextual they should have identified them and then cross-examined the respective authors of each email to give him the opportunity to deal with any doubt as to its meaning, any contention as to its authenticity, and any insinuation or allegation that it had been contrived to ‘paper the record.’ They had not done so, and should not now be heard to make any such contention or allegation.

Eventually, and as foreshadowed above, the Claimants’ substantive response was that:

(1) The simple fact that this is an example of only one or two documents that Dr Lynch has been able to find undermines rather than supports his contention as to the purpose of the hardware program being a marketing one;

(2) Especially having not (as they put it) had the opportunity of cross-examining Dr Lynch, it cannot, in the context of $200m worth of hardware being sold in the Relevant Period, alter the conclusion to be drawn from many contemporaneous documents passing between them that the primary purpose of the programme was revenue generation rather than anything else;

(3) Furthermore, the email was “pretextual”, in that (so they alleged):

i. it was sent “out of the blue” by Dr Lynch to his core management team but notably not Mr Sullivan: there was no need for Dr Lynch to convey his thoughts in this way, unless the email was designed for some ulterior purpose;

ii. it is wholly uncharacteristically formal in content and style and it reads as if it is designed for the record;

iii. its factual basis is contrived: Autonomy was not dealing at CIO level in relation to these hardware sales, which were instead co-ordinated and brought to Autonomy by EMC. Nor is there any documentary support for any dealings taking place between Autonomy and the customers at CIO level in relation to these pure hardware sales;

iv. Autonomy finance department needed something to persuade Deloitte to allow Autonomy to account for these hardware sales how they wished, and the strong inference is that this email was part of that process and designed to put in place a paper trail.

In response to those submissions of the Claimants that both emails of 2 September 2009 were pretextual, Mr Miles submitted in his oral closing that:

(1) In order to accept that it was a pre-textual email, it would be necessary to “postulate someone you’re trying to fool” : whilst the Claimants had contended that Deloitte was the intended audience, the email in question was not the sort that would ever go to Deloitte, nor is there anything to suggest that Deloitte would ever have asked or looked for emails of this kind; and there was no evidence that it had been shown to Deloitte;

(2) Contrary to the Claimants’ suggestion, there is evidence of dealings at CIO level. Mr Miles cited an email chain from 17 July 2009 to 22 July 2009:

(3) This began with an email from Mr Sullivan to Mr Hussain, Mr Egan and Mr Mooney stating:

“We want to identify our mutual large customers where we can become their EMC reseller. This will be the easiest path. DB, Citi, & JPMC are very large customers. Who are our highest level contacts there? What other companies do we have at CIO level contacts with that we can leverage?”

(4) Mr Egan’s response stated:

“Updated

Morgan Stanley-Not so good as we alreayd [sic] have Hitachi thing going and Morgan expressed their preference ot [sic] us so I would not involve EMC

Eli Lilly-Perfect, My relationship is with Mike Heim, CIO

GSK-Perfect, Ingo Elfering, VP IT Strategy at GlaxoSmithKline

SOC Gen-

BofA- I’m going to get you a name of a guy I haven’t met yet but who knows us and of me well.

Apple-Dounbt they buy EMC

AT&T-Not mine

Bloomberg-Tom Secunda, Co founder and chief of technology

Bayer-William (Bill) Dodero, GC US

MacAffee-Ron Wills, Associate General Council, and Dave Dewault CEO (former EMC) May be very tight with Billy or they may not get along??

1. JPMC- We are close to Larry Feinsmith, Paul McEwen, Fernando Castenheira, and Guy Chiarello

2. Deutsche Bank-Firstly, let Billy know we could team up on the Storage Buy Back RFP at DB coming from Rolf Riemenschnitter and Chares White. I know Charles White very well. Dan Manners is also a big champ and recently promoted…”

All this was consistent with the emails in question.

(5) Although in other contexts, Mr Egan had accepted that he had concocted, and had accused others of, concocting pretextual documents^{^[148]} , this was not such a context, and he accepted when cross-examined in these proceedings that he could not point to any evidence to support the assertion as against Dr Lynch and, furthermore, had not suggested any such thing to the US Grand Jury some time earlier than his witness statement.

(6) In fact, the emails are consistent more generally with the evidence of Mr Egan in these proceedings, to the effect that the hardware sales had benefitted the software selling by the company and that was their purpose: there was no reason to think that he was being pretextual in what he said.

(7) Mr Miles also submitted that, despite only being uploaded to the electronic repository of documents for this trial after Dr Lynch’s evidence, the email was disclosed by the Claimants years ago and was part of the Claimants’ disclosure so that they did have the opportunity to cross-examine him on this document. If the Claimants had wished to characterise the document as pretextual they should have said so and cross-examined Dr Lynch about it. It was not open to them now, having themselves disclosed the document and not tested Dr Lynch, to contend that the document was false and dishonest.

I accept that the email exchange was curiously and uncharacteristically formal. But I doubt that either Dr Lynch or Mr Egan would have thought to leave a confected exchange of emails in case of litigation, kept the exchange away from view and then contrived to pull it out of a hat in court late in the day. I have not been persuaded that the exchange was pretextual. As to its introduction into the trial bundle after cross-examination of Dr Lynch had concluded, both sides share the ‘blame’ (though the word is too strong for an oversight in respect of one document amongst 11 million or so). As a document in their own disclosure, the Claimants must be taken to have accepted its authenticity as a document; and though that does not, in my view, extend to accepting that its content was not pretextual, if they wished to allege that it was, they should have identified it and put it to Dr Lynch if they perceived he would rely on it. However, even though it was in the Claimants’ disclosure, if Dr Lynch wished to rely on it, it was for him and his advisers to see to it that it was included in the bundle in good time.

In such circumstances, I will not exclude the exchange; but I will take into account in weighing its effect that it appears to me somewhat contrived; and having had no opportunity to see whether Dr Lynch would be able to dispel that prima facie impression I cannot attach to it such importance as Dr Lynch at a very late stage came to place on it.

(6) The Defendants’ case that the strategy was not secret nor the revenue disguised

The Defendants also submitted that if there had been an illegitimate scheme to inflate revenue, it would have been carefully concealed: but they were adamant that it was not.

The Defendants contended that, on the contrary, the sales were (a) widely discussed within the company and (b) fully disclosed to and reviewed by Deloitte and the Audit Committee.

Discussion within Autonomy itself

The Defendants submitted that the strategy and the hardware sales were well known within the company, and that there is no evidence that Autonomy concealed or attempted to conceal either.

Mr Sullivan explained that when the hardware reselling strategy was discussed and formally launched at the Loudham Hall meeting in July 2009, there were at least 10 people present; and he said that the sales were well known within the company. Dr Lynch’s evidence in his first witness statement was that “a wide group of people knew about Autonomy’s hardware sales, including employees, customers and suppliers.”

The hardware sales were recorded separately in Autonomy’s ledgers, were open to analysis (including being made available to HP after the acquisition), and would inevitably have been discovered by a future purchaser. In this regard, Ms Harris noted that:

“ The ledger codings clearly separated hardware sales and costs from software. For example, 47000 referred to hardware revenue, 57000 referred to the cost of hardware, and there was also a specific marketing code for hardware costs attributable to sales and marketing. ”

Mr Sullivan said that he had “ never concealed [the fact that Autonomy was selling hardware] and never thought there was a reason to. In fact, I tried to make it clear. ” When asked whether he had been told to tell customers that Autonomy was reselling discounted hardware, he answered:

“Absolutely. In fact, it actually said that in the contracts and it said it in all the POs.^{^[149]} Or most of the POs I should say…”

When asked whether as regards the EMC and Dell transactions he told the customers that the sales were part of a marketing programme, he unequivocally answered “Yes”.

Mr Egan (who also had attended the Loudham Hall meeting) was aware of the hardware reselling strategy, though not its detail. So was Ms Gustafsson/Prentis. As I shall come back to in explaining the Audit Committee’s understanding, Mr Bloomer did not regard the hardware sales as a secret: the topic was regularly and openly discussed during his tenure as Chairman of the Audit Committee, as also Mr Webb confirmed.

Understanding of the extent and purpose of the hardware sales programme, however, was, leaving aside the Defendants, Mr Sullivan and what I have termed the clique or cabal of three (see paragraph 833 below), much more limited. The lack of any minute or other record of the Loudham Hall meeting has made it difficult to determine exactly what was stated there to be the purpose, save as it has years later been presented to have been by Dr Lynch. However, given that at that meeting hardware sales of $5 million to $10 million were envisaged to be something of a stretch, I suspect that when the far greater potential had soon become clearer ^{^[150]} , the true purpose may have changed, and become understood only by the few in control of the strategy, and, in particular, the clique or ‘cabal’.

For example, as will be apparent from my later discussion of Mr Egan’s evidence in the context of the SHI/BofA transaction in Q1 2011 (see paragraphs 1140 to 1155 below) that by then at least he had lost touch with the programme, which was being conducted by Mr Sullivan; and his evidence in his witness statement, which on this point was not challenged or unsettled in cross-examination, was that as a general matter he “did not know which companies had purchased hardware from Autonomy or how much third-party hardware was being sold… ” and usually did not know of hardware sales to his software companies (so that he could not “use the fact of a prior or contemporaneous hardware sale to promote the software sale that I was trying to make.” )

Ms Gustafsson similarly made clear in her witness statement that her understanding was that “Autonomy’s hardware sales were fairly limited and made available only to key customers” (though it may be noted that her understanding was that these sales were “with the genuine commercial purpose of maintaining a customer relationship and generating further software sales” ).

The overall impression I have formed was that though hardware reselling was not a secret, and nor were there any efforts to keep it so, the hardware reselling strategy had a very low profile within Autonomy. It was directed by the small group I have called the ‘cabal’ and Dr Lynch, and implemented almost entirely by Mr Sullivan, without any substantial discussion or coordination with Mr Egan or anyone else involved in software sales. The email exchanges concerning the details and targets of hardware sales appear to be almost entirely confined within that group. It does not appear from the evidence that, beyond that small group, it was a matter which aroused either interest or concern. That in itself may be indicative of the fact that the hardware reselling strategy had very little day to day nexus with the mainstream software business.

Disclosure to and review by Deloitte and the Audit Committee

However, there is no dispute about the fact that the nature and extent of Autonomy’s hardware sales were fully disclosed to Deloitte and scrutinised in detail. Deloitte had detailed knowledge of the sales and their avowed purpose; and the fact and avowed purpose of the hardware sales programme was also explained to the Audit Committee.

I am aware that the successive lead partners, Mr Knights and Mr Mercer, have been severely criticised in a FRC Report in disciplinary proceedings, where it was determined that their approach was insufficiently sceptical, and their ultimate judgments were flawed.^{^[151]} But that does not signify that the audit process, as distinct from the judgements made by lead partners, was deficient. The evidence before me revealed a very careful and open audit process, with full cooperation in it by the finance department and all concerned within Autonomy.

Mr Bloomer statement in his witness statement that he “found Deloitte to be thorough and diligent in their audits” was not challenged; and their reports struck him as “detailed, open and direct”.

Ms Gustafsson stated in her witness statement that she considered that the process was thorough and robust. Subject to the caveat I mention below as to whether Mr Knights and Mr Mercer lost objectivity and the requisite degree of scepticism ^{^[152]} , I would accept that.

I also accept her unchallenged evidence that throughout the audit, Deloitte (who were often on-site at Autonomy) had unfettered access to all Autonomy’s files and to the finance department functions and its sales and technical teams. It is a fair inference that the Defendants would have expected any accounting records relating to the hardware reselling to be carefully scrutinised.^{^[153]}

In cross-examination before me, Mr Welham confirmed the truth of the following evidence he had given in the US criminal trial against Mr Hussain:

“Q. …From 2009, 2010, and 2011, Deloitte knew that Autonomy was reselling EMC and Dell hardware; correct?

A. Correct.

Q. And Deloitte, the Deloitte audit team, knew how much hardware Autonomy was reselling; correct?

A. Yes.

Q. Probably down to the penny? I mean did you know exactly how much hardware was being sold?

A. We would test the majority of the hardware transactions. They would have good - good level of understanding, yes.

Q. And Deloitte knew that Autonomy was reselling the hardware at a loss; correct?

A. Yes, for the most part.

Q. And Deloitte also knew that some of the hardware that Autonomy was reselling was sold without adding any software to it?

A. Correct.

Q. Deloitte knew that Autonomy was reselling hardware through hardware resellers as well; correct?

A. Correct.

Q. And all of this was known to and discussed, not only with the Deloitte audit team, but also the audit committee and the board as well?

A. Correct.”

Mr Welham summarised what Deloitte understood the purpose of the programme to be in cross-examination in the US criminal proceedings as follows ^{^[154]} :

“Q. Deloitte, the Deloitte audit team, understood Autonomy's strategic rationale for selling this hardware at a loss; correct?

A. Yes.

Q. And is it fair to say that in the third quarter of 2009, management explained that Autonomy wanted to meet existing and potential customers' demand for one-stop shopping?

A. Yes.

Q. In other words, to be able to supply hardware along with software?

A. Yes.

Q. To big customers?

A. Yes.

Q. Like the banks?

A. Yes.

Q. And Autonomy also wanted to develop a relationship in Q3/2009 with a particular hardware supplier, EMC; correct?

A. Yes.

...

Q. And in sum, you understood that big picture, Autonomy was selling hardware at a loss to increase its software sales down the road; correct?

A. Correct.

Q. And the audit team, applying your professional scepticism, you considered and you challenged Autonomy on that rationale; right?

A. Yes.

Q. But in the end, you weren't aware of any evidence contradicting management's explanation for the reason for why they were doing these sales; right?

A. Correct.”

Mr Welham agreed with the following comments made by Mr Mercer in an interview with the FRC about the use of hardware sales as a means of driving software sales and in cross-examination, he confirmed that they accorded with how he saw things at the time:

“Yes, I think the most important background or context to make on hardware is that Autonomy had no desire just to sell hardware. This was not a different business. This was not something that they were doing as an end in itself. The only reason they ever sold hardware was to sell more software. That’s massively important to get that understanding for all the questions that you - the two or three questions that we need to concentrate on.

We were - Richard was and I was completely aware that Autonomy did not want to sell hardware as an end in itself. They were conscious that some of their competitors were offering hardware and software and combinations thereof. They were conscious that the marketplace was becoming ever more competitive. They didn’t want to lose out on software sales because their offering was short of this area.”

[My emphasis]

Mr Welham also confirmed the statement in the Deloitte Defence in the FRC Proceedings that Deloitte considered that the motivation for the hardware sales as being to drive software sales appeared commercially plausible, and after consideration was accepted as such by the Deloitte audit team. He gave the following evidence:

“Q. But the audit team's position was that there was no problem as such with Autonomy selling hardware for strategic reasons and that also what is called their plain vanilla hardware sales are also okay?

A. Yes.”

In the absence of evidence from any other of those involved in Deloitte, I take and find that to be Deloitte’s understanding, as well as Mr Welham’s own, of the purpose of the hardware reselling strategy. That is reinforced by what Autonomy’s management told Deloitte in discussions on the connected issue of the allocation of the costs of hardware sales (see paragraph 1262 et seq. below).

In summary, I consider and find that Deloitte understood from the material they had:

(1) The full extent of the hardware sales, the terms on which they were contractually undertaken and the identities of all suppliers and end-customers concerned;

(2) That the strategy included, indeed largely comprised, the sale of hardware without Autonomy software;

(3) That the effectiveness of the strategy in marketing terms was largely a matter of assertion without documented analysis (except the “Linkage Analysis” as to which see below);

(4) The typical structure of hardware sales, including the introduction of Autonomy into existing transactions and the payment by Autonomy of subsidies to end-users, resulting in a loss to Autonomy because the price it paid to the hardware suppliers exceeded the net (of subsidy) price it received from end customers;

(5) That the finance department repeatedly demonstrated their determination that no distinction be drawn between that revenue and revenues from Autonomy’s IDOL Product core business, and that the revenues were combined when described in Autonomy’s published information.

In my judgment, Deloitte did not understand (or perhaps, through Mr Knights, chose not to understand) that the “strategy” was not to drive software sales, but to use hardware reselling as a flexible resource from which to make up for shortfalls in software sales. Deloitte (who, as I understand it, did not see Autonomy’s internal emails) could not know or (for whatever reason) did not focus on what truly lay behind the hardware sales, nor the careful calculation of what hardware sales were necessary in light of the identification of any shortfall with growing precision as each quarter came to its end. If Deloitte were misled, or fed a line which they were prepared to swallow, it was not as to the basic facts: it was as to the real over-arching purpose of the hardware sales in the minds of the Defendants and the ‘cabal’. The key part (which I have underlined) of what Mr Mercer told the FRC in the passage quoted in paragraph 841 above, and the part which he himself said was “massively important” , was the perception that “The only reason they ever sold hardware was to sell more software” . Led by Mr Knights and subsequently Mr Mercer in the Cambridge office, they allowed themselves to use the lens that Mr Hussain and the ‘cabal’ provided to them. If Deloitte were at fault, it was in accepting that justification and thereafter in continuing to accept it despite growing concern as to the lack of documentary support for it, and management’s severe allergic reaction against any revelation of the strategy in any published information or public calls. How they came to accept it, and how Mr Knights came to abandon any scepticism, is at the heart of the hardware case, for the strategy depended on it.

Turning to the Audit Committee, Mr Bloomer (who became Chairman of the Audit Committee in August 2010) made clear that he did not regard the hardware sales as a secret. His evidence in his witness statement was that the existence and disclosure of hardware was regularly and thoroughly considered by Deloitte and discussed at every meeting of the Audit Committee whilst he was its Chairman: it was “always one of the key accounting issues we considered.”

That the Audit Committee was fully kept informed is apparent from Deloitte’s reports to the Audit Committee quarterly, half yearly and yearly on their review of Autonomy’s accounts and notes prepared by Mr Hussain, with Mr Chamberlain, also quarterly (which were always circulated to Dr Lynch in advance). By way of example:

(1) In their Q3 2009 Report, Deloitte, having set out that revenue from hardware sales for the quarter ($36 million) amounted to 19% of the total revenues for the period and that the sales (disclosed to be at a loss of some $9 million) did not include any IDOL software component, outlined the explanation they had themselves accepted: this was that the sales reflected “early targeting of the merging market of appliance solutions”. This was elaborated later in the Report as involving “the provision of appliance related solutions to leading multi-national financial institutions” and it was stated that to that end Autonomy had entered into a “significant hardware and marketing purchase” with EMC to provide (a) hardware for delivery to existing Autonomy customers (b) ongoing “joint sales and marketing support to promote further sales in this emerging market” and to (c) “begin to develop an appliance based hardware and software configuration whereby Autonomy software might be fully integrated into EMC hardware for products aimed at the appliance sector and major institutions” .

(2) In their Q1 2010 Report, again under the heading “Key Risks” , Deloitte noted that included in revenues for the quarter was $12.2 million of hardware sales (mostly sales of Dell hardware to Morgan Stanley, SHI and Fanny Mae, but also sales to “other blue-chip companies in order to become the preferred supplier for all archiving requirements, including both software and hardware”) at an overall loss . In a further explanation which is of more particular relevance to the related issue as to the allocation of costs between COGS and Sales and Marketing expenses (see paragraphs 1412 to 1419 below) Deloitte stated that Autonomy had had to pay a price considerably higher than they would normally have to pay in order “to gain a strategic partnership and become the preferred hardware reseller with EMC, Dell, SHI and HDS.”

(3) In their Q2 2010 Interim Report, Deloitte (again under “Key Risks” ) reported hardware sales of $27.5 million at a loss of approximately $3.8 million. They explained management’s rationale in entering into these loss-making contracts as being “that Autonomy is seeking to develop a long term strategic relationship with the end-users in order to secure future profitable software sales.” Deloitte referred for the first time to a Linkage Analysis (see paragraphs 1477 to 1496 below) “demonstrating strong linkage between the loss making hardware sales and highly profitable software sales” and stated that having reviewed it they had accepted that the loss of $3.8 million should be allocated to Sales and Marketing.

(4) In their Q3 2010 Report, under the heading as ever of “Key Risks”, Deloitte signalled loss-making hardware sales of $26 million (10% of total revenues). Deloitte again reported that they had reviewed and accepted “management’s analysis of the linkage between the loss making strategic sales and subsequent highly profitable software sales…”;

(5) In their FY 2010 Report, (as always under “Key [Audit] Risks” ) Deloitte noted (mostly loss-making) hardware sales of $29 million for Q4 2010, representing some 11% of Group revenues. The rationale for this was stated to be “in order to procure future, profitable software sales.” Deloitte also notified a change in proposed accounting treatment of the associated costs of the sales. They explained that given that Autonomy had purchased and on-sold $110 million of hardware during 2010, “management now considers that the level of sales being made is equivalent to that of a hardware reseller” and proposed that as:

“an equivalent sized reseller would expect to make a modest profit on such sales…it would better reflect the nature and volume of the strategic hardware sales if Autonomy recognised an equivalent margin on these sales, estimated at 5%...[which] would then reflect the true sales and marketing expense incurred in making these sales in order to procure further, profitable software sales.”

(6) Deloitte added to this that the total amount thus proposed by management to be allocated to sales and marketing for the year would be $16.4 million, which would be lower than for 2009 ($37.8 million) because:

“those initial [2009] hardware purchases (mainly from EMC) including an amount paid for additional marketing services and future development costs as part of an effort to build strategic relationships with those suppliers…No equivalent marketing services or future development costs were purchased with the 2010 hardware, which is evidenced by the much smaller net loss made on these sales.”

(7) Deloitte stated its response to these proposals which was that (a) they would accept the allocation to sales and marketing expenses of a sum of $4 million equal to the loss on hardware sales in Q4 2010 but (b) this would be justified as a “judgmental adjustment” not as proposed by Autonomy’s management because (c) management’s proposal did not, in Deloitte’s judgement, “better reflect the nature of these transactions” and was “inconsistent with management’s assessment that the group has just one Operating Segment, being sales of IDOL software.”

(8) In their Q1 2011 Report, Deloitte reported loss making hardware sales of $20.4 million for the quarter, again using the description “strategic” to connote that the sales were “only made at a loss in order to procure further, profitable software sales” . They also noted that “Management has further extended its analysis determining the strong linkage between the loss making hardware sales and subsequent highly profitable software sales” and responded to it only by stating that “given the scale and consistency in allocation with the prior quarters, [we] accept the decision taken by management to allocate the loss of $2.0 million to sales and marketing expense in Q1.”

(9) In their Q2 2011 Report, Deloitte noted loss making hardware sales in the quarter of $20.9 million (some 8% of Group revenues, down from the previous quarter) and recorded the same rationale for those “strategic sales”.

Like Deloitte, and in reality following their lead, Mr Bloomer and the Audit Committee accepted that hardware sales were purely to promote and protect the software business. Satisfied of that, and satisfied in consequence that no separate business requiring segmental accounting was required, the Audit Committee was content.

(7) Defendants’ support of and belief in the purpose asserted and its accounting treatment

Two warning notes at the outset of this section are appropriate.

First, not only did the Claimants reject the honesty of the Defendants in respect of the asserted purpose or rationale of the hardware reselling strategy, but they also relied, as a further manifestation of dishonesty, on what they regarded as the Defendants’ deceptive presentation of the programme to Deloitte, especially to obtain the accounting treatment of its costs which they needed to perpetuate its disguise. I return to deal with those matters in greater detail later. This section addresses the Defendants’ case that they honestly believed in the rationale of the hardware reselling strategy they avowed and were reassured in that by Deloitte’s acceptance of it, and of Autonomy’s accounting treatment of the sales revenue it generated (albeit at a stand-alone loss).

Secondly, especially given the fact that the question of whether Deloitte properly understood the relevant requirements (which the Claimants submitted and Mr Holgate was definite they did not) was one of the matters in issue and determined against Mr Knights and Mr Mercer by the FRC’s Disciplinary Tribunal, I should at the outset clarify that, in my view, what is of relevance in the present context is what Deloitte and the Audit Committee understood at the time the Standards to require, and their approval of the accounts and advice to Autonomy on the basis of their understanding. Whether they were right or wrong is a matter relevant to issues numbered (2) to (4) in the Claimants’ summary of the main issues in the hardware case (as set out in paragraph 638 above), which I address in turn later; but it is not relevant in the present context. Even if Deloitte were wrong (and Mr Holgate is adamant that they were) the fact remains that they did give the approval and advice then; and if the Defendants really did rely on it then, they are entitled to rely on it now, unless of course they knew that Deloitte had given it no proper consideration, or that their conclusion was plainly and obviously wrong and/or based on a material misunderstanding of the underlying facts.

The Defendants’ case is that they honestly believed, in reliance on the approval of both Deloitte and the Audit Committee throughout the Relevant Period, that:

(1) having regard to the purpose of the programme, there was no requirement to distinguish hardware revenue from software revenue either in the narrative to the accounts (what the Claimants called the ‘front-end’) or in the accounts themselves (what the Claimants called the ‘back-end’);

(2) the disclosure provided was in accordance with the requirements set out in Accounting Standards devised to enable the presentation of a true and fair view of a company’s financial position;

(3) the further disclosure they provided, in particular in ‘ Supplemental Information’ in the ‘front-end’ , was fair and approved by Deloitte; and

(4) the decision whether or not to give further voluntary disclosure either in accounting or narrative form was a commercial one taken for good commercial reasons and in good faith, after careful iterative exchanges with Deloitte.

Deloitte considered the issue of disclosure from two angles:

(1) they considered whether or not the Accounting Standards mandatorily required disclosure of revenue from hardware sales separately from other revenues; and

(2) in their Reports to the Audit Committee, they identified for discussion anything they considered merited further disclosure and made recommendations in that regard.

As to (1) in the preceding paragraph (paragraph 853(1)), there is a variety of contemporaneous documents which show that Deloitte considered the relevant Accounting Standards relating to the separate identification and disclosure of sources of revenue at the time:

(1) Their principal focus was on the issue of segmental accounting and the application of IFRS 8. Mr Welham said that the audit team considered IFRS 8 particularly carefully, because it was a new piece of guidance, introduced in 2009. Deloitte approved management’s assessment that Autonomy had just one operating segment. Their analysis specifically considered the hardware sales, in order to assess whether they would meet the criteria to be classed as a separate operating segment under IFRS 8, and concluded that they would not. For the FY 2009 audit, this conclusion was reviewed by Mr Barden, in the context of a “ Consultation on difficult or contentious matters. ” The Financial Reporting Review Panel was also happy with the analysis. The Claimants accepted for the purpose of these proceedings that this determination was correct.

(2) As well as considering whether there was more than one operating segment, Deloitte’s reports to the Audit Committee for FY 2009 and FY 2010 showed that Deloitte considered whether there was a need for entity-wide disclosures (under IFRS 8 §§32-34).^{^[155]} Those provisions apply even where a company has only one operating segment for the purposes of IFRS 8. This point was also considered in consultation with Mr Barden, who agreed with the audit team’s determination that no analysis of revenues by product type was required for the purposes of §32.

(3) Deloitte considered the application of IAS 18 §35 and (as Mr Welham confirmed when cross-examined) determined that Autonomy was required to break out services, as that was mandatory (under IAS 18 §35), but that there was no mandatory requirement to disclose hardware separately in the accounts.

In considering these matters, Deloitte were well aware of the extent and nature of the hardware sales (see above). Deloitte specifically addressed their attention to the extent of the sales: the working paper on operating segments for the FY 2010 audit stated that:

“ The level of these sales has continued to increase on a quarterly basis and now contribute approximately 12% of the group's revenue in FY 2010. On the basis that it now represents a relatively significant proportion of Autonomy's business, we must consider whether it would meet the criteria to be classed as a separate Operating Segment under IFRS 8. ”

In assessing whether or not entity-wide disclosure was required under IFRS 8 §§32-34, Deloitte were fully aware of all the various details identified by Mr Holgate as relevant to such an assessment as being liable to “influence decisions that users make on the basis of financial information” : absolute amounts of hardware sales, the amounts of hardware sales relative to total revenues, the significant effect that hardware sales had on growth percentages, the very different gross profit percentages on hardware and software, as well as the fact that Autonomy described itself as a “pure software company”. Deloitte also knew that Autonomy was rated in the market as a “beat and raise” stock with particular importance being attached to (and its share price being liable to fall in the event of any material decline in) its organic revenue growth and high gross margins.

Mr Welham confirmed in cross-examination that his view (and Deloitte’s) at the time was correctly summarised in the following passage of Deloitte’s Defence in the FRC Proceedings:

“ There was no need for Autonomy to disclose the hardware sales in the financial statements because Autonomy was a single segment business ”;

and further that:

“ there are no mandatory disclosure requirements for the financial statements prescribed by IFRS (or other legislation) to disclose the ‘existence, nature and extent of pure hardware sales’. ”

In relation to FY 2009, the absence of disclosure in respect of hardware sales was a point specifically considered by Ms Lisa Bennett of Deloitte’s professional standards review team. Her note raised the question why there was no mention of the hardware sales in note 2(e) to the financial statements. The response given (by, so it appears, the audit team) was “ Not material. ” Mr Welham confirmed that this assessment, founded it would seem on a numerical consideration, rather than on any broader consideration of what it might tell about Autonomy’s quality of earnings, was a further reason for Deloitte’s advice that there was no need to disclose hardware in the back-end of the accounts.

Whether the Accounting Standards required separate disclosure of hardware revenue in the accounts was also a matter repeatedly considered with the Audit Committee (especially under Mr Bloomer’s chairmanship from mid-2010 onwards):

(1) According to Dr Lynch, careful consideration was given to advice from Mr Knights in Q3 2009 that disclosure of the hardware sales should be discussed since there was a prospect that it could be mandatory at year-end “particularly if they became material to the numbers”. His evidence in cross-examination was that although he had no specific memory of it, he thought that there had been a decision not to disclose at that time because “we didn’t expect the numbers to be material”.

(2) In 2010, shortly after joining the Audit Committee in mid-2010, Mr Bloomer had a long discussion with Mr Mercer, addressing the decision not to separate out hardware as a separate operating segment. Certainly during his tenure as Chairman of the Audit Committee, the Audit Committee discussed the level of hardware sales regularly with Deloitte, in considering whether further disclosure was required. Mr Bloomer stated in his witness statement that the hardware sales and their disclosure were regularly and thoroughly considered by Deloitte, and discussed at every Audit Committee he chaired: indeed, “ It was always one of the key accounting issues we considered. ”

In cross-examination, Mr Bloomer explained that:

“A. … it was a regular topic at the audit committee, not least because we were discussing, certainly at least two of the audit committees, whether there was a need under what was then a relatively recent accounting standard to split out separate segments of the business, and there was clearly a view for a range of reasons that are set out that there was no need to do that for hardware. We also considered it for a geographical split where there was some information given but again that wasn't felt to be an operating segment. It was quite clear in both Deloitte's mind, management's mind, Deloitte's and mine, that we had one operating segment.

Q. Did Deloitte give any indication that you can recall as to the level of sales at which it would then become an issue?

A. No. No. And as I say, the levels of -- the proportion of hardware sales started to fall anyway so it became less relevant.”

Mr Bloomer, whom the Claimants described as “a straightforward and reliable witness”, explained his view that, in circumstances where, as he emphasised his perception to be, “hardware sales were not a strategic goal of the company (as opposed to a tool to further software sales)” and the assessment had been made that hardware was not a separate operating segment, the decision not to disclose hardware sales separately naturally followed. Viewing the matter as an issue relating to segmental accounting and whether the hardware sales undermined the conclusion that Autonomy was a single segment (software) business, he was adamant that:

“The decision not to disclose hardware sales separately was not an attempt to hide anything. It just did not make sense to separate out hardware sales because we did not consider hardware sales a separate part of Autonomy’s business. Had hardware been a higher portion of sales, say 25% or more, we would have considered whether hardware sales had become more important such that it was, in fact, a separate operating segment or, in any event, significant enough to warrant further disclosure.”

What I take to be the Audit Committee’s overall view was summarised by Mr Bloomer in his witness statement, in two passages which were not challenged, as follows:

(1) “I really did not see the disclosure of hardware sales as a major concern at the time. I saw it as a tool Autonomy used to develop key relationships, get big sales over the line and maintain the headline price of its software. The purpose of hardware sales, as I understood it, was to drive Autonomy’s software sales. This did not necessarily mean there was always a direct link between any particular sale of hardware and a sale of software. Sales of hardware and software were not necessarily simultaneous but, as I understood it, the underlying strategy behind hardware sales was always to further software sales.”

(2) “Had hardware been a higher portion of sales, say 25% or more, we would have considered whether hardware sales had become more important such that it was, in fact, a separate operating segment or, in any event, significant enough to warrant further disclosure. ”

In short, neither Deloitte nor the Audit Committee ever considered that further disclosure of the hardware reselling strategy was mandatorily required by the Accounting Standards; and both were content to approve Autonomy’s published information in respect of its financial position in the Relevant Period accordingly. Mr Webb QC, Autonomy’s Chairman, gave the following unchallenged evidence:

“I can remember occasionally discussing hardware sales and accounting for hardware. I was not a member of the Audit Committee, but I attended Audit Committee meetings once or twice and met with Deloitte from time to time. I do not recall the context of the conversation, but I remember someone explaining the sales of hardware using an analogy along the lines of “if you sell long playing records, sometimes you have to sell a few gramophones”, which I thought was a light-hearted dig at my generation. I do not recall the discussions around the disclosure of hardware being particularly heated. There was a discussion and the conclusion was to account for hardware sales in whatever way Deloitte said the company was required to account for it.

If Deloitte had any concerns, they could have contacted me easily. I feel confident they would have done so had any issues arisen. They never did.”

Dr Lynch was equally clear that if Deloitte or the Audit Committee had said that the hardware sales had to be disclosed, they would have been disclosed.

However, as the Defendants recognised and accepted, there is a distinction between these decisions as to whether a separation between revenue streams in the accounts was needed (and in particular that there was no need to account for hardware revenues separately), and whether nevertheless some narrative or other non-IFRS disclosure of the nature and extent of hardware sales was required in the front-end of the accounts, or at least advisable.

The Defendants countered the two limbs of the Claimants’ case in this regard, on the basis that:

(1) Autonomy always carefully considered any recommendations or suggestions made by Deloitte as to whether further ‘narrative’ or ‘front-end’/non-IFRS disclosure should be provided; but Deloitte always accepted that this was a matter of balance and commercial judgement, and thus ultimately a commercial decision for Autonomy’s directors. The Defendants rejected as “wrong” the Claimants’ argument that Autonomy ignored repeatedly advice from Deloitte to provide further narrative or other non-IFRS disclosure.

(2) Autonomy provided Supplemental Metrics from Q3 2009 to Q2 2011 to assist a better understanding of its business. Nothing was kept from Deloitte. When Deloitte considered the form of disclosure to be potentially misleading, there were discussions between Autonomy and Deloitte which resulted in an amended presentation suggested originally by Deloitte themselves, which Deloitte and the Audit Committee approved.

Further as to (1) above, the Defendants accepted that Deloitte did repeatedly press the Audit Committee and Autonomy’s management to consider whether further disclosure was required. The following examples illustrate this pressure:

(1) On 14 October 2009, following receipt of the Strategic Deals Memorandum, Mr Knights wrote to the Defendants noting that there would be an issue at year-end as to whether the hardware sales needed to be disclosed and raised the possibility of disclosure in October 2009, as something that Autonomy might like to consider voluntarily, in case disclosure later became mandatory:

“One additional point to be considered at the year end will be whether under IFRS you could be required to disclose hardware sales- particularly if they became material to the numbers. Whilst this is a year end matter, if disclosure did become necessary and in the absence of any previous indication through the year, it would be the first time that this information would be made available to your investor and analyst community. This might be worthy of some consideration at Q3?”

(2) Deloitte’s report to the Audit Committee for Q3 2009 dated 16 October 2009 noted that hardware sales represented 19% of the total revenues for the quarter and that the Autonomy board:

“ should consider how best to communicate this new opportunity to the shareholders as these revenues are not driven from the organic IDOL technology of the Group ”.

Deloitte added that, if hardware sales in the year were significant, there might be a requirement to disclose them in the year-end financial statements.

(3) Deloitte’s report to the Audit Committee at Q2 2010 advised that:

“ Given the increasing significance of hardware sales to the Group’s revenues, and the resultant impact on the gross and operating margin in the quarter and half year results we would expect appropriate explanation to be given in the Q2 2010 press release ”.

(4) Deloitte’s report to the Audit Committee for Q3 2010 again advised that given the:

“ increasing significance of the hardware sales to the Group’s revenues, and the resultant impact on the gross and operating margin in the quarter and half year results we would expect appropriate explanation to be given in the Q3 2010 press release ”.

(5) Deloitte also suggested that it was likely that in light of questions raised at the Q2 2010 press conference:

“it would be helpful to include narrative regarding the nature of these revenues in the quarterly report.”

(6) Deloitte’s report to the Audit Committee on the 2010 audit stated that “ Given that Autonomy has purchased and on-sold $110 million of hardware during 2010, management now considers that the level of sales being made is equivalent to that of a hardware reseller ”. Deloitte expressed the view that:

“ Given the increasing significance of hardware sales to the Group’s revenues, and the resultant impact on the gross and operating margin in the quarter and full year results, we expect appropriate explanation to be given in the 2010 Annual Report ”.

However, there was always iterative discussion and according to Dr Lynch, management would always seek to find wording that addressed the concern in discussion with Deloitte. The Defendants provided illustrations of what they presented as a typical iterative approach as follows.

The first illustration related to Deloitte’s Report to the Audit Committee in Q2 2010 (see paragraph 867 above) and their statement that given the increasing significance of hardware sales to the group’s revenues, and the resultant impact on gross and operating margins “ we would expect appropriate explanation to be given in the Q2 2010 press release. ” As to this:

(1) Following the Audit Committee meeting, at which management had agreed to give disclosure in respect of the impact of strategic hardware sales in the front end of the accounts (the narrative part), Mr Chamberlain prepared some proposed wording for Deloitte. Mr Chamberlain’s proposal was to say that “ The fall in gross margins in Q2 2010 was in line with our expectations due to Arcpliance sales as discussed last quarter .” Deloitte did not regard that first draft of the wording as adequate.

(2) Further wording, with what Mr Chamberlain described as “MRL’s tweaks” was accordingly suggested. The revised wording was that “ The small variation in gross margins in Q2 2010 was in line with our expectations due to the sales mix as discussed last quarter. ”

(3) Discussions took place between Deloitte and management in which the wording was further revised, as shown in Mr Welham’s email dated 21 July 2010 to Mr Robertson and Mr Lumb (each of Deloitte):

“Please find attached the final draft of the Autonomy press release. There have been a few tweaks but nothing substantial since the previous version you saw. The one significant point is the wording on the appliance/hardware sales. We have spent some time going through this with management and the proposed wording is as follows:

The small variation in gross margins in Q2 2010 was in line with our expectations due to the sales mix including appliances as discussed last quarter.

As an engagement team, we are comfortable with the wording because it makes reference to appliance sales and the resultant impact on gross margin. Chris Brough has also confirmed that he can accept this wording.

Can you confirm by return that you are happy from a PSR perspective.”

(4) Mr Welham confirmed that the discussions with management referred to in this email were with Mr Hussain and Mr Chamberlain. The PSR (Mr Robertson, then the professional standards reviewer) provided the confirmation sought by Mr Welham and on 22 July, Deloitte signed off its independent review of the results, with an unqualified opinion.

(5) The results contained the phrase approved by Deloitte, which had added the words underlined:

“ The small variation in gross margins in Q2 2010 was in line with our expectations due to the sales mix including appliances as discussed last quarter. ”

(6) In cross-examination, Dr Lynch was criticised about the sentence quoted above from the Q2 2010 results. His response was that Deloitte had a detailed knowledge of the nature of the hardware sold by Autonomy; they had pressed for further disclosure, which Autonomy agreed to give; and Deloitte had discussed various forms of wording with management (not with Dr Lynch), and arrived at a form of words with which they were comfortable, as Mr Welham confirmed they were in his oral evidence.

Another illustration suggested by the Defendants related to Deloitte’s report to the Audit Committee for the Q3 2010 review, when they (Deloitte) again stated that given the increasing significance of the hardware sales to the Group’s revenues, “ we would expect appropriate explanation to be given in the Q3 2010 press release. ” In this regard:

(1) A sentence was added by management, as noted in Mr Welham’s email to Mr Brough dated 18 October 2010:

“Hardware — they have added in an explanatory sentence similar to that included in the Q2 release which states that the gross margin has once again being [sic] impacted by sales mix as highlighted in Q2. In Q2 they talked about the strategic appliance sales, so this is a direct reference to this.”

(2) Mr Welham confirmed that, having had a discussion with management, Deloitte was comfortable with the wording in the Q3 press release. Again, Deloitte issued an unqualified review opinion.

The third illustration the Defendants advanced concerned discussions on the wording of the Q4 2010 press release. As to this:

(1) Mr Welham had informed Mr Chamberlain by email on 30 January 2011 (cc Mr Mercer) that it would be “ necessary to have some comment on hardware sales in the year and the impact of these on GP margin .” Interestingly, I think, he added “At present, this section states that the principal reason is IDOL cloud but surely this would result in fewer hardware sales.”

(2) There were then discussions between management and Deloitte.

(3) In the final version of the press release, in a section discussing gross profits and gross margins, the press release approved by Deloitte noted:

“ During the year Autonomy has seen success in addressing the urgent needs of a small number of customers with package solutions, constructed of services, hardware and software, such as Arcpliance. The gross margin in these cases is lower than the normal business. ”

(4) Again, Dr Lynch was cross-examined about this, and it was put to him that the statement was only “ about Arcpliance” and as such was misleading. Having pointed out that the words were actually “such as Arcpliance” (connoting a subset of “hardware and software” ) his answers more broadly exemplified his defence:

“Well, that’s language that was worked out with Deloitte and I don’t think Deloitte would have done anything that was misleading…

…

…we have some disagreement about which hardware is which type, which we haven’t really addressed, so there is a difference between us. But, again, I just come back to the fact that Deloitte were completely alive to this issue. They consider it - it actually is not just the audit team, it’s something that goes up to their technical experts and there’s a conversation and we agree a form of words, and this is it, and you know, I’m relying on my finance department and I’m relying on the audit committee and most of all I’m relying on Deloitte and it looks like a reasonable sentence to me.”

The Defendants instanced the following evidence of ultimate consensus after constructive discussion:

(1) Mr Welham confirmed in cross-examination that Deloitte had discussions with Mr Hussain, Mr Chamberlain and the Audit Committee from time to time about potentially including further narrative concerning hardware sales, but that they never considered Autonomy’s disclosure was false or misleading because it did not contain further detail on this point:

“ Q.… We've seen a couple of references so far to Deloitte from time to time discussing with management potentially putting further explanation into the published information.

A. That's correct, yes.

Q. If we can just turn to that for a moment. Is it fair to summarise it this way, that Deloitte had discussions with management and the audit committee from time to time about potentially including further narrative concerning hardware sales because Deloitte thought it would be preferable to say more?

A. That's a valid statement, yes.

Q. But Deloitte never thought that the quarterly releases or the annual reports misrepresented things or were false or misleading because they didn't contain that further narrative?

A. That's correct, yes.

Q. And when you had those discussions with management, those were discussions with Mr Chamberlain and Mr Hussain?

A. Yes, and with the audit committee as well. ”

(2) As to the Audit Committee’s perception in 2010, Mr Bloomer emphasised again, with I think his focus on the issue of segmental accounting, that:

“…from a Deloitte point of view, and clearly from the previous audit committee’s point of view, hardware was not seen as a big topic actually to flag.”

(3) In re-examination, Mr Bloomer added this:

“As I said at the time, Deloitte were not - although they flagged that they’d like to see some mention of it, it wasn’t a huge point for them. It was more of a point, the essence of it, that looking forward, if these get much bigger, then we’re going to have to talk more about them. In practice, the volume of hardware sales certainly in the first two quarters of 2011, dropped somewhat and were starting to become a smaller percentage of total revenue and that was the way I interpreted it at the time. It was certainly not a big issue that this must be. If it had have been, Deloitte would have insisted more on it and they didn’t and I didn’t.”

Dr Lynch also stressed that further voluntary disclosure was kept under review: he accepted that where hardware sales represented a substantial percentage of total revenue (say, 20%) disclosure was “certainly something one needs to think about” but “it depends on the purpose”; and in circumstances where the Accounting Standards imposed no obligation to provide further disclosure, they contended that it was a matter of commercial judgement.

What he presented as the “judgement call for management” was whether voluntarily to disclose (a) the amount and/or (b) a breakdown of the type of the hardware it sold and the nature of the sales. He regarded that, as any such judgement call, to require:

“balancing the market’s desire to know as much as they can find out about the company’s business with Autonomy’s commercial interest in keeping confidential information away from its competitors and counterparties.”

The Defendants supported the decision not to provide further narrative disclosure as commercially rational and justified in circumstances where (a) after careful investigation by Deloitte and its reviewers and by the Audit Committee it had been determined that hardware sales did not constitute a separate operating segment of Autonomy’s business and accordingly that it did not have to break out hardware revenue separately in its accounts whatever proportion of revenue hardware sales comprised in any financial year and (b) though both (and especially Deloitte) favoured further disclosure, neither Deloitte nor the Audit Committee had ever suggested that it was required under the Accountancy Standards nor that determining against it would be improper.

In a sense, the most surprising aspect of the issue of disclosure is the Defendants’ consistent efforts to avoid or minimise it and Deloitte’s passive acceptance of anodyne wording (see paragraphs 869 to 871) which obviously did more to obfuscate than to clarify and reeked of reluctance to give any real insight into a strategy justified as being of such benefit to Autonomy’s overall business. Mr Knights and in his turn, Mr Mercer, seem to have been quite extraordinarily reluctant to stand back from the rules and the mantra of management judgement to ask themselves and press management really hard as to quite what was the reason for such reticence.

After all (as the FRC put it to Mr Mercer when they interviewed him), if the purpose of the hardware reselling strategy was as avowed and successful, it would be more natural to expect the Defendants to “want to sort of shout this strategy explicitly from the rooftops.” Mr Mercer’s response was that Autonomy’s management would not necessarily “from a competitive point of view, want to shout from the rooftops the way we are being successful at driving our software sales by selling them hardware. So there’s a commercial sensitivity point, I suspect…” . This was worse than lame: the hardware sellers were, on the Defendants’ case, collaborating to that end. The fact remains, however, that both Deloitte and the Audit Committee were fully aware of what the analysts and the market were being told, or rather, not being told; and although they occasionally pressed for more disclosure, both clung to the belief that the hardware reselling was all part of the software business and in any event immaterial, and ultimately shrank from insisting on more than the opaque wording which Autonomy was persuaded upon from time to time to include.

Dr Lynch, in his witness statement and subsequently when cross-examined:

(1) Stressed what he regarded as the “defensive nature” of the strategy, and his concern that if Autonomy were to break down its revenue into hardware and software, “let alone into different types of hardware” , competitors would perceive Autonomy’s concern about its weakness in the appliances sector and about its vulnerability to the risk of being removed from strategic supplier lists or to others copying its strategy from a position of greater strength;

(2) Explained that he was also concerned lest revelation of a special relationship with one hardware supplier could preclude Autonomy from working with another, and also lest if news of special deals were broadcast, everyone would be demanding the same;

(3) Stated his assessment that the “more details we disclosed about our hardware sales, the more we would expose these commercial sensitivities” ;

(4) In cross-examination, added to the above (a) concern about “most favoured nation situations that we have to be very careful about” and (b) his understanding that “EMC wanted to keep the arrangement confidential, as did the customers.”

(5) When it was suggested to him that the truth was that he did not want to disclose the programme because he preferred that the market should be misled about Autonomy’s real source of revenue rather than being told the true position, he was adamant:

“No, there was no intention to mislead the market.”

The Defendants maintained that this was also consistent with a comment made by Dr Lynch at the time: in his email to Mr Ariko dated 23 October 2009, after Mr Ariko voiced his concern and wish to review further hardware disclosure, Dr Lynch noted that it was “ also very commercially sensitive to our partner us and our customer and concerning new product developments. ”

It is, of course, important to strip out the effect of hindsight. But even having done so, Dr Lynch’s attempted justification appears to me paper-thin. Furthermore, Dr Lynch’s reply to Mr Ariko dated 23 October 2009 (also wishing him well for an excellent retirement) sought to mollify him by assuring him that (a) “the market is already aware we sell hardware, something we have done for 5 years or so…” (b) Autonomy would be issuing “a new press release next week on these matters relating to hardware” and (c) that “if this approach were to develop into a larger part of our business on an ongoing basis we would review the current master press release template.”

The falsity of each point is elaborated later: the exchange with Mr Ariko is an important part of the story and provides very considerable support to the Claimants’ case that the hardware reselling strategy was deliberately concealed: see paragraphs 1397 to 1403 below. A summary suffices for the present. Dr Lynch’s email to Mr Ariko was false in every respect. The market did not know of the hardware sales. The press release did no such thing: it talked in general terms of Arcpliance. The reassurance provided was never fulfilled.

Nevertheless, to return to focus on the Defendants’ case, they rejected also the second part of the Claimants’ allegations, which was an attack on the non-IFRS Supplemental Metrics included in Autonomy’s Quarterly Reports from Q3 2009 to Q2 2011, and in the 2010 Annual Report, which (for the first time) broke down Autonomy’s revenues into IDOL Product, IDOL OEM, Services and Deferred Revenue Release, and at a later point IDOL Cloud.

Hardware revenue was included in these categories, but the Claimants argued that they were defined in a way that could not properly include hardware, and thereby both overstated software revenues and at least implicitly represented that there were no material revenues from hardware sales. In other words, the Claimants contended that the further disclosure the Defendants did decide to provide was in truth a further exercise in disguise: the non-IFRS Supplemental Metrics in what was called the “front-end of the Accounts” further disguised the fact of the hardware sales.

The Defendants rejected these criticisms of the non-IFRS Supplemental Metrics and maintained that:

(1) As the Quarterly Reports and 2010 Annual Report made clear, these were metrics “ provided for background information and may include qualitative estimates. ” They were not precisely defined.

(2) Deloitte scrutinised the calculation of the non-IFRS metrics each quarter, and was well aware that Autonomy’s hardware sales were included in the category “IDOL Product”. Each quarter, they prepared a working paper, whose stated aim was “ To agree the metrics used in the quarterly press release to the supporting schedules and to test the validity of these schedules. ”^{^[156]} Mr Welham confirmed that, as part of the audit or review process, Deloitte would have reviewed the underlying contracts for at least some of the transactions included in those schedules - all that were over $1m in value, or that were part of the sampling process. He said that Deloitte knew that certain deals included within IDOL Product were hardware deals: indeed, that is clear from the face of the working paper, and from Deloitte’s tickmarks.^{^[157]}

(3) Deloitte were of course aware of how IDOL Product was described in the Annual Reports, as demonstrated by the ticked off versions, where the number is ticked next to the description in question.^{^[158]} They knew that some of the hardware sales did not include an IDOL software component.^{^[159]} They did not consider that this rendered Autonomy’s Quarterly or Annual Reports misleading, as Mr Welham confirmed:

“Q. Then looking at what you did know, going back for example to IDOL Product, you knew that as part of the total amount that was being stated as IDOL Product, that included the hardware deals that we've looked at?

A. Yes.

…

Q. … So you knew those facts, you didn't think that the way that then Autonomy presented itself to the financial markets through its published information was misleading in any way, did you?

A. We did not, no.”

Thus, the Defendants presented the non-IFRS Supplemental Metrics as intended to assist in the understanding of Autonomy’s core business by providing a breakdown of the main sources of IDOL revenue: they contended that they were not intended to mislead, and certainly did not suggest any such improper purpose and dishonesty as was alleged.

In summary, the Defendants contended that they were entitled to rely on the fact that, knowing all the details of the hardware sales, and with the means and commercial acumen to test the commercial justification which the Defendants advanced in support of them, Deloitte and the Audit Committee:

(1) were persuaded, as according to the Defendants was the fact, that Autonomy’s directors and management (and thus the Defendants) had conceived and were implementing the hardware reselling strategy for genuine commercial purposes;

(2) approved Autonomy’s financial statements (sometimes known as the ‘back-end’ of the accounts, for which Deloitte had full auditing responsibilities) as compliant with all applicable Accounting Standards and as showing a true and fair view of its financial position;

(3) did not object to the content of the ‘front-end’ of the accounts (which Deloitte did not audit but had to review for fairness and consistency) and approved the Supplemental Metrics provided by Autonomy;

(4) accepted that what further disclosure to give was a commercial judgment;

(5) never required further disclosure in Autonomy’s published information than was from time to time provided.

The Defendants relied also on the fact that when KPMG spoke to Deloitte during the due diligence process on 17 August 2011, Deloitte confirmed that there were no disagreements with management regarding accounting policies and accounting conclusions.^{^[160]} They could not have given that confirmation if they considered that further disclosure was required which Autonomy was refusing to make.

Summary of the Defendants’ Hardware case

In conclusion on this section, therefore, the Defendants:

(1) Insisted that the purpose of the hardware reselling strategy was clearly elaborated, objectively plausible, intended in good faith to protect and promote the software business, and accepted as such by the Claimants’ own principal witnesses on the hardware part of the case, Messrs Sullivan and Egan.

(2) Submitted that Autonomy would not have engaged in the programme simply for the revenue benefit, since (to quote from Dr Lynch’s supplemental witness statement), “hardware created a negative impact on metrics that were of greater importance to me and the market” (and especially profitability and earnings per share).

(3) Maintained that the programme was successful in that regard, and that its costs (which Mr Mercer estimated^{^[161]} as in total less than 10% of the sales and marketing budget) were minimal compared to the value of the software business secured.

(4) Relied on the approval of Deloitte who had full information and knowledge of the nature, extent and costs of the hardware reselling strategy, in relation to both the narrative description and the accounting treatment of the hardware sales, including the costs associated with them.

(5) Made (as they claimed Deloitte had advised they were entitled to make) a commercial judgement on the extent of any other disclosure, having established that no more detailed disclosure was required, and determined in that context that more disclosure could damage the company.

(6) Accepted that a collateral benefit of the programme was increased revenue which could be recognised in Autonomy’s accounts, but pointed out that Autonomy would not have chosen the programme had revenue pumping been its purpose, given the availability of other means of achieving that purpose (including sales of third-party software) which would not have had the adverse effect of the programme in adversely reducing gross margin and gross profits, and which would have been considerably more flexible.

(7) In all the circumstances rejected the Claimants’ case theory that the sales were intended to bring in extra revenue without disclosing its source.

(C) Elaboration of Claimants’ case on the real purpose of the hardware reselling strategy

The Claimants’ overall response to the Defendants’ case was that even if the commercial rationale of the hardware reselling strategy, as articulated by Dr Lynch and echoed by Messrs Egan and Sullivan, was theoretically reasonable and sustainable, that rationale:

(1) was not in reality what informed, or at least quickly came to inform, the programme as in fact it came to be implemented; and

(2) could not explain, and indeed raised the question, why the Defendants actively sought to disguise and hide from the market what was presented as a commercially beneficial programme to promote and protect Autonomy’s core business.

To recapitulate the principal elements or limbs of the Claimants’ case:

(1) What the hardware reselling strategy was directed towards and entailed was the sale of third-party hardware (at a loss) in order to generate revenue which would be presented and accounted for as derived from Autonomy’s software business: its purpose was to cover and disguise shortfalls in software sales;

(2) The intended benefits asserted by the Defendants were never more than side-benefits used as a pretext to persuade Deloitte and the Audit Committee to include revenue from hardware sales as part of the revenues of a single segment software business without disclosing their true source;

(3) The Defendants were well aware that if the true source, nature and extent of the hardware reselling strategy was disclosed or discovered, the market (and in due course HP) would have considered it to be improper and an erosion of the quality of Autonomy’s earnings: and that this would have adversely impacted Autonomy’s share price and the value placed on IDOL and Autonomy’s software business as a whole;

(4) The extensive steps and representations taken and made by the Defendants, and further steps and representations they were prepared to take and make, to ensure that the hardware reselling strategy was neither disclosed nor discovered, confirm its impropriety and their knowledge that Autonomy’s published information was false in this regard.

The Claimants supported that case by reference to:

(1) The lack of any documentary support for Dr Lynch’s explanation that the hardware reselling strategy was conceived as a response to changes in the market which posed real threats to Autonomy’s core business;

(2) The almost immediate increase in Autonomy’s appetite for hardware reselling transactions between Autonomy and EMC well beyond anything contemplated at the Loudham Hall meeting;

(3) The resort to hardware sales purely as a response (which became routine) to shortfalls in software revenue which were identified at the end of a quarter, and the use of, and accounting for, revenue from such sales as if it were derived from software sales;

(4) The development of ways of using that source as a discretionary fund without any discernible appraisal whether such expansion and use would proportionately benefit Autonomy’s core business, and if so, how;

(5) The Defendants’ internal written communications between themselves and others in their team^{^[162]} , which focused, not on how hardware sales might generate additional software sales, but on how much revenue would be needed and raised from that source to ‘plug the gap’ as regards Autonomy’s aggregate revenue targets.

(6) EMC’s sudden withdrawal as a supplier to Autonomy of hardware for reselling, and Dr Lynch’s misrepresentation of the nature of Autonomy’s relationship with EMC and the reasons for its severance;

(7) EMC’s immediate replacement by Dell, and Dell’s participation in the programme on the terms of a hardware reselling agreement between Autonomy and Dell which in name and substance provided for no more than Autonomy being interposed as a reseller of Dell products to Dell’s own customers (albeit in the expectation that most if not all would also be customers of Autonomy) at a discount funded by Autonomy.

(8) A chronological summary, based on documentary evidence, of the dealings between Autonomy, Dell and Hitachi and their customers between Q1 2010 and Q2 2011, demonstrating in particular that:

i. It became routine for Autonomy to rely on loss-making hardware reselling transactions as the means of achieving revenue forecasts calibrated to meet, and if possible “beat and raise” , market expectations.

ii. In almost every such transaction Autonomy was simply interposed into a hardware sale originally intended to be between Dell and the end-user, with the end-user seldom (if ever) being made aware of the fact that Autonomy was funding the discounted price.

iii. Recognition of revenue from hardware reselling was manipulated so as to minimise the adverse effect on Autonomy’s ‘bottom line’ and thus reduce the visibility and the chances of the programme being revealed or discovered.

iv. Autonomy developed and soon came to depend upon the practice of transacting sufficient loss-making hardware reselling to cover any anticipated shortfall in high margin software sales, but then deferring recognition of revenue arising from such hardware reselling transactions if at the end of the relevant quarter software sales (which, being high margin, maintained its gross margins and EPS) in the event proved sufficient to meet forecast.

(9) The way Mr Sullivan was incentivised solely by reference to hardware revenue generated and recognised, without any reference or regard to software sales resulting from discounted hardware sales nor to any other criteria relevant to the software business.

(10) The absence of contemporaneous documentary evidence to suggest that:

i. Autonomy’s marketing department was in any way cognisant or involved in the hardware reselling strategy;

ii. Autonomy’s software sales team was instructed to use the loss-making hardware sales (with discounts funded by Autonomy) as a bargaining chip when negotiating sales of software to Autonomy’s hardware customers or in any other way as part of any software sales pitch to customers of Autonomy (whether prospective or existing);

iii. apart from the Linkage Analysis (see further in paragraphs 1477 to 1496 below and paragraphs 1034 to 1307 below), there was ever any internal discussion or assessment of any measurable impact or effect of the hardware reselling strategy as a marketing strategy.

(11) The evidence that the Defendants, in conjunction with Mr Kanter and Mr Chamberlain, repeatedly sought to and often did mislead Deloitte and the Audit Committee, or circumvent their concerns and recommendations, and in particular, the steps they took:

i. to persuade Deloitte and the Audit Committee (including by pressing Mr Sullivan to extract a false depiction from EMC of the nature of its relationship with Autonomy) to approve the allocation of costs and losses referable to the hardware reselling strategy as marketing expenses rather than COGS in order to reduce the adverse effect on Autonomy’s apparent financial results and prevent revelation of the strategy;

ii. to present to Deloitte and the Audit Committee information designed to show a “strong linkage” which did not exist between hardware sales and software sales especially to answer Deloitte’s concerns in 2010 that the hardware reselling strategy had become “business as usual” ;

iii. to convince Deloitte and the Audit Committee, when the increase of marketing expenses itself began to invite scrutiny, that R&D and increased sales and marketing expenses were referable to the development and marketing of a new product, SPE.

(12) Autonomy’s management’s determination not to make the full disclosure which the Claimants submitted it is obvious they would have actively wished to make had the purpose of the hardware reselling strategy been as the Defendants asserted it to be.

INTENTIONALLY LEFT BLANK.

These points are elaborated and assessed below.

(1) No documentary evidence of threats to Autonomy’s software business

The Claimants put to Dr Lynch that his oral evidence that Autonomy adopted its hardware strategy in response to two market shifts which seriously threatened Autonomy’s core software business, being (as he summarised them in his supplemental witness statement) (a) vendor consolidation and (b) a perceived shift towards consuming software as an appliance, was not supported by any documentary evidence prior to October 2009.

When further pressed that there was “not a single document…and you know it, that supports your suggestion that the sale of hardware at a loss was in response to the two trends in the market you identify” Dr Lynch’s response was that he disagreed with that; and he told me that he thought there were (a) “multiple emails” (b) analysts reports (c) “commentary inside the company” and (d) information easily available on the internet (especially as to the perceived threat from Google’s Search appliance), all of which should be “very easy to find.” Mr Rabinowitz challenged the Defendants to identify them. None was, either during cross-examination or in re-examination.

Nor was there any record of any meeting or discussions within Autonomy about these market shifts or how to address and respond to them. Dr Lynch’s explanation for this was that it was not the practice within Autonomy to record discussions in writing. For example, when it was put to him that there was no record of any discussion of any of this either before, at or after the Loudham Hall meeting ^{^[163]}, he accepted that this was “ unlikely. We didn’t take notes at the meetings.”

The question is whether this lack of any documentary support indicates, as the Claimants contended, that the alleged risks were simply manufactured by Dr Lynch as an overall justification for the hardware reselling strategy (additional to the promotion of software sales) and had no real basis. That possibility is to some extent supported by the fact that (as I elaborate later when addressing the development of the Strategic Deals Memorandum) Mr Hussain did not make any reference to those risks when first seeking to explain and justify the relationship with EMC.

Nevertheless, I doubt that Dr Lynch would have cited repeatedly an entirely fanciful justification. A characteristic of the justifications for the various stratagems deployed by the Defendants was that they appeared outwardly or apparently reasonable and had enough substance to persuade Deloitte and the Audit Committee, and key employees such as Mr Sullivan and Mr Egan. It seems to me to be plausible and even likely that financial organisations would have wished to reduce the number of IT suppliers and, to the extent possible, to have all their IT requirements met from a single supplier; and likewise it seems plausible that, before development of the Cloud, appliances might have been seen as a necessary part of a software company’s offering as the means of making available its software products.

To my mind, however, the real question is whether it was the alleged risks which really prompted the programme; or whether the risks were relied on as convenient context for what in reality was a response to concerns about shortfalls in software sales in Q2 2009, and the need to find other sources of revenue to maintain Autonomy’s record as a “beat and raise” stock ^{^[164]} and shore up its share price.

(2) Expansion of and dependence on the programme with EMC in Q3 2009

Dr Lynch told me that no fixed amount of hardware transactions to be undertaken had been agreed at the Loudham Hall meeting ^{^[165]} . But Mr Sullivan’s evidence (which I accept) was that the figure he had suggested as easily achievable in a quarter was $5 million to $10 million (which had elicited Dr Lynch’s joke that if he could do this, Dr Lynch would buy him a Porsche). It is reasonably clear, and I find, that this was the range contemplated at Loudham Hall; and it is to be inferred (and I find) that any assessment of the cost/benefit of the hardware reselling strategy when incepted was by reference to that range.

It quickly became apparent from Mr Sullivan’s meetings with EMC after he returned from the Loudham Hall meeting, and as by email dated 15 July 2009 Mr Sullivan reported to Dr Lynch and Mr Hussain, that there would be no difficulty in arranging sufficient hardware transactions with EMC to deliver “all the revenue we need per our discussions at Loudham.” He said he already had commitments for up to $20 million of hardware. He added that he “could probably get twice that if you want” .^{^[166]}

This apparently considerable source of revenue seems to have opened the Defendants’ eyes to the potential of hardware reselling as a more significant resort; and the hardware reselling strategy with EMC soon increased very considerably beyond what had been contemplated when it was incepted. The Claimants cited an exchange of emails in early September 2009 between EMC’s Mr Joe Profeta and Mr Sullivan in this regard:

(1) In an email on 2 September 2009 Mr Profeta set out the deals that EMC was “ chasing for the quarter ” , listing the customers and the deal values as being (a) Bloomberg ($7-10m) (b) Citi ($8-12m) (c) Sony Corporation ($4-7m) (d) JPMC ($10-20m) (e) Deutsche Bank ($8-15m) and (f) DTCC ($4-6m). The customers appear to have been selected by EMC. Most, though probably not all, were existing customers of Autonomy. There is no reference to any identification of their software needs.

(2) Mr Sullivan forwarded this email to Mr Hussain, noting that there was a “ Big range from total $46MM to $80MM ”.

(3) The following day, Mr Hussain forwarded Mr Sullivan’s email to Dr Lynch^{^[167]} stating that he had “ listened in on a call with the head of sales at EMC and Sullivan yesterday ”, that they (EMC) were “ committed internally, their selling style is v aggressive ”, and that they (EMC) believed that “ $60m will be the eventual number (equates to around $73m cost) ”.

(4) That was of course some six times the value of hardware sales which Mr Sullivan recollected had been canvassed at the Loudham Hall meeting. Although Dr Lynch was at pains to emphasise that this was a maximum potentially available, and not a target, the fact that such a considerably increased figure was in contemplation as a recourse seems to me to demonstrate how far the programme had extended from its originally conceived size without any re-assessment of its value as a marketing exercise (if that was what it was).

The Claimants emphasised especially a number of aspects of this exchange:

(1) First, the focus of Mr Hussain was on the amount of revenue that could be obtained by Autonomy through this arrangement, coupled with the cost to Autonomy of participating. Dr Lynch had to accept this when cross-examined.

(2) Second, Mr Hussain did not appear to mention or be concerned with the marketing potential of these sales, nor does he suggest that they were designed to achieve anything other than meeting Autonomy’s revenue targets. Dr Lynch also accepted this, but sought to dismiss it on the basis that (a) discussion about marketing effect would have taken place in a different context and (b) this was just a discussion about “salespeople getting the revenue” and simply showed “the sausage-making machine making sausages”. However, there was no record provided of any discussion about marketing effect in a different context; and the email did not relate to a discussion between “salespeople” : it was a communication between Dr Lynch and Mr Hussain solely focused on revenue generation.

(3) Third, it is plain that responsibility for securing the hardware deals was to lie with EMC alone, utilising their “ v aggressive ” selling style. Consistent with this, it was EMC, rather than Autonomy, that had visibility as to the level of hardware deals that could be secured. The hardware was almost invariably to be physically delivered directly by EMC to the customer, as Dr Lynch accepted^{^[168]} .

(4) Taken together, the Claimants described Autonomy’s role as therefore being “ limited to being slotted into the transaction to provide the customer with the discounted pricing.” In that context, the Claimants rejected a somewhat half-hearted suggestion made by Dr Lynch that EMC and Autonomy co-ordinated in determining the selection of customers. When tested against particular transactions it was clear that the customers were simply selected by EMC and Autonomy was inserted into the transaction thereafter. Dr Lynch had to admit that the selection was not by reference to any marketing potential for Autonomy: it was, he said, “the sausage machine making sausages”.

The Claimants submitted that all this, and the apparent absence of any evidence of any attempt to identify or monitor whether this expansion of the programme with a single supplier to the same pool of customers offered any or any commensurate incremental marketing benefit to Autonomy, or occasioned any review of the programme such as might be expected when a marketing programme potentially expands by a factor of six or seven times its originally conceived size and cost ^{^[169]}, strongly supported their case that “relationship building” and “marketing potential” and the benefits avowed by the Defendants were nothing more than a pretext, and that the real purpose was

“simply to, in effect, ‘buy’ (at a substantial cost) recognisable revenue that would be included in the revenue figures reported to the market without revealing the true source (or cost) of this additional revenue stream.”

The Claimants added that “It is inconceivable that Dr Lynch and Mr Hussain, intelligent individuals, could have considered that what they were doing was honest.” I return to the issue of “guilty knowledge” later.

(3) Emails showing strategy’s purpose and use as flexible source of revenue

There is no doubt, and it was not disputed, that the Defendants were heavily focused on revenue growth as a key performance metric. Autonomy’s share price was volatile, and as a “beat and raise” stock tended to experience sharp declines if in any quarter the Autonomy group failed to meet market expectations of revenue growth. The Claimants did not suggest that such promotion of revenue growth as an indicia or metric of performance was of itself improper; but they did contend that it became improper when it was not disclosed that in reality a material proportion of it was not generated by software sales, and overall revenue generation became an end in itself irrespective of its source.

The Claimants relied especially on a sequence of emails on 10 August 2009 between Dr Lynch and Mr Hussain under the subject heading “revenue updates” as revealing that hardware sales were simply part of a relentless drive for more revenue after disappointing Q2 2009 results, without so much as a whisper of any objective of consolidating and extending relationships with customers and suppliers with a view to enhanced software sales or co-operation towards a joint product and joint marketing. Thus:

(1) In an email at 09:14 Dr Lynch urged Mr Hussain “are you doing sms today…you need to grab it revenue revenue revenue ”.

(2) Mr Hussain responded 30 minutes later that “ EMC is key to quarter ” and that he was “ tracking $190m with EMC ”.

(3) Dr Lynch’s reply was: “NOT ENOUGH” (though in cross-examination Dr Lynch accepted that $190 million was the top end of the analysts’ revenue forecast for the quarter).

(4) Later that day Mr Hussain sent Dr Lynch a further email (under the same subject heading) setting out expected and projected revenues and confirming that he was “still targeting” $20m to $25m of EMC and that including the MS-Hitachi order of “$6m to $7m” he was now at “$196m so far but Iwov^{^[170]} is a bit week [sic]”.

Much the same obsessive focus on revenue generation is apparent from an exchange between them only a fortnight later, on 31 August 2009. This commenced with a query from Mr Hussain whether to spend some time on planning for two acquisition prospects (in fact, Coremetrics and Informatica ^{^[171]} ), and also demonstrates how even Mr Hussain as CFO felt obliged to obtain permission as to how best to use his time. Dr Lynch responded:

“Your call, I guess a day on acq is OK but in general revenue revenue revenue.”

Dr Lynch sought in his witness statement to shrug this off as “a fairly typical push on my part for progress from the salesforce” and as not being “addressed to hardware in particular”. He likewise sought to shrug off his message on 31 August 2009 as “One of my standard motivational emails…”. His explanation in cross-examination, when pressed to accept that he had pushed Mr Hussain especially hard for more revenue following the disappointing Q2 2009 results, was that:

“It was the case every month, so until we had the quarter in, it was always revenue revenue revenue”.

Dr Lynch emphasised that the focus evidenced by these emails was on revenue generally, and was not confined to hardware revenue. But the largest single component mentioned in the emails was the $20 million or more expected from sales of EMC hardware, and Dr Lynch accepted in cross-examination that he appreciated at the time that “EMC was the key to the quarter” in terms of achieving the overall revenue forecast.

The Claimants contended that what was starkly confirmed by these emails was a consistent and obsessive focus, which had become an end in itself, unconnected to any marketing drive or perceived threat, and which was fuelled by the availability of far more revenue from the reselling of hardware than had been contemplated when the programme was incepted.

(4) Use of hardware sales as a flexible source to ‘plug’ shortfalls in software sales

The financially incentivised willingness of EMC to agree to the interposition of Autonomy into existing and prospective hardware sales transactions meant that hardware reselling offered a much more considerable source of revenue than had originally been contemplated at the Loudham Hall meeting.

Total revenue forecasts for Q3 2009 were fixed before the potential of the resource was apparent, and hardware revenues were used ad hoc to make good software revenue shortfalls. In quarters subsequent to Q3 2009, once the resource had become clearer, the potential for hardware sales revenue was also taken into account in the forecasting process. Overall forecasted revenue would take into account and include a targeted sum in respect of revenue from hardware sales: and the target developed for hardware reselling revenue would typically be whatever sum was needed over and above forecast software sales revenue to meet market expectations of likely overall revenue targets.

Thus, the practice within Autonomy became that, early in each quarter, Dr Lynch and Mr Hussain would agree goals for revenue to be achieved from what they called “low margin” or “strategic sales”, meaning hardware reselling. Mr Hussain would then inform Mr Sullivan of these goals, which would also set the framework for Mr Sullivan’s incentive payments (see further below). The hardware sales were a ‘stop-gap’ in the sense that the targeted revenue from them made good the gap between what Autonomy expected software revenues to be and the expectations in the market of total revenue for the quarter which Autonomy’s “beat and raise” status had engendered.

However, ad hoc recourse continued to be made and in the course of every quarter, the target for hardware sales would often be adjusted in response to (a) indications from the hardware supplier (EMC or subsequently Dell, or less often Hitachi) as to their own expected sales and (b) Autonomy’s progress towards achieving (also) pre-determined software sales targets: whilst high margin software sales would be primarily targeted, any shortfall might need to be made good or plugged by recourse to hardware sales. In other words, the hardware reselling strategy came to be used as an available resource not only to pump up quarterly forecasting but also as and when required to “plug gaps” in software revenues.^{^[172]}

This required careful balancing of the various sources of revenue. The need to strike this balance became a feature of Mr Hussain’s exchanges with Dr Lynch in Q3 2009. As the quarter progressed - and as in all subsequent quarters - Mr Hussain closely monitored the level of revenue that had been secured or was anticipated from hardware sales and updated Dr Lynch accordingly.

Efforts to mitigate the effect and visibility of loss-making hardware sales

There was, however, a serious downside of the hardware reselling strategy: an inevitable consequence of its considerable expansion was a considerable increase in its costs (of purchases) and losses (since all, or almost all, the hardware resales were at a loss).

These costs and losses would barely have been significant at the levels of hardware purchases and resales which seem initially to have been contemplated at the Loudham Hall meeting. But as they grew, their effect on Autonomy’s bottom line, and on the prime parameters of gross profit and gross margin (which Mr Morland explained was a calculation of sales minus costs of goods divided by sales, expressed as a percentage), was potentially very considerable.

A material effect and deterioration of gross margin would inevitably cause analyst concern and give rise to questions. A hallmark of software companies is a typically very high gross margin on software sales (80 to 95%) compared to hardware sales by hardware suppliers. This is largely because the direct cost of reproducing software for licence is extremely low. The high gross margins which software companies characteristically enjoy, together with high operating margins which are a key determinant of how much cash a company ultimately generates, are important reasons why software companies tend to command high valuation multiples. Autonomy itself tended to boast very high gross margins of in the region of 90%, and high operating margins of in the region of 40%.

Expanding costs of the hardware reselling strategy might have put that at risk; and any deterioration in gross margin not only meant that the hardware reselling would jeopardise a key value parameter, but it would also increase the risk of scrutiny which would expose its nature and extent. Although, as Dr Lynch acknowledged, Autonomy took steps to confine the use of hardware reselling strategy, with its easy but high-cost revenue, to that which was necessary to fulfil consensus revenue projections, the deterioration in software sales in Q3 2009 (and even more markedly in Q4 2009) made that straightforward solution insufficient. The balance between Autonomy’s need to plug the gap and the risk it exposed was a tricky one.

In subsequent quarters (from Q1 2010 onwards) Mr Hussain and the finance department devised ways of treating revenue from such sales as a discretionary fund to which Autonomy could fine tune its recourse according to the figures for quarterly software revenue: see Q1 2010 (paragraphs 968 to 989), Q2 2010 (paragraphs 1022 to 1029 and 1033(1)) and Q1 2011 (paragraphs 1074 to 1100) below. But in Q3 2009 and Q4 2009, their efforts were directed towards attenuating the adverse effect of hardware sales on its ‘bottom line.’

As to attenuating the adverse effect of loss making sales, as I shall elaborate later, what was required was some way of justifying accounting for the costs/losses not as costs of sales but rather as Sales and Marketing Expenses which would not affect gross margin.

These, together with (so the Claimants alleged) the depiction of the increase in Sales and Marketing expenses as referable to the development and marketing of SPE are matters of principal importance in the context of assessing whether and how Autonomy misled its auditors, Audit Committee and the market, and I shall return to them each in that context.

Developing use of hardware revenues

The Claimants illustrated the emerging use of the hardware reselling strategy as a flexible source of revenue whereby to meet market expectations notwithstanding software sale shortfalls by reference to email exchanges between Mr Hussain and Dr Lynch as Q3 2009 was coming to an end:

(1) On 26 September 2009, Mr Hussain emailed Dr Lynch (under subject heading “update” ) as follows:

“Went thru deals with brent^{^[173]} and mooney this evening.

Yesterday was v bad. Iwov^{^[174]}- $2m off and US idol $6m off. Covered with part jpmc/ emc. But now worried more will fall out. Am at $189m to $190m with $30.7m of EMC stuff. We have $41m so $10.3m left to recognise…”

(2) The Claimants made two points in relation to that email. First, the words “Covered with part jpmc/emc” showed that hardware deals were being used to plug gaps in software sales. In other words, they submitted, the hardware deals were being treated as a source of income that assisted Autonomy in meeting or exceeding its revenue targets where a shortfall had arisen, or was likely, due to lost software sales. Dr Lynch himself said in cross-examination that the EMC hardware sales provided “a back-up plan” if needed. Secondly, and relatedly, it appears from the words “ Am at $189m to $190m with $30.7m of EMC stuff. We have $41m so $10.3m left to recognise ” that Mr Hussain had yet to decide whether to recognise the $10.3 million of hardware revenue within the quarter, or to hold it over until the next quarter; and the approach apparently taken was that recognition would be dictated by the Defendants’ wish to meet the market’s revenue expectations: in other words, the amount of hardware revenue to be recognised would be set in order to achieve that end.

(3) The next day, 27 September 2009, Mr Hussain emailed Dr Lynch with a further update. He said:

“ So from Thursday collapse is heavy. Excluding kraft and including 30.7m from emc derived = $188.5m. Stouff says kraft is possible so with that and $26m from emc we are looking at 190m at this stage ”.

(4) The Claimants submitted that again, this confirmed the approach suggested above: if the Kraft deal concluded, it would only be necessary to recognise $26 million of hardware revenue, rather than $30.7 million, that quarter: the extent of resort to hardware revenue again appears to be treated as (a) optional and (b) unrelated to protecting or driving Autonomy’s software business, or any of Dr Lynch’s stated purposes.

(5) On 28 September 2009, Mr Hussain provided Dr Lynch with a further update, stating “ With 41m from emc related I am at 200m plus but…. ” . That tallied with the $41 million figure for EMC hardware revenue which appeared in Mr Hussain’s email to Dr Lynch of a few days earlier (on 26 September 2009).

There is once again no sign that these fluctuations in the amount of hardware which Mr Hussain asked for were referable to any specific software objective. More generally, there is no record relating to this quarter of any projections or analysis or even discussion as to whether hardware sales could benefit software sales. The ad hoc use of hardware reselling revenue as a discretionary recourse to the extent necessary if and to the extent of a shortfall in software sales was a third feature relied on by the Claimants as demonstrating that the hardware reselling strategy had nothing to do with promoting or protecting software business and everything to do with covering emerging deficiencies.

Dr Lynch did not address any of these emails in his witness statement, though the Claimants pointed out that they had been identified and relied on expressly in the Claimants’ responses to Requests for Information. When cross-examined, Dr Lynch sought to present the figures as estimates or projections, rather than actual sales; but he did have to accept that even if that were so, the fact remained that Mr Hussain was calibrating the amount required from hardware sales according to whether or not software sales eventuated within the quarter and thus according to revenue targets or software revenue shortfalls.

It was in this context that he accepted (as mentioned in paragraph 728 above) that Mr Hussain had “a degree of flexibility” and that “on occasion” hardware revenue would be used to “plug gaps in relation to other revenue which may not be achieved.” Further, when asked whether he told Mr Hussain not to seek to use the arrangement with EMC to raise further revenue at a loss once he was confident of hitting what Dr Lynch recollected as a target of $180 million, he said:

“If I knew we were about to do 196, then that would - if we were making a loss on that hardware, it’s probably one of the first things we’d stop doing.”

Importance of hardware sales revenues relative to total revenues in Q3 2009

In the event, Autonomy’s total hardware purchases from EMC in Q3 2009 amounted to about $47.3 million, and its reselling receipts on resales to JP Morgan, Citi and Bloomberg amounted to about $37.6 million. It thus apparently adventitiously resorted to some $7.6 million additional hardware sales revenue to cover the software sales shortfall. Total revenue reported in Q3 2009 was $191.6 million, of which hardware sales revenue comprised over 19%. That met, indeed just exceeded, best expectations. It was within a range of $190m to $200m which Mr Geall had advised would “help set-off” the fear of bears in the market if Autonomy once more failed to “beat and raise” (see paragraph 65 above).

Without that hardware revenue, Autonomy would not have achieved top-end forecast, nor even the lower end forecast. The Claimants’ case was that the use made of the hardware revenues to meet the need which emerged neatly demonstrated the true purpose.

(5) EMC’s withdrawal from the programme and its replacement by Dell in Q4 2009

Mr Sullivan’s intention had been to continue reselling EMC hardware in Q4 2009. However, as explained in paragraph 784 et seq above, at the beginning of December 2009 EMC suddenly withdrew from the programme. As foreshadowed in describing in those paragraphs the Defendants’ case as to what the witness evidence revealed, the Claimants contended that the sequence of events and documentation belied the Defendants’ presentation of the relationship as primarily or at least substantially driven by the objective of promoting software sales through an OEM relationship.

It is clear that by the time of EMC’s departure, Autonomy had become dependent on hardware reselling to maintain its revenue and meet its forecasts, and needed to replace EMC as the hardware supplier to enable the programme. Mr Hussain’s emails in particular, show some considerable urgency to establish another reselling arrangement.

On 17 November 2009, Mr Hussain sent an email to Mr Mooney and Mr Sullivan with the subject “ Dell ”. He stated that he wanted a call that day as “ this is now critical ”. He continued: “ We need oem ($10m) plus appliance resell (v large). This is a priority ”. The Claimants contended that Mr Hussain’s use of the expression “ appliance ”, given the sums involved (“ v large ”) and the reference to “ resell ”, was code to refer to “pure” hardware sales. I accept that.

On 20 November 2009, Mr Hussain reported to Dr Lynch on a conversation that Mr Sullivan was due to have with Dell. That conversation took place with Bob Barris, a Dell Vice President of Sales, who told Mr Sullivan that Dell had:

“ a few pending deals in the $10m range that could potentially go through Autonomy … The catch might be that these deals are going to happen rather quickly and if there are other contingencies we may miss the window ”.

As the Claimants noted, that would suggest that all Dell had to offer at that stage were deals already in the pipeline with existing or immediately prospective Dell hardware customers, into which Autonomy would simply be injected as reseller: hence Mr Barris’ reference to Dell already having pending deals that could be made to “ go through Autonomy ”. The Claimants made the point that in such circumstances, if Dell’s customer happened also to be an existing Autonomy customer (as in fact Mr Sullivan said in direct examination in the US criminal trial they happened to be) that was just a coincidence. The Claimants relied on this as undermining the suggestion by Dr Lynch about the ‘marketing’ purpose of these sales.

Later that day (20 November 2009), Mr Sullivan reported back to Mr Hussain that there had been “ Great progress with Dell today on the reselling ” and that $10-15 million of potential revenue had been identified, though Autonomy would “ Need to move very fast on paperwork ”.

This development prompted Mr Hussain to start sketching out his thoughts on what Autonomy’s reselling agreement with Dell should say. It is plain that Mr Hussain’s objectives were to be able to point to provisions in the agreement which would reinforce Autonomy’s presentation of it as constituting something more than a resellers’ agreement, and as directed to enhancing software sales.

On 23 November 2009, Mr Hussain set out these thoughts in an email to Mr Sullivan. In that email, he stated “our ideal” would be that the agreement should (a) provide for Autonomy to have the ability to set prices, (b) rehearse that Autonomy would take risk as principal in the transaction with the end customer, (c) provide for the purchase order from Autonomy to Dell to split the price between hardware and marketing, and (d) provide for the purchase order from the customer to Autonomy to contain wording such as “ includes software ”. The Claimants submitted that, in other words, the idea was to use the form of the purchase orders to support the impression that these hardware resales were linked to sales of Autonomy software.

The idea and its purpose were further reflected in internal email exchanges within Autonomy. Thus, for example, in an email from Mr Guiao to Mr Chamberlain and Mr Sullivan (cc Mr Scott) dated 2 December 2009 which attached a travelling draft of the proposed agreement with Dell, Mr Guiao referred to the draft as “based off of the Dell-required VAR agreement” but which would have additions to:

“contemplate both a “standard” VAR relationship, as well as one where pricing is based on Dell marketing efforts…Also I had a discussion with Mike, and mentioned that each PO we submit to Dell will have a line item for Dell marketing…”

The terms of the Value Added Reseller Agreement between Dell and Autonomy

On 22 December 2009, a “Value Added Reseller Agreement” was concluded between Dell and Autonomy. This authorised Autonomy Inc to re-sell Dell branded computer hardware and related products, including software, to Morgan Stanley and SHI International (“SHI”), a company that purchased computer hardware and resold it to BofA. Thereafter, Autonomy used Dell as its primary source of pure hardware for resale.^{^[175]}

The provisions of the agreement between Autonomy and Dell included:

(1) Clause 3.1: Dell appointed Autonomy to resell “ Products ” to “ Approved Accounts ”. The “ Products ” were defined in clause 2 as “ Dell-branded computer hardware and related products, including Software ”^{^[176]} . The “ Approved Accounts ” were set out in “ Attachment A ”, which named Morgan Stanley and SHI.

(2) Clause 4.6 dealt with “ Shipping and Handling Options ” and made clear that the products would be shipped by Dell to Dell’s customer.

(3) Attachment B (which was referenced in clause 3.1) set out further terms regarding the obligations of the parties in relation to the reselling arrangements. Clause 2 made clear that Dell was to deal directly with its customer, following which Dell would issue a quote to Autonomy, which in turn would issue a corresponding quote to the customer “ based upon the agreed upon pricing between ” the Dell customer and Dell.

(4) Clause 7 of Attachment B contained a provision headed “ Joint Marketing Efforts ”. This wording was included at the behest of Mr Hussain and Mr Chamberlain so as to facilitate the allocation of the hardware costs to sales and marketing expenses. However, whether or not Dell participated in any sales and marketing efforts was “ in its sole discretion ”.

The Claimants summarised these arrangements with Dell as in effect providing for Dell marketing its own hardware to its own customers, and then, when terms had been agreed, offering its customers the opportunity of a discount, funded by Autonomy, as a nominated reseller.

On that basis, according to the Claimants, these arrangements were in substance broadly the same as the arrangements with EMC, with cosmetic changes to seek to improve Autonomy’s arguments in relation to the allocation of a proportion of the costs of the hardware reselling strategy as sales and marketing expenses rather than entirely as COGS.

As with EMC, the Defendants contended that Autonomy’s relationship with Dell went far deeper and covered much more ground than was expressed in their agreement. They emphasised that:

(1) The Claimants’ own witness, Mr Sullivan, had incepted the relationship: and he was its instigator and the person chiefly responsible for its development. Mr Sullivan had told the court in the US criminal proceedings that Autonomy not only had used Dell for years, deployed its servers in all its data centres and was buying “thousands of servers at the time” , but also Dell were:

“actually a customer as well. They were an e-Discovery customer, and we were just trying to get them to use more of our e-Discovery products.”

(2) Autonomy’s many projects with Dell included working on development of an appliance with new Dell hardware running IDOL software, under code-name “Project Blue Jay.” Discussions on this had begun in February 2009 and certainly by November 2009 had moved into “high gear”. Although Project Blue Jay eventually did not proceed, Mr Sullivan said “That would have been a very large software deal if it closed, which it didn’t”.

However, the Claimants contended that these were no more than potential side benefits of association with Dell, and were not the purpose of the hardware reselling arrangements or the hardware sales.

Just as Autonomy had swiftly become dependent on EMC to interpose it in hardware sales of EMC hardware to EMC customers (most of which were also Autonomy customers too) so too it became dependent on Dell (albeit with occasional resort to Hitachi) to do likewise in respect of Dell deals.

The inter-company exchanges between Mr Sullivan and Mr Barris show Autonomy consistently seeking to boost revenue from hardware deals, and the disappointment and growing anxiety as quarter end approached on the part of Mr Hussain (in particular) when sales were delayed or failed. None of the exchanges mentioned or even hinted that the sales were intended to achieve a marketing objective for Autonomy. As in Q3 2009, so in this quarter, there is no record of any projections, analysis or even discussion as to whether hardware sales could benefit software. In his evidence in the US criminal proceedings, Mr Sullivan confirmed that none of the sales involved software, although he also confirmed that the end-users were existing customers of Autonomy.

The continued importance of hardware sales in terms of Autonomy’s Q4 2009 revenue is apparent from a number of the contemporaneous documents:

(1) The market consensus for Autonomy’s Q4 2009 revenue was between $223 million and $225 million;

(2) Consistent with that, on 15 December 2009, Mr Hussain sent an email to Dr Lynch, with the subject “ as requested ”, attaching a document entitled “ ROUTE TO 225 ”. The attachment identified that one of the elements to achieving the $225 million upper end of market consensus was obtaining a further $5m of hardware sales through Dell/Hitachi, which “ could be ok with the MS deal ” (which was a reference to a possible deal involving Autonomy buying hardware from Dell for $6.28 million and then selling it to Morgan Stanley for $5.712 million).

(3) A few days later, on 19 December 2009, Mr Hussain provided Dr Lynch with a further update. By this point, to achieve revenues of $224.5 million, further hardware sales of $6.8 million were being targeted. That $6.8 million was comprised of two Dell hardware deals: one with Morgan Stanley for $5.7 million (referred to in the previous sub-paragraph) and a deal for $1.1 million with SHI for the end customer BofA.

(4) Dr Lynch emailed Mr Sullivan on 21 December 2009 (copied to Mr Hussain) stating “ So its [sic] all down to you and 10m from Dell ”. In cross-examination, Dr Lynch agreed that he was making it clear to Mr Sullivan that whether market expectations as to Autonomy’s Q4 2009 revenue could be met depended on getting a hardware deal with Dell.

(5) Mr Hussain emailed Dr Lynch with further “quick updates” on 22 December 2009, which highlighted the possibility of a $20 million Dell hardware deal with Bank of New York, and also noted that EMC had (through Mr Scannell) contacted Mr Sullivan to express interest in this large deal.

(6) Further revenue updates were provided to Dr Lynch by Mr Hussain on 23 December 2009 (“ 225m ”), 24 December 2009 (“ 224 ”), 28 December 2009, 29 December 2009, and then again later that same day.

(7) On the evening of 28 December 2009, Dr Lynch emailed Mr Hussain enquiring whether he had had “ any luck with mooney/sullivan? ” Mr Hussain’s response again illustrated the importance of the hardware revenue: “ Was on call with Sullivan - nothing, though I urged him to keep trying”. Later on 28 December 2009, Mr Hussain emailed Dr Lynch: “ US - talked with mooney (dell oem said no), stouff and Sullivan. They are doing what we agreed though Sullivan has nothing ”. The Claimants pointed out that this suggested close co-operation between Dr Lynch and Mr Hussain.

(8) In the event, Autonomy recognised $9.1 million of pure hardware revenue in Q4 2009, according to the figures in Autonomy Inc’s general ledger. Revenue from impugned VAR and reciprocal transactions was $37.8 million.

(9) The total revenue which the Claimants contended was from an improper activity or source was $46.9 million.

(10) Without that total Autonomy’s revenue for the quarter would have been $158.4 million, a miss of 29% as against the analysts’ consensus of $226.1 million, and that figure included add-back of revenue from previous quarters.

From Q1 2010 onwards, Autonomy continued to purchase and resell Dell hardware and also Hitachi hardware, without there being any software included in the sale, nor any sale of software agreed at the same time.

(6) What a chronological summary of the hardware reselling strategy with Dell by reference to the documentary evidence reveals

In a long section of their written closing submissions, the Claimants set out, separately in respect of Q1 2010 to Q2 2011 (and for good measure, Q3 2011), the facts on which they primarily relied to establish the first limb of their case, to the effect that the hardware reselling strategy was simply a means of pumping revenue, and that the purpose asserted by the Defendants was just a pretext to justify treating the hardware sales revenues as part of the software business without separate identification and disclosure.

They also included within that factual resume the second limb of their case, to the effect that in every quarter both Deloitte and the Audit Committee were materially misled in this regard. In this chapter I focus on the chronology revealed by the documentary evidence as to the first limb. I start with a chronological outline.

Chronology of Autonomy’s relationship with Dell from Q1 2010 to Q2 2011

Q1 2010

Towards the end of Q4 2009 there were, as had become the norm, discussions within Autonomy as to what level of Dell hardware resales should be targeted for Q1 2010:

(1) On 30 December 2009, Mr Mooney sent Dr Lynch an email with the subject “ Dell resale” . Mr Mooney asked, “ Should we target $20M for Q1? ”, to which Dr Lynch replied, “ yep ” .

(2) Mr Mooney passed this instruction on to Mr Sullivan within minutes of receiving it from Dr Lynch: “ We need to target $20M for next Q and get as much in January as we can ”.

(3) Mr Sullivan then emailed Dell’s Mr Barris stating: “ Autonomy’s goal is to do $20m in reselling business in Q1 with Dell. We would like to do as much as possible in January ” .

(4) On 4 January 2010, Mr Barris informed Mr Sullivan of the hardware deals that Dell was “ tracking ”, which he said “ should be able to get us to our desired targets each quarter ” .

In none of these exchanges was there any discussion of which customers might be targeted, or of any marketing benefit that might flow from these deals.

In cross-examination, in response to the suggestion that this was all directed towards generating revenue to fill any gap in software sales, Dr Lynch told me: “ once we were selling the hardware, that had to go into the forecast and once it was in the forecast, it had to be hit ”. He continued:

“So there will be a forecast -- from the point where we decided to do this, we would have our quarterly forecasts and those would be what analysts would try to coalesce around with consensus. Because we know we’re going to be selling hardware, that hardware number has to be built into the forecast because we can’t have it coming on top of it and so the forecast is actually constructed to take account of the expected amount of hardware business.”

The Claimants submitted that Dr Lynch thus accepted, at least implicitly, that the purpose of the hardware sales was to generate the revenue necessary so as to meet Autonomy’s forecasts, which drove in large part market expectations.

The chronology for Q1 2010 in this regard was as follows:

(1) As referred to above, Dr Lynch’s confirmation that Autonomy should target $20 million of hardware sales was given on 30 December 2009.

(2) It was only subsequently, from 7 January 2010, that Mr Hussain and Dr Lynch started discussing the revenue targets for each quarter in 2010.

(3) On 7 January 2010, Mr Hussain provided Dr Lynch with details of the Bloomberg consensus for the year and how that might be achieved.

(4) On 11 January 2010, Dr Lynch responded as follows:

“go with 192

map out the year

work out the acceptable q1 eps

map out the eps year

report back

maintian [sic] y o y margins”

(5) On 15 January 2010, Mr Hussain provided Dr Lynch with a forecast breakdown for the year. He suggested that for Q1 2010 they should target $194 million of revenue, with 45% operating margin and 25 cents earnings per share. He continued: “ I have analysed how to get to $194m and can get the 25 cennts [sic] EPS with around $15m or so of appliance sales ”. In other words, in order to generate $194 million of revenue in Q1 2010, with earnings per share of 25 cents, Mr Hussain estimated that they would need to include what he called “ appliance ” sales of $15 million or so. Dr Lynch confirmed in cross-examination that this was a reference to building hardware sales into the forecast to achieve the target.

(6) Dr Lynch was content with Mr Hussain’s proposals.

Pausing there, the Claimants submitted in their written closing submissions that this chronology demonstrated that it could not have been the 2010 targets that drove Dr Lynch ’s instructions to target $20 million through Dell for Q1 2010, for the “Simple reason that that instruction was given before the setting of the 2010 targets.” But that submission seems to me to be based on a misunderstanding: Dr Lynch ’s point (as I understood it) was that because hardware sales were included in the forecast, they cannot have been there to plug shortfalls in software revenues if they emerged, and that point is substantiated by the chronology, rather than upset by it. How far that takes Dr Lynch is another matter: the Claimants’ contention remains that Autonomy needed to show growth and the forecast Autonomy felt able to give was on one view “padded” with hardware sales to enable it to do so. Whether plug or padding seems to me to make little difference in the context.

To return to the factual r ésum é, throughout Q1 2010, as in previous quarters, Mr Hussain provided regular updates to Dr Lynch in relation to the achieved and projected hardware revenue. The Claimants submitted that every exchange supports their contention that the primary, if not sole, purpose of these sales was the production of recognisable revenue, rather than any ‘marketing ’ objective:

(1) On 8 February 2010, Mr Hussain provided Dr Lynch with an update on projected revenue and earnings per share for Q1 2010. This forecasted revenue at $197 million and earnings per share of 26 cents. Mr Hussain made clear that in order to achieve $197 million, a number of hardware sales would need to be secured, including “ Hit[a]chi / MS + BofA $5m ” and “ Dell (Ubs) $10m (in negotiations) ”. Mr Hussain concluded that he was “ happy that we have sufficient backup to hit the numbers at this stage ”.

(2) On 11 February 2010, Mr Sullivan sent an email to Mr Hussain identifying “ the more probable deals for the qtr ”. As can be seen from the attachment to that email, by that time, Mr Sullivan was anticipating the possibility of Dell hardware deals in Q1 2010 totalling $30.7 million.

(3) In Mr Sullivan ’s commentary on each of these deals there is no suggestion that any of these transactions involved the sale of an appliance or a package that included Autonomy software; nor that the sale of Autonomy software was of the remotest interest to Mr Sullivan, as Dr Lynch accepted in cross-examination.

(4) Further, as the Claimants especially drew to my attention, the entry in relation to a $10 million deal with Citibank indicated that Autonomy ’s involvement in the deal was unlikely to be known by Citibank. Dr Lynch recognised that is hardly consistent with these deals forming part of an Autonomy marketing strategy:

“ I wouldn ’ t be happy with that, unless we could tell Citi. … I wouldn ’ t accept that order if I had known - if I knew that it was not possible to tell Citi, but I don ’ t know how they would stop us telling citi”.

(5) On 23 February 2010, Dr Lynch sent an email to himself which included details of secured and anticipated revenue. By that time, $100 million of revenue had already been achieved in the quarter, of which $5.5 million was “ low margin ”, i.e., loss-making hardware sales. The email appears to note that further hardware deals were projected for the quarter ( “ 3 dell/hiatachi [sic] ”, “ 10 HW ***** ”), with total revenue forecast totalling $200 million.

(6) By 1 March 2010, Dell deals of only approximately $5 million had actually closed. An email from Mr Sullivan to Autonomy ’s Mr Matt de Luca of that date, noted that by that point $50 million of deals had been talked about with Dell but that “ much of the $50m are bids that dell may not win or won ’t hit this qtr ”. The concern appears to have been lest the revenue targets were not hit: and the Claimants made the point that had the purpose of these sales been ‘marketing ’, then whether or not the deals would conclude that quarter would have been a matter of little moment for Mr Sullivan.

(7) On 12 March 2010, Dr Lynch sent himself a further email containing an update in relation to the revenue position as it stood at that time. Revenue from deals “ done ” at that time totalled $108 million, which included $9.3 million of “ low margin ” hardware sales. Further hardware sales were forecast at $8 million ( “ 8 HW ***** (Silicon G, rita„ Dell 3, emc?, Hit[achi]/ms 4.5, dell5 2, Target) ”) in order to arrive at a total revenue figure for the quarter of $195 million.

(8) A few days later, on 15 March 2010, Mr Hussain told Mr Sullivan that they would “ definitely need ” a further Hitachi deal and that he would also “ like ” a further Dell deal, which were two of the hardware deals included in Dr Lynch ’s 12 March update to himself. On the same day, Mr Hussain provided an update to Dr Lynch stating that revenue was “ currently at 195.5m ”, but “ 5-8 more ” was needed.

(9) The following day, 16 March 2010, Mr Hussain sought an update from Mr Sullivan on the status of hardware deals with Dell and Hitachi. Mr Sullivan responded as follows:

“ HDS will be about $2.1m – final orders coming in this week – I need to calc the final number based on final margin agreement with Morgan. Morgan is not willing to place the additional $4m order this qtr.

- should have another $500k to $1m from Dell/SHI. discussing some other orders as well – we could provide extra incentive to accelerate this.

- also working on a deal to sell DB a dedicated safe. this could be about $300k to $400k and could be delivered this qtr. There is no rep involved as Dan Manners came directly to us (Roger, Rob and I). I am in Pleasanton and we are trying to cut a deal with him ASAP.

other longer shots for Q1

- Will get orders from UBS in March for up to $1m, but unlikely to ship. Calls later today on this should provide more info.

- I am meeting Bank of NY in NYC on Thursday re: potential dell orders.

- We will probably get an order from Citi through a reseller called insight in March. Around $1m. Shipping will likely make this a q2 deal. ”

(10) There is no mention of any sales of Autonomy software to those customers. Mr Hussain ’s response was that he “ was counting on hds so need ubs and citi orders to be shipped” . Mr Sullivan immediately actioned that.

(11) Shortly after receiving Mr Sullivan ’s email, Mr Hussain updated Dr Lynch, informing him that it was “ getting more difficult, low margin is down for hds - confirmed for 2.1m out of 4.5m (but with 2m from ubs and citi as possibles). have 1 m from bofa dell so need another c 1.5m from somewhere ”.

In a pattern becoming familiar, as Q1 2010 came to an end, Mr Hussain became increasingly concerned about whether enough hardware reselling revenue would be raised, and the Defendants continued to monitor whether $20 million of hardware revenue would suffice to meet aggregate revenue expectations:

(1) On 24 March 2010, Mr Hussain, in an email to Mr Sullivan, expressed concern that two hardware orders that Mr Sullivan had mentioned might not ship in time to enable revenue to be recognised in the quarter: “ … I need $1m to ship ”.

(2) The next day, Mr Hussain told Mr Sullivan and Mr Egan, in connection with a contemplated hardware sale to Morgan Stanley, “I just want the revenue gents”.

(3) On 27 March 2010, Mr Hussain wrote to Dr Lynch, “Keeping you informed that I have increased the appliance sales to $19.215m (with another $1.7m possible)”.

(4) Dr Lynch disagreed that this was simply a false description for what were in truth pure hardware sales and suggested in cross-examination that some of these sales would have been appliance sales.

(5) However, in his evidence to the US court, Mr Sullivan stated that he did not consider any of the Dell sales in 2010 to have been sales of appliances; and there is no documentary evidence to suggest that any part of the $19.215 million of potential revenue related to appliances. Furthermore, the amount of appliance sales was stated to be very small. I accept Mr Sullivan ’s evidence.

On 29 March 2010, Dr Lynch sent Mr Hussain an email attachment recording in short form the Q1 2010 aggregate revenue figure. This stated that $132 million had already been achieved “ INCL 20 Low margin ” . Dr Lynch ’s email noted “- 5 replace low margin in closed with hi” which the Claimants maintained was a suggestion that $5 million of the loss-making hardware revenue should be replaced with “ hi ”, i.e. revenue from high margin software sales, if such sales could be achieved.

This was put to Dr Lynch in cross-examination. After patient but rigorous cross-examination, he said he was “having trouble following this” and subsequently that this was ten years ago, that it was one of 100 million pages in the case, that he did not know the context, and that the suggested interpretation “might be on the right lines, or it may be something completely different, I don ’t know”. This reluctance to speak to selected documents from 10 years ago is understandable; but it was uncharacteristic of Dr Lynch; and his equivocation was likewise uncharacteristic and unconvincing. There is nothing of substance to contradict the Claimants ’ suggestion, which I thus accept: Dr Lynch ’s equivocation does not suffice .

The following day, 30 March 2010, Dr Lynch sent Mr Hussain, in a format which had become habitual for Dr Lynch of a column of numbers and opposite each number an abbreviated reference to a counterparty (such as “5: PMI Partner” ) or description of type of business or activity (such as “- 5: replace low margin in closed with hi” ), a revision of his suggested list of transactions to achieve forecast. The first entry was “138: Done (INCL 20 Low margin)”. The list following now included two alternatives, the second catering for the possibility that no Vatican Library VAR transaction (discussed in paragraphs 280 to 397 of the Schedule of Impugned VAR Transactions) were to materialise; in that event, Dr Lynch envisaged “ 22 Low margin ”, i.e. $22 million of pure hardware revenue. When cross-examined on this, Dr Lynch accepted, that this anticipated the possibility of Autonomy including anywhere from $15 million of pure hardware revenue ($20 million less $5 million) to $22 million.

Later that same day, 30 March 2010, Dr Lynch sent Mr Hussain a further revision of his earlier notes on sales, again containing two alternatives, stating as follows :

“ Ok heres my list

note the without vat list

good to hear vat looking fine but even if vat were compromised, even if you could get 5 from Valueteam it might just scrape it at 193.”

On 31 March 2010, the last day of the quarter, Mr Hussain sent an email to Dr Lynch entitled “ route ”. The attachment, entitled “ SH route to 25c ”, identified that, if “ Low margin cost ” (i.e. if the cost to Autonomy of purchasing hardware sold in the quarter) were $20 million, total revenue of $200 million would be required in order to achieve earnings per share of 25 cents in Q1 2010. This reflected the fact that although the hardware sales boosted revenue, they were loss-making, and thus had a deleterious effect on earnings per share. As Ms Gustafsson had explained:

“ Selling hardware …, particularly hardware at a loss, involved a cost that hit Autonomy ’ s earnings per share. … While higher hardware sales may have led to a higher quarterly revenue figure, they led to a lower EPS figure given they resulted in a negative margin ”.

Dr Lynch responded to Mr Hussain later that day, attaching a revised version of the “SH route to 25c” document. Based on Dr Lynch’s alternative calculations, earnings per share of 25 cents could be achieved with total revenue of $195 million (“EPS 25c 195”). Dr Lynch’s apparent intention was to suggest that the target earnings per share could be achieved if less hardware revenue (and thus less cost) was recognised in the quarter. He appears to have reached this view having checked with Mr Goodman where market consensus was at for Q1 2010. Mr Goodman informed Dr Lynch that a “more accurate reflection of where Q1 consensus really is, excluding the two clear outliers that are above $200m” was revenues of $193.12 million with earnings per share of 25 cents. Dr Lynch forwarded this on to Mr Hussain.

Numerous further versions of the “SH route to 25c” schedule were exchanged between Dr Lynch and Mr Hussain on 31 March 2010, which the Claimants suggested was a demonstration of the importance that they attached to ensuring that an appropriate balance was struck between achieving revenue through (loss-making) hardware sales and maintaining an adequate earnings per share.

In the event, Autonomy reported revenues for Q1 2010 of $194.2 million and earnings per share of 25 cents in its Q1 2010 Quarterly Report. This was exactly the Reuters consensus figure. Reported revenues included $11.8 million of revenue (reduced from the original target of $20 million) from the hardware sales programme (as well as some $26.8 million which the Claimants alleged was improperly recognised in respect of five impugned VAR transactions, $8.5 million from reciprocal transactions which they also impugned, and a further $12.7 million from hosting arrangements which they alleged ought not to have been recognised in Q1 2010). That left $149 million from sources not impugned by the Claimants.

That $11.8 million of revenue for the hardware sales programme was less than the value of the hardware sales achieved and shipped in Q1 2010, which was just over $19.2 million.^{^[177]} In the days following the end of the quarter, Mr Chamberlain, at the instance of what he described as “powers greater than me”, directed “adjustments to revenue”. The Claimants alleged that this was improper and illustrated both the Autonomy management’s attitude to hardware revenues and their propensity to mislead the market, Deloitte and the Audit Committee. I turn now to the issue of revenue manipulation.

Q1 2010: alleged post-quarter end manipulation of the hardware revenues recognised

I have mentioned previously the problems to which an expansion in Autonomy ’ s recourse to hardware sales to make good shortfalls in software sales gave rise in Q3 2009, and the ‘ solution’ adopted at that time of persuading Deloitte to approve the allocation of a substantial proportion of the costs/losses as marketing costs (see paragraphs 917 to 922 above). The continued material recourse to such sales after EMC had withdrawn and Dell had replaced it as hardware supplier for the programme necessitated the development of other ways of minimising the adverse effect of the programme on Autonomy ’ s ‘ bottom line’ figures.

That was, however, easier said than done; and the problem was exacerbated by two factors:

(1) The problem of identifying final figures for quarterly software sales revenue. Delivery of software could be achieved instantaneously over the internet. A considerable proportion of software sales tended to be concluded at the very end (often on the last day) of the relevant quarter, and so software revenues could not accurately be measured until the last moment. Hardware sales, by contrast, required physical delivery and in any event, EMC (or subsequently, Dell and/or Hitachi), would typically wish to fulfil the sale contract in the ordinary course of the quarter.

(2) In the case of both software and hardware, once delivery was effected unconditionally, revenue from the sale was mandated under IAS rules.

The only ‘ solution’ available to this mismatch was to find some way of justifying deferring recognition of hardware revenue which had proved to be, as it were, excess to requirements. The expedient developed by Mr Hussain was to find some basis on which to assert that the hardware sales had not been completed, usually by reference to some alleged delay or imperfection in delivery.

According to the Claimants, Q1 2010 was the first quarter in which Autonomy deferred revenue recognition from hardware sales. The Claimants ’ case in closing was to the effect that Autonomy, at the instigation of the Defendants, treated hardware sales revenue as a discretionary fund, recognising or deferring recognition of such revenue as they pleased in order to achieve (or appear to be achieving) targeted revenue growth whilst at the same time protecting Autonomy ’s bottom line and consequently key parameters including organic growth figure and EPS.

The Claimants contended that this was not only improper in itself, but also was another powerful indication that the hardware programme had nothing to do with protecting and promoting Autonomy ’s software business and was never calibrated or pursued for the purposes avowed by Dr Lynch: it was always about “ revenue revenue revenue” and the need for and use of it was a function of revenue targets.

There was a dispute, however, not only as to whether this reduction in recognised hardware revenue was dishonest manipulation (rather than an appropriate decision against recognition), but also as to whether any such claim was open to the Claimants on their pleading. Before turning to the dispute in more detail, it is convenient to determine whether or not the allegation against the Defendants was sufficiently pleaded.

The Defendants had to acknowledge that there was a reference in a different part of the RRAPoC (under the heading “Involvement of Lynch and Hussain in the transactions themselves” ) but Mr Miles submitted that in the allegations of false accounting for hardware in the RRAPoC there was no express allegation that the deferral of revenue was illegitimate. Nor was it alleged that the matters referred to resulted in an inaccuracy in Autonomy ’s published information.

The Claimants maintained that the plea was plain, that it was well understood by the Defendants to extend to an allegation of impropriety, and that the parties had as a matter of pleading and submission joined issue on it. The pleading on which they relied was expressed as follows:

“On occasion, Lynch and Hussain also discussed and (it is to be inferred) agreed to defer recognising revenue from pure hardware sales from one quarter to the next, and to reduce the level of pure hardware sales that could otherwise have been achieved (and thus to postpone or avoid recognising the associated costs of hardware) where they anticipated that they could meet market expectations with a lower level of hardware sales than had been, or could be, achieved.”

The Claimants submitted that:

(1) Dr Lynch and his lawyers plainly understood the allegation that he (Dr Lynch) faced: when pleading to that paragraph, he stated that the email communications which the Claimants went on to refer to in PoC §135 “ did not relate to any improper deferral in recognising revenue from existing sales … ”. Indeed, Dr Lynch sought further particulars of the allegation, which the Claimants provided;

(2) The same is true in relation to Mr Hussain, stating that “ [a]ll revenue was appropriately recognised in accordance with the relevant accounting standards ”;

(3) Contrary to Mr Miles ’ assertion, this issue was in fact addressed by Mr Welham in his witness statement, in passages on which he was not cross-examined: It was also addressed by Dr Lynch ’s own witnesses.

(4) The point was also addressed in the Claimants’ written opening submissions.

To my mind, there is force in Mr Miles ’s objection to the extent that such an allegation should have been pleaded so as expressly to state in the more relevant part of the pleading that impropriety was being asserted. However, I do not think the Defendants can realistically claim to have misunderstood what was being alleged, or to have been taken by surprise. The evidence engaged with the point. In my judgment, the allegation cannot be disposed of on the basis of a pleading point; and the allegation is relevant to probity in any event.

As to the substance of the allegation, the Defendants argued that (a) the decision was taken because additional information became available such that the finance department considered that part of the hardware resales revenue could not properly be recognised in that quarter; (b) late changes such as these were subjected to intense scrutiny by Deloitte and approved by them; and (c) Autonomy was in any event entitled to manage its sales pipeline carefully to ensure predictability of its results, and was open about doing so in its Annual Reports.

Dr Lynch contended further that there was no basis demonstrated for the Claimants ’ contention that he was involved in the decision. However, when cross-examined, although he could not speak to the actual facts, he justified the principle and practice of such changes as entirely normal: enquiries post-quarter end would often reveal some impediment to recognition such as late delivery or failure to deliver “in good stead”. He insisted that “There ’s nothing inappropriate about that happening past the quarter end”.

To demonstrate the contrary, the Claimants relied on the following demonstration of manipulation, in addition to the matters I have referred to above:

(1) On 31 March 2010, Mr Sullivan provided Mr Hussain with “ a preliminary estimate of how it will come down based on updates from Dell ”. He set out the status and details of the Hitachi and Dell hardware sales that had been achieved in Q1 2010. He stated that he would not have the final “ shipping numbers ” until the end of the day “ so these numbers could change a little ”, but that the preliminary estimate was that $18,583,976 of hardware had been shipped in Q1 2010 (with a further $7,043,598 falling into Q2 2010).

(2) Later that evening, Mr Sullivan provided updated numbers to Mr Hussain and Ms Cynthia Watkins (Autonomy’s Corporate Controller). These showed that $19,224,547.76 of hardware had been shipped in Q1 2010, with a further $7,595,180.74 falling into Q2 2010.

(3) On that basis, just over $19.2 million of hardware sales had been achieved, and shipped, in Q1 2010.

(4) That, therefore, is the amount of hardware revenue that should have been recognised in Q1 2010: but it was not.

As to the thinking behind the sudden reduction in hardware sales in the Q1 2010 Quarterly Report the Claimants contended that:

(1) On 8 April 2010, over a week after quarter end, Mr Chamberlain sent an email to Ms Watkins, copied to Mr Stephan and Ms Harris, subject “ Revenue adjustments ”, stating as follows:

“Firstly, I apologise for the constant changing of your numbers. Powers greater than me are making these decisions and whilst I understand them I know they will be causing you a lot of pain. I will make sure this is remembered when it comes to sorting out Q1 bonuses.

We need to make adjustments to revenue which affects hardware revenue and costs as well as normal licence.

1) Defer Capax (FSA) - $4,285k

2) Defer additional hardware deals as per attached - some ins and outs from last nights schedule

Once processed can you resend TB, consol pack and revenue sheets.

Thanks and apologies again”.

(2) The attachment to Mr Chamberlain’s email was a schedule which showed the percentage of resold hardware which had been shipped in Q1 2010; and the totals after applying the percentages equate to those which appeared in Mr Sullivan’s email of 31 March 2010 referred to in paragraph 980(2) above.

(3) The Claimants submitted that it was to be inferred that what was under consideration was that revenue in relation to certain deals should be deferred for some reason other than delay in shipment.

As to who was responsible for or involved in this revenue manipulation, the Claimants contended:

(1) It was plain from his covering email of 8 April 2010 that Mr Chamberlain was not deciding off his own back what revenue should or should not be recognised in Q1 2010. He was being directed by “ powers greater than me ”.

(2) At the very least must have included Mr Hussain. (Ms Harris thought it likely that it was Mr Hussain, but did not know the extent to which he would have discussed the matter with Dr Lynch). Indeed, that is apparent from Mr Hussain’s email to Mr Chamberlain of earlier that day (with the subject “ costs ”): “ If we can defer $4m from HDS (Morgan Stanley) then I believe we get there ”.

(3) Mr Hussain would have wanted to check with Dr Lynch, and given the close interest that Dr Lynch clearly took in ensuring an appropriate balance between revenue and earnings per share (see paragraph 965 above), as well as his previous (and subsequent) input into how much hardware revenue to recognise, it is improbable that he would not have been involved on this occasion too.

The Claimants contended that given the familiarity that each of them had with the revenue recognition rules (and this would have involved the most basic application of those rules), absent any good reason for such deferral, it would have been obvious to Mr Chamberlain, and also to Mr Hussain and Dr Lynch, that what was under consideration was wholly improper, but that this was not thought by them to be an impediment.

The Claimants sought further to support this contention by reference to various crosses against hardware shipments made on a succession of versions of the schedule prepared by Mr Sullivan (see paragraph 980 above), which appeared to show changing plans in succeeding versions as to what hardware sales (marked with a cross) should be stripped out. They suggested, for example, that:

(1) As at 7 April 2010, the plan had been to defer the revenue (and corresponding costs) in relation to eight hardware sales with a total revenue of $4,472,575, meaning that $15,054,428.69 would be recognised in Q1 2010 (column G, row 38); whereas

(2) By 8 April 2010, only five crosses were included, but the ‘crossed’ deals now included a $5.6 million deal with Morgan Stanley showing a loss to Autonomy of 30% (much higher than the standard 10% and therefore a considerable negative in respect of EPS): this meant that $7,748,041.69 (column H, row 36) of hardware revenue would be deferred, with $11,779,144 being recognised in Q1 2010 (column H, row 38).

A few days later the position had changed again. On 12 April 2010, Mr Chamberlain sent a further email to Ms Watkins, copied to Mr Stephan and Ms Harris, with the subject “Further changes”:

“ We have had to make further changes to your numbers.

Capax (FSA) – back in

Hardware - have had to recognize more HW

The revised hardware sheet is attached again with a column identifying the changes. We have manually updated the consol packs and TBs and processed the journals so you should tie in to these revised numbers.”

The Claimants submitted that by 12 April 2010, it would have been entirely clear what hardware had been sold and delivered in Q1 2010: and that these latter changes cannot sensibly be explained by reference to changes emerging from 8 April onwards. They also pointed out that the one person - indeed perhaps the only person - who would have known exactly what hardware had been delivered was Mr Sullivan, but he was not copied in on any of the above exchanges and does not appear to have been involved in the discussions or decisions as to what hardware revenue to recognise.

Furthermore , they pointed to later documentary evidence that Mr Hussain, at least, was concerned to ensure that Mr Sullivan was not involved or made aware of the late changes. Thus, on 26 April 2010, in an email with the subject “ Low margin business ”, Mr Hussain instructed Mr Stephan that he (Mr Stephan) was responsible for “ keeping right on top of how much has been sold, how much is in deferred revenue, how much in the pipeline ”, requesting that Mr Stephan provide him with “ daily updates ” . In response, Mr Stephan enquired whether he could “share with Sullivan the changes we made to what was recognised/deferred in Q1”. Mr Hussain’s reply was categorical:

“No - you do not give this info to mike S - why would he need it? just get what he has sold on a daily basis - create a spreadsheet - I do not want any inventory at the end of this q”.

T he Claimants concluded with the point that in any event, it is clear from the delivery notes for the deferred deals that (a) the hardware was indeed shipped in Q1 2010 and, save for one, was delivered in Q1 2010 too; and (b) in respect of many of the deferred deals, Autonomy had actually received payment in Q1 2010, disproving (if it were otherwise contended) that these deferrals were anything to do with collectability or issues as to whether the customer had accepted the goods. Payment had been received by Autonomy in Q1 2010.

In my judgment, the decision to defer recognition of this hardware revenue was taken because the revenue was (a) not needed for that quarter to make good any shortfall in software revenue (they had already been covered in other ways); (b) ‘expensive’ in that the costs of it exceeded its sale price and thus had an adverse impact on gross margin, so that (in pursuit of the objectives of the strategy) it was best kept back for use when more necessary; (c) likely to prove useful in future quarters if and when shortfalls arose which had not been covered. That was improper, and the Defendants knew it.

Q1 2010 - the inventory

A consequence of the decision was that the hardware of which the sale had not yet been recognised for accounting purposes needed to be shown as an asset on Autonomy ’s balance sheet as at 31 March 2010 until it was purportedly sold. This meant that there was a significant increase in Autonomy ’s inventory for Q1 2010:

(1) The inventory appearing as a sub-heading “Inventory” under the heading “Assets” in Autonomy’s balance sheet as at 31 March 2010 (and included in the Q1 2010 Quarterly Report) was $10.250 million. That represented an increase of over $9.75 million as compared to the previous quarter.

(2) In cross-examination, Ms Harris did not dispute that the increase in inventory as at the end of Q1 2010 was referable to the revenue from certain of the hardware sales not having been recognised in that quarter.

(3) That can also be seen from the contemporaneous documentation.

i. On 16 April 2010, Ms Harris sent an email to Ms Watkins with the comment: “ We want to classify the 3 invoices at the bottom of the attached list (4,170,350; 4,666,928 & 1024,093) as inventory rather than prepayments ” ). Those three entries - $4,170,350, $4,666,927.90 and $1,024,093.37 - totalled $9,861,371.27. In a further email to Ms Watkins later that same day, Ms Harris stated that the “ equipment was shipped directly to the customer which is stated on the invoices ” .

ii. On 17 April 2010, Ms Watkins responded: “ This is for the balance of 8.3M as an additional 1.5M was recognized. Helen will be sending to you. Its more like 15 invoices ☺ ” .

iii. Following accounting revisions on 12 April 2010, $8,209,780.60 of costs relating to the four hardware deals was deferred (column M, row 36); with $1,508,340 of additional costs being recognised as a result of a further revision in the 8 April 2010 version of the spreadsheet. (column M, row 39).

iv. The fact that the inventory related to these hardware deals can also be seen from an attachment to Mr Chamberlain’s email to Mr Hussain of 16 April 2010.

The Claimants contended, and I accept and find, that Deloitte were misled in relation to the inventory.

(1) On 20 April 2010, Mr Chamberlain sent an email to Mr Stephan, which he copied to Mr Welham, referring to the hardware deals in respect of which revenue was not being recognised in Q1 2010 and stated:

“ to cut a long story short we were not able to prove delivery on these deals, each of them having problems and so we have deferred until Q2 2010. The future receipt of cash and satisfactory delivery notes will provide me with the evidence I need to recognize. They are currently deferred as goods in transit/inventory. ”

(2) In response, Mr Welham stated:

“ The issue we have is ensuring that you have the asset at the balance sheet date. Can you prove delivery now, i.e. post period end to show that you had the hardware and delivered just after period end? ”

(3) Mr Chamberlain responded:

“ We have many delivery notes that show delivery leaving in March but not reaching customer until early April. ”

(4) This was false: as set out in paragraph 988 above, all bar one of the hardware deals had been delivered by quarter end (31 March 2010) and, in many cases, Autonomy had actually received payment for it by that date.

However, the substantial increase in inventory therefore required some explanation to the market.

(1) On 16 April 2010, five days before the results were released, Autonomy issued a trading statement, stating that it expected to:

“ report record first quarter 2010 results in line with analyst consensus estimates of revenues of approximately $193 million and fully diluted EPS (adjusted) of approximately $0.25 ” .

(2) The trading statement continued:

“ In Q1 the company took advantage of discounted offers to purchase stock for the Arcpliance product in advance of Q2 sales, which affected the cash position. These sales have now been completed ” (emphasis added).

(3) Dr Lynch and Mr Hussain were both involved in approving this wording prior to its release.

(4) In a similar vein, the Q1 2010 Quarterly Report stated:

“ Movements in cash flow during the first quarter of 2010 of note included: … Purchasing of inventory of $10 million for Q2 2010 sales, most of which have now completed. ”

However, as both Defendants knew:

(1) The increased inventory in Q1 2010 was not due to stock having been purchased in Q1 2010 for sale in Q2 2010: as explained above, it was due to the decision not to recognise the revenue in Q1 2010.

(2) Further, it was not “ stock for the Arcpliance product ”. Arcpliance was sometimes referred to as ‘Digital Safe in a Box’, as Mr Avila confirmed in cross-examination. It involved Autonomy pre-loading Digital Safe on to hardware and then pre-configuring it.^{^[178]} In contrast, the inventory related to pure hardware which had already been sold and, as Ms Harris confirmed in her email to Ms Watkins on 16 April 2010, had been shipped directly by the manufacturer to the customer.

(3) Nor was it stock that was purchased by Autonomy to take “ advantage of discounted offers ”, but was instead hardware that Autonomy was selling (indeed, had sold) at a loss.

The story was false: but once spun it had to be kept going. As I describe later, the false representation that the $10 million of inventory related to Arcpliance was subsequently reinforced by the Defendants on the Q1 2010 earnings call: see paragraphs 1573 to 1588 below.

My assessment re Q1 2010

In my judgment, these facts and those set out in paragraphs 951 to 994 above demonstrate that:

(1) In Q1 2010 a pattern of manipulating recognition of revenues from the hardware reselling strategy to achieve the best balance between revenue growth and gross margin/EPS is clearly evident;

(2) The deferral of hardware revenue recognition gave rise to the problem that the amount of the deferral was (logically and properly) posted as an asset to Inventory: but the truth as to its source could not be told without revealing the hardware reselling strategy;

(3) At the Earnings Call for the quarter both Defendants provided an explanation for the Inventory which was patently false: see below.

Q2 2010

The hardware reselling strategy continued in Q2 2010. A particular feature of this quarter, which served to increase Autonomy ’s reliance on hardware, was the dramatic reduction in VAR transactions.^{^[179]} There were no allegedly ‘ reciprocal ’ transactions in Q2 2010 either. It was in this quarter that the Hogenson episode occurred, which placed the spotlight on the propriety of both.

The Claimants contended that Q2 2010 provides a particularly striking illustration both of the Defendants ’ reliance on hardware sales and of the way the Defendants tailored their sales, or at least recognition of revenue arising from them, in order to meet market revenue expectations.

On 16 April 2010, Mr Sullivan sent the Defendants a summary of where things stood regarding hardware sales for the quarter. He referred to an “Additional Pipeline of approx. $67m” with caveats that (a) some was “low probability” (b) about $17m was “70% probability or greater” and (c) “Some will likely book or ship outside Q2 2010”. He provided a schedule of prospective deals assigning to each a probability percentage. He said he was:

“ Looking for any guidance on where to draw the line on new pipeline or target Q2 Low Margin Revenue range ” .

On 19 April 2010, Mr Hussain sent Dr Lynch an email referring to the “ consensus ” market expectation of aggregate revenue of $222 million and earnings per share of 30 cents for Q2 2010. Mr Hussain then set out a number of different models:

(1) Taking the cost base excluding hardware as $94.5 million, in order to hit earnings per share of 30 cents (sometimes referred to in shorthand as “30c”), software sales of $190.5 million would be required.

(2) If it was assumed that there would be $15 million of hardware sales and if various software deals could be concluded, revenue of $201.5 million for the quarter would be achieved with earnings per share of 28 cents.

(3) In order to hit aggregate revenue of $215 million with earnings per share of 30.6 cents, Autonomy would need “$20m hardware” along with various software sales; or, if aiming for aggregate revenue of $220 million with earnings per share of 30 cents, “[we] will need either 5m more h/w and lower cost or more s/w”.

He suggested two alternative targets, explaining that the question was “more of how much h/w to take…”

(1) “So could go for $215 m and 30c (most likely outcome) but 10% organic growth means share price likely off a lot”;

(2) “Or could go for $220m and 30c (13% organic growth) - will need either 5m more h/w and lower cost or more s/w”.

Mr Rabinowitz asked Dr Lynch about this in cross-examination, and put to him that “the consideration of how much hardware deals to do was all about what revenue figures you want to get to, correct?” Dr Lynch answered this as follows:

“Yes, I think that’s reasonable. What’s happening here is we’re working out what the forecast is going to be, so, as we were discussing the other day, we lower and raise the forecast as we think prospects are. And, yes, we are using, or at least in one of his options, the option in the middle, he’s using the possibility of 5 million of hardware flex to fit that forecast together.”

Mr Rabinowitz followed this by asking him to accept that “ it’s all about revenue figures…it was about producing revenue, albeit at a cost, that might be used to achieve particular revenue targets?” Dr Lynch’s calmly delivered answer encapsulates the essence of the defence:

“ There’s no doubt that they come with revenue and there’s no doubt that that does have the advantage of giving us a bit of flexibility. Not much because of the delivery issue. But that’s not the primary reason why this was chosen as the strategy. If the primary reason was revenue, there would be better things to do.”

The next day, 20 April 2010, Mr Hussain emailed Dr Lynch some further thoughts as to how to achieve the target . He explained how earnings per share of 30 cents could be achieved by reference to different overall revenue levels with differing levels of hardware sales:^{^[180]}

“ FURTHER THOUGHTS on outlook – tax rate in Q2 is higher:

1. Tax losses mean full year rate will be 24.4%, means Q2 rate will be 21.8% (catch up on Q1)

2. Costs exc. Hardware = $94.5m

3. To hit 30c in Q2 without h/w need $194.8m revenue. My current forecast (inc 15m from jpm and 13m from bav) is $186.5m so $8.3m short

4. To hit 30c with $215m of revs = $18.3m of hardware, means $10m short

5. To hit 30c with $220m of revs = $23.2m of hardware, means $10.5m short

So overall we will need to find $10m or so if we forecast 195, 215 or 220

How can I find it?

- CA deal should yield $6m to $7m

- Grow Safenet, BlueArc etc

My suggestion is to go with 220m and 30c but we can discuss when you waken.”

Later that day, Mr Hussain sent an email to Dr Lynch with the subject “ Interesting point on deals for hardware ”, stating: “ Would be $40m to have no effect on gross margins - i.e. big visibility on q2 ”. It is unclear quite what Mr Hussain meant, but the Claimants suggested (and I accept) that the effect of what he was saying was that if $40 million of hardware sales were made in the quarter that would have no effect on gross margin, but that sales at that sort of level would be very visible to the market.

It appears that, shortly thereafter, Dr Lynch and Mr Hussain agreed the amount of revenue required from hardware sales in Q2 2010. Thus, following a discussion he had had with Mr Hussain on 21 April 2010, Mr Sullivan emailed Mr Chamberlain (with subject heading “Low Margin Revenue” ) that the pure hardware sales revenue goal for Q2 2010 was now $30 million. When cross-examined, Dr Lynch said he assumed that the $30 million figure was one tha t they had agreed would be the objective.

In an email to Dr Lynch dated 28 April 2010, Mr Hussain referred to the pure hardware sales revenue earned to date in Q2 2010 ( “Deferred was $15m New so far is $15.5m Expect $10m more minimum” ) and stated “ I am slowing it down ok ?” Dr Lynch replied two minutes later, “ ok ”.

The Claimants suggested that all this once more confirmed that hardware revenue, or at least its recognition by Autonomy, was a tap which the Defendants turned on or off more or less at will, without regard to any business consideration other than meeting market expectations for aggregate revenue and earnings per share. The Claimants also relied on exchanges between Mr Hussain and Mr Barris of Dell, which showed Mr Sullivan (a) seeking to control the hardware revenue and (b) suggesting deferral to the next quarter once the $10 million which he had mentioned to Dr Lynch seemed likely to be considerably exceeded.

On 5 May 2010, Mr Hussain sent Dr Lynch a document entitled “Getting to 220 SH.doc”. This was an updated routemap to achieving $220 million of aggregate revenue in Q2 2010. The document referred to $16.4 million of “ low margin ” revenue being needed that quarter, and noted (as Dr Lynch accepted) that $13.2 million of such revenue had not been recognised in, but had been deferred from, Q1 2010. This version of the routemap anticipated $13.5 million of revenue from “ BAV ”, i.e. the Vatican Library.

Dr Lynch reviewed and amended the routemap. On 12 May 2010, he sent a revised version to Mr Hussain. The following day, 13 May 2010, Mr Hussain sent Dr Lynch a further version of the routemap. In this 13 May version, the amount of $13.5 million previously shown for BAV was reduced to zero, recognising that no revenue could be achieved that quarter from a VAR transaction with the Vatican Library (discussed in paragraphs 280 to 397 of the Schedule of Impugned VAR Transactions) . Instead, a new line-item “Low margin (cover for BAV)” of $13.5 million was added. As Dr Lynch accepted in cross-examination, Mr Hussain was proposing that, if the sale of software to Vatican Library (through a VAR) was not concluded in this quarter, it could be covered by an equivalent dollar amount of low margin hardware sales: as Dr Lynch put it “that gives him his certainty of 220…”.

On 25 May 2010, in a further version of the “Getting to 220 SH.doc” document sent by Mr Hussain to Dr Lynch, next to the line item “Low margin (cover for BAV)”, the comment “BP can cover this” had been added. This was a reference to a hosting contract which Autonomy hoped to conclude with BP that quarter. The comment appears to signify that if the hosting contract materialised , the pure hardware sales would no longer be needed to meet Autonomy’s revenue target for Q2 2010.

On 2 June 2010, Mr Sullivan informed Mr Hussain that “ We have hit $30m in reselling for the qtr ”, with “ Still more to come ” .

On 7 June 2010, Mr Hussain sent an email to Dr Lynch attaching a document entitled “ Getting to 220 SH 7^th june 2010 (snap shot) ”. Mr Hussain said in his covering email, “ Positive moves on Msft and deloitte allows lower low margin deals means 29c ”. The Claimants suggested that this signified that if increased software sales were to be made, that would give room for lower hardware sales and thereby allow Autonomy to report higher earnings per share of 29 cents.

In this context, Mr Hussain ’s attachment reduced the amount shown for “ Low margin (cover for BAV) ” from $13.5 million to $5.8 million, and annotated the change with the comment, “ Have closed 38m, recognising 23.1m ”. That, as interpreted by the Claimants, suggests that, even though Autonomy had in fact closed ‘ pure ’ hardware transactions which would entitle it to recognise $38 million of revenue in Q2 2010, Mr Hussain was proposing that Autonomy should recognise only $23.1 million of that revenue within the quarter, leaving an amount of further pure hardware revenue that Autonomy could recognise in a later quarter as and when required.

Dr Lynch did not agree with this interpretation and suggested that all that Mr Hussain meant was that “he may have some discretion in terms of saying to the customer Can I deliver that to you in three weeks ' time….” That was not convincing: and Mr Hussain did not suggest that his proposal had anything to do with what hardware had been delivered in the quarter, because he was plainly assuming that it would all be delivered before quarter end, which was still over three weeks away.

A few days later, on 11 June 2010, Mr Hussain sent Dr Lynch a further version of the routemap. As before, $38 million of low margin hardware deals had closed, with $23.1 million of the revenue to be recognised.

On 16 June 2010, Mr Hussain provided Dr Lynch with another update: by this point, with revenue for the quarter at $217 million, Mr Hussain planned to include low margin hardware sales of $23 million, meaning that earnings per share would be at 28.3 cents. As at 18 June 2010, the plan remained to recognise $23 million of hardware sales in Q2 2010. The same was true as at 21 June 2010.

On 29 June 2010, Mr Hussain sent Dr Lynch a further update tracking Autonomy ’s success in concluding large deals prior to the end of Q2 2010. The update gave a snapshot of the status of each deal as at five dates between 16 June 2010 and 29 June 2010 and also identified the amount of revenue proposed to be recognised in respect of “ Low margin ”, i.e. pure hardware sales. As stated above, at 16 June 2010, 18 June 2010 and 21 June 2010, the proposal was to recognise $23 million in respect of pure hardware sales. However, by 28 June 2010 it had become apparent that “ NO DEAL ” could be secured on three hoped-for software transactions. As a result, the figure for “ Low margin ” was increased to $30 million to compensate for the corresponding shortfall.

At the foot of the spreadsheet, Mr Hussain identified the effect on earnings per share of recognising more hardware revenue. He stated that if $30 million of hardware revenue was recognised, earnings per share would be 26 cents (row 38); whereas if only $23 million of hardware revenue was recognised, with a further $7 million of software sales instead ( “ EPS (with $7m s/w) ”), earnings per share would be 28.7 cents (row 39). Dr Lynch accepted that this reflected the balancing act that he and Mr Hussain were having to perform :

“ Q. And that really reflects the balancing act that you and Mr Hussain were having to perform with how much hardware sales should be included or recognised because, although it could help you get to a revenue figure, it had a knock-on deleterious effect on your ability to hit the particular earnings per share figures, correct?

A. Yes, that ’ s accurate. We ’ re having to balance not only the revenue, the EPS, but cash positions, a whole series of metrics that we ’ ll be trying to optimise in the balancing.

Q. And that is why, if you could have a software deal which you could recognise, you would not recognise the hardware?

A. If that was an option, if we had the ability to not recognise the hardware, then in that situation where we had too much revenue then we wouldn ’ t do the loss-making hardware transaction, that ’ s correct. ”

Mr Hussain provided Dr Lynch with a yet further update on the afternoon of 30 June 2010 (the last day of the quarter). As before, $30 million of hardware revenue was to be recognised in the quarter.

In the event, Autonomy recognised $31.1 million of pure hardware revenue in Q2 2010. That was one of the highest quarterly amounts in the Relevant Period and constituted over 14% of total revenues reported for the quarter ($221.1 million). It enabled Autonomy almost to hit the consensus projection of $223 million notwithstanding the restrictions on VAR sales and the continuing depression of software sales revenues.

I was not shown and am not aware of any document or exchange between Dr Lynch and Mr Hussain relating to these contemplated sales and Mr Hussain’s projections and modelling, that suggests that any of this had anything to do with marketing.

Q2 2010 - alleged post-quarter end manipulation of the hardware revenues recognised

The revenue recognised for Q2 2010 was once again, however, not the full amount of revenue generated from the hardware reselling strategy. As in the previous quarter, management again determined not to recognise all the hardware revenues from hardware transactions that had taken place in Q2 2010.

On 7 July 2010, a week after the close of the quarter, Ms Cynthia Watkins sent an email to Mr Chamberlain with the subject “ H/W Cost ”. She stated that “ H/W cost is coming in @ 45M ”, providing a breakdown of “ H/W Orders (Shipped) ”. That included “ Q1 related 9,914,858 not accrued in Q1 ” and “ Q1 Deferred 9,861,371 ” which reflected the fact that certain hardware revenues and costs had been deferred, rather than being recognised, in Q1 2010. Mr Chamberlain forwarded Ms Watkins ’ email on to Mr Hussain with the comment: “ disaster ”.

Mr Hussain ’s responded to Mr Chamberlain as follows:

“ You need to check this.

We deferred $15m or so last q (cost $20m or so - including the expensive Morgan Stanley). We are recognizing $16m more ($17.5m or so) - unfortunately that gives $37.5m.

If you can defer the more expensive ones then we hit the $34.2m I have.”

The Claimants contended that, after the end of the quarter, Mr Hussain was thereby instructing Mr Chamberlain to defer the recognition of certain “ expensive ” hardware deals in Q2 2010. They submitted (and I accept) that it is plain that Mr Hussain ’s focus was solely on the amount of revenue that needed to be deferred to meet the targets he had set. That was why he had selected “the more expensive ones” . The Claimants added that there is no suggestion of deferral being based on whether the revenue recognition criteria were met in respect of them, except what may conceivably be implicit in the words “If you can…” , though both Defendants knew that if the revenue recognition criteria were met, Autonomy had no discretion as to whether or not to recognise the revenue.

It would appear that Mr Hussain ’s instruction was actioned. Two days later, on 9 July 2010, he emailed Ms Harris, Mr Chamberlain and Mr Stephan, with the subject “ US costs for the low margin business ”, stating as follows:

“ The US team has put in the complete costs but of course they don ’ t know that we are deferring some revenues.

I have the US pack giving us total costs (EXC. R&D, OPTIONS & BAD DEBTS) from Lisa of $72.9m

If I take out the relevant costs from the latest forecast then we have $27.5m - the difference being $45.4m. I would like you to confirm that this relates to the low margin business and then we will put in the relevant $35.6m number.

Please confirm and report back as the difference is quite large!!”

The Claimants emphasised especially, and described as “striking” , the comment that the “ US team ” (including, I assume, Mr Sullivan) did not know that some of the hardware was being deferred. The “US team” had operational control of the transactions. If there had been problems delivering some of the hardware sold in the quarter, then the “ US team ” would have known that the revenue in question would have to be deferred. This email was not copied to Mr Sullivan; nor to Dr Lynch, who maintained that he was not involved in or aware of this aspect of the Q2 2010 hardware programme.

The Defendants contended that there was nothing odd in the fact that Mr Sullivan was not involved. They cited their witness Ms Harris ’s explanation that revenue recognition was for the accountants to sort out, “not the salesmen who conducted the sales.” Ms Harris told me she “didn ’t have contact with the salespeople.” But after some equivocation as to whom she would approach for the basic facts about delivery and the like, and after suggesting that Mr Hussain might be the more likely contact as he was more senior and “these are big strategic deals”, she conceded in cross-examination that in fact she did not know: she would not have been the “person having those conversations” because she was not in the revenue recognition/revenue accounting team. In any event, the disconnect between the accounting function and the sales operation seems to emphasise the point that the deferrals were driven without reference to the facts.

In terms of the facts, I was not shown any evidence of any factual reason for deferral of revenue recognition from the Q2 2010 hardware resales in question.

Deloitte Report to the Audit Committee on the Q2 2010 Review

In its report to the Audit Committee for Q2 2010, Deloitte concluded:

At the Audit Committee meeting on 20 July 2010, as reflected in Mr Welham ’s email of later that day, it was agreed that disclosure of the impact of the hardware sales should be included in the front-end of the press release for Q2 2010.

I address the exchanges which followed, with Deloitte pressing for transparency and principally Dr Lynch pushing back hard, in paragraph 869 above. For the present I can summarise the result as being that Dr Lynch got his way. All that the market was told was:

“ The small variation in gross margins in Q2 2010 was in line with our expectations due to the sales mix including appliances as discussed last quarter ”.

My assessment of Q2 2010

In my judgment:

(1) The facts relating to Q2 2010 as set out above repeat and reinforce the pattern evident in Q1 2010 of covering shortfalls in software sales revenue compared to forecast, and of also manipulating recognition of revenues from the hardware reselling strategy to achieve the best possible balance between revenue growth and gross margin/EPS;

(2) Dr Lynch resorted to (at best) half-truths to (as the Claimants put it):

“put the market off the scent as to the existence of the pure hardware sales and their use as one of the means by which Autonomy was able to meet market expectations.”

Linkage Analysis

Before turning to the hardware transactions in the next quarter (Q3 2010), I should mention in passing ^{^[181]} the fact that it was from Q2 2010 that Autonomy started providing to Deloitte what were known as “ Linkage Analyses”. This was its response to Deloitte ’s expression of concern about the accounting for hardware sales, and in particular their perception that what (according to Mr Welham ’s witness statement) they had perceived to be a one-off strategic initiative with EMC was becoming “business as usual”.

It was Mr Welham ’s evidence that Deloitte requested (it is not clear when) what he described as:

“an analysis…setting out, on a customer-by-customer basis, what management said was the evidence of the linkage between the loss-making hardware sales and software sales and hosted revenue to the same customers”.

A dispute developed as to what prompted this request, and whether what was provided by Autonomy properly addressed it. Again, I return to consider this in greater detail later. I mention it now primarily as a marker: (a) Deloitte ’s request and the concern that gave rise to it further demonstrates how the hardware sales were being presented to Deloitte as ancillary to the software business; (b) since there was no real linkage, any response was, necessarily, an exercise in disguising that fact; (c) linkage analysis for every quarter became routine from this time on, and became an important part of the way Autonomy presented the hardware sales to Deloitte; and (d) the Defendants were required to devise explanations of the hardware selling for Deloitte which were increasingly detached from reality.

This was matched (indeed exceeded in terms of detachment from reality) by what Dr Lynch chose to tell analysts in the Earnings Call for Q2 2010 which took place on 22 July 2010. I address this further when assessing the evidence that the Defendants consciously sought to conceal the hardware reselling strategy and the inclusion of hardware revenues without differentiation in the revenues attributed to Autonomy ’s software businesses.

Q3 2010

The hardware reselling strategy continued in Q3 2010. Mr Sullivan was set revenue goals which were calibrated, as before, simply by reference to the volume of recognised revenue and incentivised according to their achievement.

By now, Mr Hogenson ’s concerns had apparently been disposed of, and VAR transactions were accordingly restarted after the pause in Q2 2010 in light of his enquiries. This slightly reduced the requirement to make good shortfalls in software revenue out of revenues from hardware sales, since part of the shortfall could, once more, be covered by VAR sales. Thus, in Q3 2010, hardware revenues, though still considerable, reduced to $26.7 million from $31.1 million in Q2 2010, and VAR revenue increased from $2 million in Q2 2010 (having been $26.8 million in Q1 2010) to $23.3 million in Q3 2010. These correlations tell their own story.

An exchange of emails when Mr Sullivan canvassed Mr Barris at Dell as to the value of deals it would be prepared to transact in Q3 2010 revealed an interesting thread of earlier emails:

(1) On 1 July 2010 Mr Sullivan asked Mr Barris to “encourage your team to show us more deals” to which Mr Barris replied the next day “How much do you want?”

(2) Mr Sullivan responded “An additional $20 million.” Mr Barris said that they would “do what we can here…”

(3) The thread is all subject-headed “BofA SW”. That reflects the fact that the first email in the chain was dated 10 June 2010. It was from Mr Thomas Carlisle, Global Account Manager at Dell, to Mr Sullivan. It read:

“I think we can pull off a large SW deal if you are interested. It seems everyone I ’ve talked to agrees we can do this. Let me know. It would be 10-20M.”

(4) Mr Sullivan expressed immediate interest. But after Mr Carlisle had emailed Mr Barris, the latter was less than enthusiastic in an email to Mr Sullivan the next day:

“ Hello Mike

Thanks for your interest. However, I don ’t want a deal like this to lessen your interest in hardware transactions. This would be less strategic to me than hardware.”

(5) Mr Sullivan did not respond: the next email in the print-out of the chain is that referred to in (1) above.

I should acknowledge that none of the parties made any point about those earlier emails in the chain. But, as it seems to me, they appear to indicate a clear focus on Mr Barris’ part on hardware sales which is difficult to square with the depiction of Dell assisting Autonomy to protect and promote its software sales and coordinating its choice of customers to that end; and Mr Barris was Mr Sullivan’s main point of contact in this context.

In any event, the familiar pattern of Mr Hussain asking for more or less “low margin” according to what he from time to time forecasted or feared he would require to meet forecast overall revenue is evident from email exchanges in this quarter, as in the last.

Thus :

(1) On 30 August 2010, Mr Hussain sent an email to Mr Sullivan with the subject heading “I’ll need more margin-thought there were $4m in the pipe for q3?”. There was no suggestion of any link to software sales nor of any advantage to Autonomy’s software business.

(2) On 6 September 2010, Mr Hussain sent Dr Lynch an email entitled “group revenue” attaching a spreadsheet “6^th sept group revenue MRL format”. Within the spreadsheet was a column stating:

	2^nd sept	3^rd sept	6^th Sept	Commentary
Memo-low margin	15	15	17	$24m done

(3) It appears from that spreadsheet that at that time Mr Hussain was envisaging the recognition of $17 million of revenue from hardware sales in the quarter (column E, row 3), as compared to the $15 million included in the spreadsheet as at 2 and 3 September 2010 (columns C and D, row 3), though the spreadsheet also recorded $24m as “done”.

(4) On 28 September 2010, Mr Hussain emailed Mr Sullivan with the subject “deals” stating “DB is fine So need EMC please And probably $2m more low margin.”

In the event , Autonomy recognised about $26.7 million in revenue from hardware sales in Q3 2010, which amounted to about 12.7% of total revenue reported in the quarter ($211 million). Autonomy ’s Quarterly Report stated that this was “the highest Q3 revenues in the company ’s history” and also that “IDOL business grew organically at 18% during the first nine months of 2010”.

No mention was made of hardware sales: in cross-examination Dr Lynch accepted that they were not disclosed to the market.

Q4 2010

Autonomy ’s hardware sales increased again in Q4 2010. In a quarter where revenue had been forecast at about $236 million, which was a considerable increase, Mr Hussain ’s apparent need for the additional revenue from them became especially marked, particularly in December (see paragraph 1054 below).

The pure hardware sales goal for the quarter was $30 million of revenue, as recorded in an email from Mr Hussain to Mr Sullivan dated 8 October 2010 headed “ low margin ”. Mr Hussain wrote: “ Need 30m this q as it ’s a big q. Can you get EMC to do some? ”.

On 15 October 2010, Mr Hussain sent Dr Lynch an email on 15 October 2010 entitled “ getting to 236 ” which specified the targets that had been provided to each part of the business, including “ Strategic sales $30m ”. The subtext is plain: in order to achieve aggregate revenues of $236 million in Q4 2010, Autonomy needed pure hardware sales of $30 million. There is no record of an objection from Dr Lynch.

At around the same time, Dr Lynch and Mr Hussain agreed that Mr Sullivan should be paid a bonus of $50,000 for Q3 2010 because of the “ good job for us ” he was doing. The “good job” was that “last q he got us the emc deal plus the low margin strategic sales”. There is no suggestion of his job extending to any results in terms of building relationships or driving software sales.

Similarly , the group revenue updates sent by Mr Hussain to Dr Lynch (among others) on 31 October 2010, 7 November 2010, 18 November 2010 and 7 December 2010 identified, in the worksheet entitled “ MRL sheet ” intended for Dr Lynch, $30 million as the target set for “ Strategic ”, “ low margin ” revenue.

An email dated 13 October 2010 from Dell ’s Mr Michael Faughnan to Mr Sullivan, subject headed “Dell/Autonomy Top Line” expressed Dell ’s perception of the purpose as being to drive revenue. Mr Faughnan:

“….wanted to know if Autonomy is still looking to drive top line in Q4 - and if so, what conditions you have for the incentives being offered - we are putting end of year plays in front of JPMC & MS before month end and will bundle with your incentives if the appetite for top line on your side is still there.”

Mr Sullivan did not dissent from this perception, replying on 18 October 2010 that Autonomy had a “ large appetite for Q4 deals ”. In a similar vein, on 5 November 2010, Mr Sullivan told Dell’s Mr Barris that Autonomy “ was very interest[ed] in doing more Q4 business with you. Are there more things we can work on together? ”

Dr Lynch accepted that “top line” in the email quoted in paragraph 1051 above was a reference to revenue; but with a characteristic illustration of his point by an analogy, he shrugged it off as a conversation between people in the engine room being told to get the engine running and it is “not the same as the reason as why the captain is taking the ship in a particular direction”. As to Mr Sullivan not having suggested any other reason in his response to Mr Faughnan, Dr Lynch said that this is “the sales people executing on selling”, and that a sale to JPMC and Morgan Stanley would be very much in line with a software business development strategy as asserted by the Defendants. Not unusually, Dr Lynch’s explanation appeared to me worldly and reasonable; but my education in the case has made me more sceptical, and in my view, there was little more to it than “appeared on the tin”: Autonomy needed to sell hardware to “drive top line” and Dell were keen to sell it hardware for that purpose.

The acuteness of this need was demonstrated by a despairing email sent on 10 December 2010 by Mr Hussain to Dr Lynch (under subject heading “ US idol” ):

“Really don’t know what to do mike. As I guessed revenue fell away completely yet SMS report shows massive activity. But I speak with the vp’s [sic] who are far more accurate. Also stouff, Joel and mike I think keep separate sheets and unless I am v wrong don’t discuss the sheets hence plane crashes and they don’t know. We’ve covered up with bofa and hopefully db and doi but if latter two don’t happen it’s totally bad. There are swathes of reps with nothing to do maybe chase imaginary deals.

So radical action is required, really radical, we can’t wait any more.

Everywhere I look at us idol it ’s bad.”

Dr Lynch accepted that he would have discussed this cri de coeur with Mr Hussain, but said that it had to be seen in the context of what he depicted as “a sales management issue in the US, which is about 25% of our revenue, where forecasted deals are turning out to have a much lower probability of closing…” and that Mr Hussain ’s unhappiness was “about the forecasting”.

When cross-examined as to what he had thought Mr Hussain had in mind in suggesting the need for “really radical” action, Dr Lynch initially suggested that this was a call to “basically reorganise the US IDOL sales management” . That may also be so: but the underlying concern expressed is that software sales in the US have fallen away, and hardware sales have been and will be required to plug the difference.

As it was, the action which in fact immediately followed was that Mr Hussain authorised Mr Sullivan to find more hardware deals even if this meant improving the incentives and so increasing the cost to Autonomy. When this was put to Dr Lynch, Dr Lynch said that he did not know whether that was the case “but it may well be”. On 13 December 2010, just three days after Mr Hussain ’s “ radical action ” email, Mr Sullivan informed SHI ^{^[182]} that Autonomy would “ extend an additional 5% discount to SHI provided that SHI orders an additional $3m in Dell product prior to 12/29/2010 ”.

SHI ’s initial reaction was that “ 3 million is an awful lot for us to bring in by the end of the year. Especially when it would sit in our warehouse for 3 months ” . As a result, they enquired whether any of the existing orders could count towards the $3 million. Mr Sullivan responded that he was “ trying to incent for net new revenue so anything that is already in does not help ” . Following further discussion with SHI, Autonomy offered, with Mr Hussain ’s express approval, the additional 5% discount to SHI, provided that it placed hardware orders totalling $2.5 million before the end of the quarter. The goal of “ new net revenue ”, even if it would cost more to bring in, appears clear.

The revenue goal remained the same as of early December. On 3 December 2010, Mr Hussain sent Mr Sullivan an email headed “ what i need ”: “ Low margin - $30m delivered ”. Mr Sullivan then informed Dell of Autonomy ’s appetite for hardware deals that quarter, indicating that it had “ a goal to do $35m in Q4 deals with Dell ”.

The incentives appear to have worked. By 21 December 2010, Mr Sullivan was forecasting that he would secure “ low margin ” sales “ in the range of $28.5m ” and said that “ [o]rders from SHI will be coming in right until the last day of the year ”. Within two days, Mr Sullivan ’s forecast had increased to $29 million, with $27.86 million already having been secured. On 29 December 2010, Mr Sullivan informed Mr Hussain that, with further “ SHI orders coming tomorrow ”, hardware sales should be over $29 million.

In the event, Autonomy recognised pure hardware revenue in Q4 2010 of $35.4 million, which constituted almost 14.5% of total revenue reported for that quarter ($244.5 million). The Claimants noted that a further $25 million of revenue came from impugned VAR transactions, and some $6 million from impugned reciprocal transactions, and some $15.2 million from impugned “Other” transactions, so that on their case, revenue from legitimate sources was $161.4 million, which would have been some 33% short of consensus estimates.

Q1 2011

The pattern of hardware sales continued also in Q1 2011, for which the market consensus projected overall quarterly revenue of $215.9 million (a slight decrease from the previous quarter).

On 10 January 2010, Mr Hussain emailed Dr Lynch with the subject “Targets for 2011” stating:

“Mike,

Below shows achievement for 2010:

EMEA (Perachio, Murray, Hutchinson) $102m

ASIA (Aurora) $13m

US IDOL (inc. OEM Latam) $114m

Protect (Neil) $37m

Promote (Rafiq) $44m

Strategic low margin $101m

Management sales (stouffer) $75.5m

Maintenance, hosted etc $381m

Totals $868.5m

The target for 2011 I propose to be $978m (up 12.6%). I believe that the analyst consensus is $972m with UBS at $979m.

EMEA (Perachio, Murray, Hutchinson) $125m (up 22.5%)

ASIA (Aurora) $13m (up 92%)

US IDOL (inc. OEM Latam) $105m

US Latam $13m (up 50%)

Protect (Neil) $45m (up 22%)

Promote (Rafiq) $55m (up 22%)

Strategic low margin $110m (up 9%)

Management sales (stouffer, Mooney) $100m

Maintenance, hosted etc $400m (up 5%)

Totals $978m

…I have told Mike S that the target is $25m minimum for Q1

Please confirm you are ok with these targets for 2011.”

That set a “Strategic low margin” target for the (whole) year of over 11% of the aggregate revenue target of $978 million. Dr Lynch responded on 12 January 2011 stating, “your call but build in some headroom”.

On 16 January 2011, Mr Hussain emailed both Dr Lynch and Mr Kanter suggesting a target of $977m but still with $110m of “Strategic low margin”. He followed this with a further email to Dr Lynch dated 26 January 2011 under subject heading “draft forecast” also attaching a “forecast for analysts”. The email set out the following, which included a comparison with 2010:

“$975m revenue (up 12%)

125c fully diluted (up 12.5%)

Gross margin 88% (vs 87% ’10)

Operating margin 46% (vs 43% ’10)

Strategic Sales $97m vs $100m ‘10

Q1 ’11 Revs $217m (up 12%), EPS 26 cents (vs 25 cents Q1 ’10 - remember no dilutive effect in Q1 ’10 from the convertible)”

The Claimants invited me to note especially, as being “difficult to square with the Defendants’ contention as to the primary purpose of this activity”, that these targets were set without any analysis, or even mention, as to how they would be achieved, the identities of the parties to whom the hardware would be sold, or how the hardware sales would be used to drive software sales. I am not persuaded that to prove its substance or intent it was necessary for the Defendants to rehearse its purpose, or that whenever a target was mentioned, it would be expected to be accompanied by a statement of “how and to whom” in a forecast for analysts. But I accept that the overall impression given is of the measurement, balancing and use of strategic sales by reference to an objective of growing revenue whilst at the same time preserving gross margin and EPS.

Also clear, to my mind, is that the pressure for hardware sales continued to fluctuate inversely according to the success of software selling, and that the calls over the course of the period to increase or find another source for hardware selling were prompted not by success in encouraging software sales, but on the contrary by any diminution in Mr Hussain’s confidence in them eventuating, or simply and more immediately, if an anticipated hardware transaction fell away.

As an example of the latter, on 25 March 2011, after Mr Sullivan had informed him that a sale to Morgan Stanley of Hitachi hardware which Mr Hussain had described to Mr Sullivan as “the last part of the low margin jigsaw” would not proceed, Mr Hussain emailed Mr Sullivan:

“Need a plan b

Anything you can do on hitachi is good

Aggressively pursue SHI, JPMC etc. Really need to hit $25m”

Mr Hussain’s email the same day to Dr Lynch to report the position spoke only of missed revenue :

“We have done deals many times and were confidently expecting $3m this time round. Although I pressed we didn’t get it and MS have given the orders ($4m) directly…

V v v v frustrating as its $4m of revenue missed.”

The exchanges thereafter also speak loudly of Mr Hussain’s by now almost desperate focus on revenue, with blame and recrimination following any apparent failure to meet his expectations. Thus:

(1) Mr Hussain provided Dr Lynch with a further update early on 28 March 2011. In relation to further hardware revenue, the position remained as it was on 25 March 2011.

(2) Shortly thereafter, Mr Hussain enquired of Mr Sullivan “how are you getting on with replacing MS / Hitachi. Need as much as we can get from SHI”. Mr Sullivan responded that “SHI can’t do anything”, but that there was “[p]ossibly another $1m from BNY, but not looking good to get to $25m”.

(3) Mr Hussain was not impressed. On 29 March 2011, Mr Hussain sent an email to Dr Lynch stating that “Sullivan’s psychology is interesting. MS / Hitachi was a complete cock up yet he isn’t trying to get a replacement for it as his life depended on it”.

(4) Dr Lynch chose to send that email on to Mr Sullivan with the comment “WOW Sushovan says you have interesting psychology................ ” .

(5) Mr Sullivan replied to Dr Lynch saying:

“Sushovan is wrong. I have been burning up the phone lines trying to find more revenue. Was I suppose [sic] to lie and give him a more rosy outlook? He just doesn’t like my forecast. It is hard to drum up $3.5m in revenue in just a few days. But In fact I have found more revenue and am pushing on the reps, Dell managers and resellers I know. Revenue is now up to $23m in to date. If you check with Chamberlain he will tell you that I was working a deal just yesterday in Ireland (It fell out of the qtr). I tried to pull in some JPMC business in Europe - no go. I have offered incentive to resellers (no takers). We processed about 300 orders this qtr so far. I am basically running a small side business and it takes a lot of time, networking, paperwork etc. I am not sitting on my ass.”

(6) On 30 March 2011, Mr Hussain sent further updates to Dr Lynch. By this stage, $152 million of revenue had closed, with Mr Hussain saying that it was unlikely that there would be more than $250,000 of further hardware revenue (row 18).

(7) Later that day, Mr Hussain’s only response to an email from Mr Sullivan reporting on progress with a prospective $1.6 million software deal which he thought would not come in before the (imminent) end of the quarter was to ask Mr Sullivan whether he had found “ some more hardware revenue? I need it ”.

(8) Mr Sullivan responded (presumably sarcastically, as the Claimants said) “ I did not know that”, and indicated that he was expecting “ another approx $100k tomorrow ”.

The Claimants presented these exchanges as further demonstrating that (a) the hardware reselling was all about generating revenue, with no other objective discernible from any of the emails and being evident from Mr Hussain’s almost panicked reaction to any sign of impending shortfalls, and (b) Mr Sullivan’s perception of himself to be running a “small side business” was in substance accurate and revealed that there was no properly integrated sales approach, with the two sides of the business working together and utilising contacts and discounted hardware sales to sweeten and encourage software deals , such as surely would have been readily apparent if the strategy really had been to use hardware sales to drive software sales.

In the event, and once again, Mr Hussain’s panic was unwarranted. Autonomy beat the market consensus for Q1 2011 of $215.9 million with reported revenue for the quarter of $220 million, gross margins of 88% and adjusted earnings per share of 26 cents ^{^[183]}. Of this, reported revenue of some $20.1 million represented what the Claimants described as ‘pure’ hardware sales in Q1 2011, which constituted just over 9% of total revenues of $220 million reported for the quarter.

There is, however, a story behind the figures as presented, and in particular, as to how the balance was so successfully struck between low margin hardware sales and the important metrics of gross margin and EPS.

Q1 2011 and alleged revenue deferral

The story arose out of an incident at around this time which the Claimants contended was a further demonstration of dishonesty on the part of the Defendants in their approach to the question of revenue recognition and their use of hardware revenues as a discretionary fund.

On 9 April 2011, over a week after the end of the quarter, Mr Hussain sent an email to Dr Lynch and Mr Kanter with the subject “ management meeting analysis ”. He attached a spreadsheet - entitled “ revenue analysis SH MRL AK only ” - which he said could “ be used as the basis for data analysis at the management meeting ”. That spreadsheet showed, in the “ Q1 actuals ” tab, that, as compared to his hardware revenue target for Q1 2011 of $25 million, Mr Sullivan had achieved $22.1 million. However, as in previous quarters, discussions took place as to whether all that hardware revenue should be recognised in Q1 2011.

On 11 April 2011, Mr Hussain sent an email to Dr Lynch, with the subject “ Q/e ”, stating:

“If we defer prisa^{^[184]} then we are at 218.1m but 24c and 85% [gross margin]. If we don’t defer prisa but defer equiv low margin we are at same revs but now at 25c and 88%. To discuss when I land or you can discuss with steve.”

In short, the Claimants submitted that Mr Hussain was identifying various options that he considered might be adopted in relation to revenue recognition at this time: h e was questioning whether to recognise revenue in Q1 2011 on the DiscoverTech transaction (which was, according to the Claimants, not entered into until early April 2011, but was backdated to 31 March 2011), or instead to recognise hardware revenue.

The former course would have an advantageous effect on gross margin and earnings per share. If hardware sales were made and hardware delivered, the associated revenue had to be recognised. If a deal with or for Prisa had not been completed before the end of the quarter, there was no revenue to recognise.

What, according to the Claimants, Mr Hussain was engaged on and seeking approval for was a ‘pick-and-choose’ approach to revenue recognition, reverse engineering their decision from the desired level of Q1 2011 aggregate revenue, gross margin and earnings per share, regardless of the underlying facts. Mr Welham had told the US criminal trial court that he was “troubled” by Mr Hussain’s suggestion that, two weeks after the end of the quarter, there was an open choice whether to defer recognition of revenue from the Prisa transaction or defer recognition of a hardware sale.

In cross- examination , Dr Lynch stated that he did not accept that Mr Hussain was “ saying that there’s necessarily a choice in all of this ”. He suggested that there might be issues and a judgment to be made, for example, as to the reseller’s credit, or, in the case of the hardware, as to whether delivery had been accepted. However, this was not convincing: Mr Hussain certainly appeared to be offering a choice, making clear to Dr Lynch that they could do one thing (“ defer prisa ”) or the other (“ defer equiv low margin ”). Dr Lynch said that he did not know whether he had a discussion with Mr Hussain following the email, but I accept that it is unlikely that, having raised the point with Dr Lynch, Mr Hussain would then have acted on his own.

In any event, the Claimants submitted, the fact that Mr Hussain was presenting a choice for discussion is put beyond doubt by Mr Chamberlain’s email to Mr Hussain the following day, 12 April 2011, with the subject “ Numbers ”:

“Three options:

1) recognize Prisa, defer $3.6m Hardware - $218.1m, 88%, 24.8c

2) Defer Prisa, recognize $3.6m HW - $218.1m, 86%, 23.7c

3) recognise BBC $1.6m, defer Prisa, recognize $2.0m HW - $218.1m, 87%, 24.2c

Need to speak asap to lock this down. We announce in 9 days and if we don’t stop moving I cannot deliver timetable.”

The Claimants interpreted the “ Three options ” that Mr Hussain and Mr Chamberlain presented as being the following:

(1) The first option was to recognise the Prisa VAR transaction and to defer the equivalent amount of hardware revenue ($3.6 million), which would result in total revenue for the quarter of $218.1 million, a gross margin of 88%, with earnings per share of 24.8 cents.

(2) The second option was to defer Prisa and instead recognise hardware revenue of $3.6 million, which would result in the same overall revenue ($216.1 million), but a lower gross margin (86%) and earnings per share (24.8 cents) due to the hardware deals being loss-making.

(3) The third option was to recognise a transaction with the BBC with revenue of $1.6 million, again defer Prisa, but recognise hardware revenue of $2 million, which would generate the same overall revenue, but mean that gross margin and earnings per share were at 87% and 24.2 cents respectively.

The need to decide between the options was emphasised by Mr Chamberlain ’ s concluding paragraph, in which he said that he needed to speak to Mr Hussain as soon as possible so as “ to lock this down ”.

Shortly after sending his email to Mr Hussain, Mr Chamberlain forwarded it to Ms Gustafsson and Ms Harris, stating “ These are the options and the impacts if he calls again ” (emphasis added).

Dr Lynch told me in cross-examination on Day 43 that this was not how he would interpret the “three options” email. As the BBC contract contained a services component where the customer had to inform Autonomy that milestones had been met, he thought it may well be that Mr Hussain and Mr Chamberlain were trying to get the milestone confirmation out of them. He added:

“What Mr Hussain I think is doing is saying: this is the minimum ways, but obviously we could end up in a situation where we get confirmation from the BBC, the VAR pay us some money and we can take that, in which case we have the upside.

I don’t think there’s an idea that these things are completely moveable parts. Some are judgement, but some aren’t .”

Dr Lynch was further cross-examined on this aspect a few days later (Day 52). By then, he seems to have felt able to be more certain and more specific, although he was a little unclear whether he was aware at the time and could now remember these matters from that time, or whether they were reconstructions from documentation he had seen, or whether he had gathered the information from emails or other such sources. He clarified that he did not have “an explicit memory of it” but the best he could do to explain his source was to refer to an email he thought he had seen “where there was a discussion of the fact that there were multiple purchase orders and what they were going to do” and “some sort of email that was sent to me about which deal was going to be taken or not”.

When further cross-examined on this issue, he asserted that “The BBC was an acceptance issue” and that Dr Menell was talking with someone called Chris Westmacott at the BBC “to get them to sign off on and acceptance within the quarter” ; and as to the multiple purchase orders issue he now clarified that DiscoverTech had submitted three orders in the same quarter, and “there was a judgment to be made about whether DiscoverTech’s credit was worthy of having three or two or one, or which of the two, and that decision had to be made”. He added that at the same time there were issues about whether all the necessary proof to meet acceptance criteria had been met and stressed that “this is part of the normal process that happens past the quarter end where all the paperwork and the terms and the evidence is reviewed.”

I do not consider that I should place any substantial weight on Dr Lynch’s evidence as to particular factors which might have justified deferral of revenue recognitions. My impression was that his second, more detailed, version was a reconstruction worked out by Dr Lynch by reference to other witness statements and documents. There are echoes of (and it may be that Dr Lynch may have borrowed or half- remembered ) Ms Gustafsson’s witness statement, where she stated that the “ Three options ” “ email does not strike me as problematic ”, and suggested that there may have been accounting judgements under consideration bearing upon what revenue could be recognised in the quarter.

I do not consider Dr Lynch was knowingly misstating the position; but I do consider that he was trying to varnish his previous evidence on the same issue, and that the truth is that he had no reliable evidence to give of the actual events. In a way it is surprising that he even attempted to suggest otherwise: for his ultimate defence was that he was not involved in the deferral process, except as a recipient of and probably brief consultee on Mr Hussain’s email of 11 April 2011 (see paragraphs 1076 to 1080 above), which (as explained below) he submitted was a report updating him as to different possible outcomes of deferral, rather than an invitation to choose between options.

In general terms I accept that the final resolution of matters such as whether goods shipped have been duly delivered and accepted without objection, and whether revenue may properly be recognised accordingly, will take place in the days after the end of the quarter, and in the event of real doubt may take some time, as for example, Mr Bloomer confirmed, (and Mr Welham had, after much equivocation, and contrary to the impression he sought to give before me, also agreed when cross-examined in the US criminal proceedings). In my judgment, therefore, the Claimants’ supposition that everything should have been determined by the end of the quarter is unrealistic, given the typical end of quarter flurry; and the Claimants’ criticism of that process is misplaced. I also consider it likely that the decisions may be influenced in fact by the animus with which the issue is addressed: if more revenue is not required there may be, without real impropriety, a tendency to resolve any small doubt as regards the fundamentals such as time of actual acceptance in favour of the preferred result.

Nevertheless, the question is whether in the particular context the exercise and delay were in fact occasioned by the need to check matters relevant to proper revenue recognition (“the fundamentals”), or whether the exercise was simply one of how best to manipulate the timing of revenue recognition without regard to those fundamentals, with the objective of improving Autonomy’s figures, and especially EPS. It is not an issue as to the ordinary process; it is a question whether (a) there was any doubt as to the fundamentals and (b) whether the cabal ignored the fundamentals and simply chose the best solution from the point of view of EPS.

There was no documentary evidence to suggest any real doubt as to the fundamentals. In addition to the “options” emails, the email exchanges suggest that the figures given for hardware revenue change according to the objective of increasing EPS whilst maintaining revenue growth, rather than by reference to any new emerging evidence or concerns. Thus, for example, the Claimants made the point that on 14 April 2011, Mr Hussain sent an email to Mr Chamberlain, Ms Gustafsson and Ms Harris, with the subject “ Q1 ”, in the following terms:

“Apologies upfront for this, i wouldn’t be pushing if it wasn’t important:

Please could you look at:

- Any more maintenance revs (maybe $250k)

- Defer similar amount of low margin

- R&d capitalisation / Depreciation / bad debt / accruals of $750k

- Taxes - i wonder if you can hit 26.75% ($500k save?)

Does this get 25.6c?”

There is no sign that any of this was being driven by any new evidence about fundamentals or revised accounting judgements: the plain focus is on getting to EPS of 25.6c.

Perhaps a different slant might have been put on the matter had any of the three persons definitely and directly involved (Mr Hussain, Mr Kanter and Mr Chamberlain) appeared before me as witnesses; but none did. There was no documentation to support the theory, as Dr Lynch acknowledged. Mr Chamberlain did not mention any matters of accounting judgement in his email. Mr Sullivan, who would be the person most likely to know the operational details, was not copied in on the emails and was very definitely not invited to the “management meeting” (see paragraph 1075 above).

I have concluded and find that the management meeting and the exchanges described, which smacked of decision-making by a clique or cabal of three (with Dr Lynch as the éminence grise) determined to and did choose between the options, not according to the fundamentals, but according to the objective of minimising any adverse financial effect of the loss-making hardware sales and maximising earnings per share.

Had the issue been about ‘fundamentals’ it seems to me likely that the advice of Deloitte and approval of the Audit Committee might have been sought. It was not. None of the ‘cabal’ (or Dr Lynch) consulted or shared the issue or their decision with either Deloitte or the Audit Committee. I accept the evidence of Mr Welham and Mr Bloomer that neither was informed or aware that this had occurred.

Defendants’ involvement/knowledge

Mr Hussain did not dispute that he well understood that this was not a choice properly open: if delivery had been completed the revenue arising had to be recognised. He did not need Deloitte to tell him that; and he was in a better position than they were to determine the matter. In his written closing submissions, Mr Hussain appeared to rely on a defence that the revenue was properly deferred and that this was confirmed by Deloitte’s approval. It was submitted that Deloitte checked every major sale as part of its quarterly and annual audit testing and had never identified any revenue which was improperly deferred. However, Deloitte were obviously not party to the cabal, or aware of it; and it is not clear what they were told from time to time about what I have called the fundamentals. I do not think any such defence avails Mr Hussain.

Except for Mr Hussain’s email of 11 April 2011 (see paragraph 1076 above) setting out the varying effects (Dr Lynch said “possible outcomes”) according to whether the Prisa transaction or the “equiv low margin” were to be deferred, none of the other emails was sent or copied to Dr Lynch, and he was not at the meeting of the cabal of three. He was adamant that he would have played no role in such decisions. He felt able to offer his understanding of the process in general terms, supplemented by his reading of the documents for the purpose of his evidence; but his evidence was that he was not involved in any of the details of such matters at the time.

Ms Harris confirmed, when she was cross-examined, that Dr Lynch was not really involved in this level of detail; but this was a somewhat generalised statement, and, for example, she accepted that she had no way of knowing whether Dr Lynch and Mr Hussain had discussed the decision between options presented in Mr Hussain’s email. Dr Lynch himself was less than certain in this regard: he told me that he did not know “if [he] would have done or not” .

I have described Dr Lynch earlier as the éminence grise of the clique or cabal. It seems to me most unlikely that he did not discuss the options carefully with Mr Hussain, especially since Mr Hussain had expressly asked for such a discussion. In any event, this was not the sort of decision Mr Hussain would have made without Dr Lynch’s approval. For reasons to which I turn and elaborate when discussing later all the instances of backdating and the issue of what Dr Lynch knew about them, I consider that Dr Lynch was aware that Mr Hussain would take steps necessary to ensure recognition of revenue or not according to the election made, and that this would depend on whether the revenue was needed in the previous quarter to make up a shortfall, or was better reserved for the subsequent quarter. It may be that Dr Lynch did not know how the election would be documented, and it is possible that Mr Hussain did not refer back to Dr Lynch before finalising that (the die having already been cast), but I cannot accept that Dr Lynch was not aware that the decision was going to be made on the basis not of chronological truth but revenue need, and that if the latter prevailed, any documentation of the decision would be false. I consider and find that Dr Lynch (a) discussed the matter with Mr Hussain (b) was party to and/or approved the decision ultimately made and (c) knew that the choice made involved dishonest backdating and the improper deferral of low margin revenue in Q1 2011, falsifying Autonomy’s accounts and published information.

In my view, that is the more likely pattern of the events in question; but it is reinforced by my general view that no decision of real substance regarding revenue recognition was done without Dr Lynch’s approval.

Q2 2011

The hardware sales goal for the whole of 2011 set in early January 2011 was $110 million (as part of an aggregate target revenue of $978 million). This had been revised later in January: on 26 January 2011 Mr Hussain had sent Dr Lynch a further spreadsheet with revised targets for overall revenue of $975 million (a reduction of $3 million) and for hardware (described as “strategic sales”) of $97 million (a reduction of $3 million also, so that the entire reduction was to come from reduced hardware sales). That same spreadsheet had targeted “strategic sales” of $30 million for Q2 2011.

The hardware sales goal was then revised down to $27.5 million at the beginning of Q2, as recorded in an email from Mr Hussain to Dr Lynch on 9 April 2011. At some point in May or early June it was further revised down to $25 million: it was a difficult quarter for hardware sales. That target too proved to be unattainable, and in the end, Autonomy recognised revenue from hardware reselling totalling $20.8 million for Q2 2011 (compared to $20.1 million in the previous quarter).

As before, most of the hardware business this quarter was done through Dell in the usual way, with Autonomy acting as reseller. The pattern was much the same as in previous quarters, but the following points were singled out by the Claimants as further demonstrating their case that the only true objective of these sales was revenue, and not the promotion of Autonomy's software business (which is not mentioned in the relevant correspondence).

First, the Claimants submitted that a good illustration of the typical way in which these deals worked can be seen from an email chain relating to a transaction involving Dell (as hardware supplier), Autonomy (as reseller) and an insurance company called Progressive (as end customer):

(1) On 2 March 2011, Autonomy quoted for the supply of certain Dell hardware to Progressive.

(2) Three months later, on 2 June 2011, Dell informed Autonomy that they had “ ended up losing this round ”, but that:

“They have another potential purchase coming up later this month. Will you be able to participate? They told me we are close on your pricing. I would imagine that another $25 off will win the business. This order may be for approximately 2000 systems.”

(3) Thus, it was Dell rather than Autonomy that was dealing with Progressive; indeed, Autonomy had no knowledge of the deal. It was Dell that was seeking to make the sale. Dell invited Autonomy to participate in the deal, for its benefit: Autonomy was to be interposed into the deal and its involvement limited to funding the discount so as to make Dell’s bid more attractive.

(4) Mr Jorge Salazar, an Autonomy Contracts Manager, responded the same day to say that he was sure that Autonomy could “ make ourselves available to participate on a new bid ”, but that the decision lay with Mr Sullivan who he copied to his email.

(5) On 3 June 2011, Mr Sullivan enquired as to the total deal value and timing of the orders. He made clear that Autonomy’s willingness to increase the discount depended on it being able to count this hardware sale as revenue in Q2 2011. Mr Sullivan concluded: “ We can only get more aggressive if the orders come by June 30 ”.

(6) There is nothing to suggest that Mr Sullivan displayed any interest in using the prospective hardware sale as a way of promoting a related software sale to Progressive: it would appear that he was entirely focused on what he had described as his “small side business” .

Secondly, the push to secure revenue in Q2 2011, with undisguised emphasis on the need for revenue without reference to any other objective, was directed from the top. So, for example, on 1 May 2011, Dr Lynch informed Mr Egan and Ms Eagan that they both “ must be focused on this Q revenue now and NOTHING else ”.

Thirdly, when Mr Sullivan, after being pressed by Mr Hussain by email dated 10 June 2011 whether it was “still on for 25m?” , replied negatively, citing “Way more moving parts than normal at this time of the qtr”, complaining that “We were tied to a lot of deals, but Dell keeps losing them”, and stating that “I don’t see a way to get to $25 without absolutely everything coming in on time which is a longshot” , Mr Hussain the next day reported to Dr Lynch somewhat hysterically that Mr Sullivan was “backtracking on his commitment of $25m” and “being slippery” . This arguably illustrated that Dr Lynch was very much involved, and that Mr Hussain (who had on 28 May 2011 told Dr Lynch that “Low margin should hit $25m…”) was increasingly desperate both to please him and avoid blame.

Fourthly, the Claimants cited also Mr Hussain’s reaction which was to increase the target for pure hardware revenue. He directed Mr Sullivan to generate $29 million of pure hardware revenue: “ As you know this is a key quarter. So what we need from you is as follows: … Low margin $20.5m net new ($8.5m done already, you need to close Morgan Stanley and JPMC) ”. In order to provide an additional incentive, he offered Mr Sullivan a $50,000 “ spiff ” (i.e. bonus) if he achieved this goal as well as a goal of $12 million of new licence revenue. Dr Lynch accepted that he and Mr Hussain offered Mr Sullivan this bonus in the hope that he (Mr Sullivan) might be able to land further hardware deals so as to reach the increased revenue target.

Fifthly, the email reports seem to indicate the compulsive focus, with ever changing forecasts, and collapses met with immediate offers of further incentives, despite the cost, and without any sign of any measurement against benefit in terms of anything other than revenue. Thus:

(1) On 16 June 2011, Mr Hussain informed Dr Lynch that the “ low margin ” deal with JP Morgan “ was strong again yesterday so 20m total should be achievable ”.

(2) On the same day, Mr Hussain sent a spreadsheet to Ms Gustafsson asking that she “ have a look at the low margin and the management sheet and check closed deals ”. The attached spreadsheet contained a tab labelled “ Sullivan ”, which detailed hardware deals. Row 40 in that spreadsheet noted that Autonomy should “ consider throwing some incentive at JPMC to place $10m in orders before June 30 - an extra 1%...? ”.

(3) Two days later, on 18 June 2011, Mr Hussain provided Dr Lynch with a further update, stating that the JP Morgan hardware deal was “ difficult again - likely $16m to $17m ”. By this time, Mr Hussain was projecting $24 million of hardware revenue in the quarter of which $18 million was “ new ”: see the fourth entry in the table.

(4) On 25 June 2011, Mr Hussain sought an update from Mr Sullivan, including in relation to “ Low margin ”. Mr Sullivan’s response set out the “ reselling ” deals that had been secured and the status of the deals that were still in the pipeline. He identified a likely deal with Morgan Stanley/Hitachi for $2 million, stating that it had “ [r]educed $ this week but is coming in ”. He reported that he was trying to secure a hardware deal with BofA by “ [g]iving some extra incentive to get [it] ”.

(5) On 29 June 2011, Mr Sullivan informed Mr Hussain that, “ Out of left field ”, the Morgan Stanley deal had gone away. Mr Sullivan forwarded an email chain from Hitachi, which recorded that “ Morgan will not be doing the deal through Autonomy or Direct. They now want to trade in other equipment ”. Mr Hussain’s response to this news was “ Shit ”.

(6) On 30 June 2011, Mr Hussain provided Dr Lynch with an update on the “ state of play ”, listing out deals which it was hoped would still close before quarter end.

(7) As it was, as previously explained, Autonomy recognised revenue from pure hardware sales totalling $20.8 million, thus falling materially short of its target.

In the round, the Claimants submitted that this pattern demonstrated that the objective was plainly to produce reportable revenue, with no sign or evidence to support the Defendants’ case that the purpose was a marketing one designed to drive software sales.

In the event, for Q2 2011, Autonomy recognised revenue from what the Claimants categorised as “pure” hardware sales totalling $20.8 million out of a total reported revenue of $256.3 million, beating the market consensus estimate of $250.5 million.

The Claimants’ case was that if revenue from transactions impugned by them (i.e. from hardware sales, impugned VARs, improperly accelerated Hosting revenue and improper “Other” transactions) only some $190.5 million revenue was achieved (a miss of some 24% on consensus estimate).

Q3 2011

Although their hardware claim relates to the period from Q3 2009 to Q2 2011, the Claimants referred also to the continued focus on hardware sales to achieve revenue forecasts and to evidence of exchanges in Q3 2011 and at the time of the HP acquisition as providing further support for their case.

Thus:

(1) On 27 July 2011, Mr Hussain sent Mr Sullivan an email headed “ low margin sales ” asking “ What is your expected number for this q. As usual i will need $25m ”. Mr Sullivan replied that he was “ on it ” and that there was “ Lots of good stuff in the works ”.

(2) On 18 August 2011, the day that HP’s acquisition of Autonomy was announced, Mr Sullivan forwarded an email to the Defendants saying:

“ Getting bombarded with calls from Dell and EMC. I have a lot lined up with Dell. Given the competitive situation, I will need guidance. I have committed to a lot of Q3 reselling revenue. I assume I keep going… There are deals lined up in future qtrs as well ” (emphasis added by the Claimants).

There is no hint in any of this of any marketing purpose; the focus is entirely on the production of recognisable revenue.

Following the completion of the Autonomy acquisition (on 3 October 2011), Mr Sullivan emailed Dr Lynch and Mr Hussain (on 11 October 2011) stating, “ I am still taking reselling orders from Dell. Do you have a reselling revenue goal for the quarter (through Jan)? ”. On 7 February 2012, Mr Sullivan stated that the hardware deals were to continue for the time being as “ MRL still wants the revenue for now ”. Again, these emails simply confirm what is already apparent from the material set out above, namely that the purpose of the pure hardware sales was to generate revenue.

In a rather different vein, the Claimants referred to representations made by Dr Lynch before and at the time of the acquisition which they contended demonstrated the extent to which the hardware reselling strategy was kept quiet, and Autonomy was presented as being differentiated by its exclusive focus on software sales.

In particular, they cited (a) a copy of a Q&A script posing and answering questions which might be put by the press on announcement of the acquisition, which Autonomy prepared and shared with HP on 17 August 2011 (which was the day before the acquisition) and (b) a rebuttal prepared by Dr Lynch to a negative description of the HP acquisition in an article in the ‘eDiscovery Journal ’ which was sent to him in draft. As to this:

(1) In the Q&A script, Question 31 was “ What is the product fit between the businesses? ”, to which the planned response was “ Perfect fit, no overlap. Barcelona [HP] is a hardware and IT Services business and Arsenal [Autonomy] is a pure software company ”. Dr Lynch reviewed earlier drafts of the script that contained this planned answer. The Claimants cited this as demonstrating a contrast of what was to be stated then and the case now being advanced by the Defendants: then, the distinction was being drawn between HP as “ a hardware and IT Services business ” on the one hand and Autonomy as “ a pure software company ” on the other. The assertion that there was “ no overlap ” between the two companies was designed to reinforce the message that Autonomy was not a company engaged in the sales of hardware.

(2) E-Discovery Journal was planning to publish a negative description of the HP acquisition, which it shared with Autonomy in draft prior to publication. Dr Lynch took umbrage in relation to the draft, describing it as “ a real mess of rumor [sic] innuendo and incorrect technical comments ”. He produced what he called a “ rebuttal ”, which he shared with others within HP and Autonomy. This contained a section headed “ Product rationalisation and overlap ”, which stated as follows (emphasis added):

“ The report poses the question:

‘ What is the go-forward plan for product rationalization, both within the Autonomy portfolio and between the Autonomy and HP portfolios? ’

But none of this is analysed either in the report, which instead makes vague references to product overlap. In fact there is very little overlap which is why the acquisition makes so much sense. The eDiscoveryJournal seems to deliberately shy away from stating the obvious: that HP is primarily a hardware company and Autonomy is a pure software business.”

(7) Incentivisation of Mr Sullivan purely by reference to hardware sales revenue

The Claimants also relied, as being consistent with and demonstrative of the Defendants focus on pumping revenue from loss-making hardware reselling, on the fact that in Q3 2009 and various subsequent quarters the Defendants pressed and incentivised Mr Sullivan to achieve revenue targets by offering bonuses calibrated by reference to revenue, without any suggestion of any regard being had to the identity of the customer, its IT needs or the volume of any sales of Autonomy software that might flow from hardware sales.

These bonus targets and promises started from the first quarter of the programme, and continued throughout the Relevant Period. Thus, in respect of Q3 2009, Autonomy agreed to pay Mr Sullivan what was described by Mr Hussain in an email to Mr Sullivan dated 15 September 2009 as “a special $200,000 bonus” for delivering $30 million of recognisable revenue by 30 September 2009, with a promised supplement of another $100,000 “should you manage to get to $50m of recognisable revenue”.

Similarly calibrated and very handsome bonus payments were promised and made to Mr Sullivan in subsequent quarters, with no stipulation or even mention on any occasion of any required correlation of hardware sales to software sales or contacts.

The Claimants pointed out also that none of these bonus payments was revealed to Deloitte or the Audit Committee, nor was the fact that they were promising incentives calibrated by reference to hardware sales targets.

In their written closing submissions, the Claimants contended that if the purpose of the hardware sales was as the Defendants contended:

“…it is incomprehensible that Mr Sullivan was promised bonuses - which were also concealed from Deloitte and the Audit Committee - by reference to one, and only one criterion: namely, the level of hardware revenue he generated, regardless of the identity of the hardware purchaser, the terms on which it transacted or the need for any software purchase to eventuate, let alone even be contemplated.”

The Claimants suggested that it would have been perfectly straightforward and, on the Defendants’ case, more logical - to tie Mr Sullivan’s bonus payments to some other measurable standard, such as the volume of software purchases entered into by customers who also bought discounted hardware from Autonomy; and that the reason why it was tied to hardware sales only was because in reality that was the only target he or Autonomy ever had.

Dr Lynch sought to counter this on the basis that the calibration by reference to revenue only was simply to ensure clarity. In cross-examination, Dr Lynch suggested that the bonus compensation ^{^[185]} had necessarily to be set by reference to “recognisable revenue” so as to ensure that it was measurable by reference to “a clear target”. In his written closing submissions, Dr Lynch submitted that any other criterion would have been unworkable, or likely to lead to dispute. ^{^[186]} Dr Lynch added that it was in his view “inherent” that the revenue achieved would be referable to the protection and/or promotion of the software business because otherwise Autonomy would not have agreed to participate in the revenue-raising hardware transaction in the first place.

Dr Lynch’s explanation struck me initially as plausible; and Mr Sullivan’s evidence in the US criminal proceedings was also in line with it. However, the fact remains that if protection and promotion of the software business was truly the underlying objective, the omission of any mention, still less provision, for Mr Sullivan to prioritise customer and hardware sales according to their propensity to engender software sales is notable; as is the complete absence from the correspondence of so much as a whisper of encouragement of such a priority.

Dr Lynch’s suggestion that it was “inherent” in Autonomy’s ability to decline any Dell transaction that sales would be confined to existing or strongly prospective customers was untested since he could offer no concrete example of a transaction proposed by Dell having been declined; in any event, it simply amounted on analysis to an assumption that Mr Sullivan would put aside the temptation of greater sales and a larger bonus by reference to the unstated objective. Furthermore, Dr Lynch’s suggestion of an “inherent” filter begged the question why it was not quite easily and clearly made express.

Again in passing, since this too is more conveniently elaborated when I address later the question whether the Defendants misled Deloitte and the Audit Committee, it is to be noted that Mr Hussain’s email dated 15 September 2009 promising Mr Sullivan bonuses for obtaining recognised revenue also promised a further:

“special bonus of $50,000 if you are able to extract an email that allows us to allocate the associated costs appropriately in my opinion” .

This was, the Claimants suggested, an extraordinary provision, which tied Mr Sullivan into trying to get some form of email from EMC sufficient to present their arrangements in such a way as to justify part of the sums paid to EMC as referable to marketing and expenses (and see further paragraphs 1272 to 1284 below). It is noteworthy also that the payments were to be made only after the auditors had signed off the Q3 results (then expected in the third week of October): demonstrating the concern that Deloitte had to be persuaded that the accounting allocation was correct.

In the event, the conditions were satisfied (though Autonomy restricted the first part of the bonus to $200,000 and refused Mr Sullivan’s request to pay a further pro rata amount to reflect the higher recognisable revenue in fact achieved). By email of 11 November 2009, which had been approved in draft by Dr Lynch, Mr Hussain confirmed the arrangements (though no Porsche) and put forward further bonus arrangements for Q4 2009, as follows:

“Hi Mike

Apologies for the delay in my response. Whilst the email [earlier in the thread] was clear in that there were 2 distinct triggers for the EMC deals ($30m and $50m), given the out performance in Q3 here’s what I would propose:

1. You receive $200,000 for delivering $30m of recognizable revenue in Q3

2. You receive $50,000 for extracting the necessary written confirmation for allocation of costs signed off by the auditors

3. For Q4 for newly contracted and recognized appliance related revenue from EMC, and HDS ~~and Dell~~ amounting to ~~$15m~~ $10m (excludes any HDS / MS revenues) you will receive ~~$150,000~~ $100,000 plus an additional linearly calculated commission from Q3 of $54,000 (see below for calculation). This will require the necessary confirmations for the allocation of costs and also inclusion of Autn software in the sale as we have already discussed.

4. For additional Q4 contracted and recognized revenue from ~~EMC, HDS~~ ~~and~~ Dell (excluding OEM related revenue) amounting to a further ~~$15m~~ $10m you will receive an additional ~~$150,000~~ $100,000.

The usual Autn ts and cs apply and good luck! I will get the cheque cut tomorrow on this basis

Regards Sushovan.”

That also serves to illustrate that the bonuses were by no means insubstantial: for Q3 2009, gross $250,000 ($200,000 for delivering recognisable revenue in excess of $30 million plus the ‘special bonus’ of $50,000); and a further $100,000 in Q4 2009 if revenue from hardware exceeded $10 million (excluding OEM-related revenue). (I was not provided with the details of every promise or payment, but note, for example, that for his “good job” including in relation to hardware sales in Q3 2010 he was promised $50,000 , and for Q2 2011 a further $50,000).

The following evidence in Mr Welham’s witness statement relating to the $50,000 ‘special bonus’ was not challenged:

“I was not aware of the existence of this bonus at the time of our audit and review work. I have never known of a situation in which an employee of one of our audit clients has been offered a financial incentive to obtain documentation to support representations being made by management to the auditor as to what should be the appropriate accounting treatment. Had we been made aware of this bonus arrangement at the time, we would have asked questions of management in relation to the rationale for the bonus.”

As a consequence, it would not have been raised by Deloitte with the Audit Committee and, accordingly, the Audit Committee would have been unaware too.

In summary, the structure and calibration of Mr Sullivan’s incentive arrangements were, at the least, entirely consistent with the Claimants’ case, even if not in isolation probative of it. They add to the indications that the hardware reselling strategy was carried on without ostensible regard to any measured benefit to the software business. The fact (as I find it to be) that the incentive payments were never revealed to Deloitte or the Audit Committee reinforces that view .

(8) No documentary evidence that discounted resales of hardware were used as a bargaining chip

It may be recalled that an important element of the Defendants’ justification for the hardware reselling strategy was the evidence of Mr Sullivan in the US criminal proceedings, and of Mr Egan in those proceedings and in this action, to the effect that they always understood that the principal purpose of the resale of discounted hardware was (as Mr Keker, Mr Hussain’s counsel in the US criminal proceedings, put it when cross-examining Mr Sullivan) to build “a bigger share of the financial market and use discounted hardware as a means to that end” (see paragraphs 726 and 743(1) above).

The Claimants relied on the documentary material in evidence to counter the Defendants’ case based on that evidence (whilst, of course, maintaining the argument, which I have in the event accepted, that the understanding of both Mr Sullivan and Mr Egan was really based on what they were told by the Defendants). They submitted that there was nothing in the documentary evidence to suggest that either the marketing department or even those in Autonomy’s sales team (apart from Mr Sullivan, and to some extent, Mr Egan) had any knowledge of the hardware sales; and that in any event, there is no contemporaneous documentary evidence to suggest that the discounted resales were ever used as leverage in promoting and negotiating the terms of software sales.

Further, in at least some of the documented arrangements (see the Zones transactions below) the contractual arrangements were structured to seek to ensure that the end-user (a long-standing customer of Dell) did not and never could find out about Autonomy’s subsidy and involvement.

The Claimants submitted that this made a mockery of the purported justification of the programme as a loss-leader for marketing purposes, as does the basic fact that (a) hardware reselling targets were fixed by reference to market expectations of overall quarterly revenues and (b) hardware reselling was, within the overall target, increased or decreased by reference to Autonomy’s success or failure in selling software offerings in that quarter.

They elaborated on this in submitting that had sales of hardware really been designed to drive software sales, matters would have looked quite different, and there would have been (as Mr Rabinowitz put it in his oral closing) “if not mountains of documents, at least some pile of documents, at least an anthill’s worth of documents” to show (for example) continuing cost/benefit analysis, consultation with the salesforce as to the impact of offering discounts, selection of customers according to likely requirement for hardware and software, and some form of input from the marketing department; but there was “no such thing”.

As the Claimants noted in their written closing submissions, there is not a single email or document in evidence passing either to or from Ms Nicole Eagan, Autonomy’s Chief Marketing Officer, to suggest that either she or anyone else reporting to her had any knowledge of, or involvement in, this supposedly important and cost effective marketing strategy. Ms Eagan was not called as a witness either and the witness statement she had made (giving a London address) was withdrawn.^{^[187]} In the recitation of the chronology, I noted Ms Eagan’s involvement only once in this context: this was that she was one of the two recipients (Mr Egan being the other) of Dr Lynch’s email on 1 May 2011 stating:

“Folks, you two must be focused on this Q revenue now and NOTHING else” (see also paragraph 1105 above).

Mr Egan’s evidence in this context was that although he was aware of the hardware reselling strategy, even he generally did not know which companies had purchased hardware from Autonomy or how much third-party hardware was being sold. In a passage of his witness statement from which I have previously quoted in part in paragraph 766 above, he illustrated this by reference to a particular transaction with Bank of America, as follows:

“I was involved in making this type of hardware sale when Autonomy first began this practice in Q2 2009, but was involved only occasionally thereafter. To my knowledge, Mike Sullivan, the CEO of Autonomy Zantaz, was tasked by Dr Lynch and Mr Hussain with managing most of the hardware reselling. I was more tasked with selling software licences. As a general matter, I did not know which companies had purchased hardware from Autonomy or how much third-party hardware was being sold. When I was involved in selling software licences to end-users, I usually did not know whether hardware had been sold to that end-user and when I was not involved in, or aware of, a hardware sale I did not use the fact of a prior or contemporaneous hardware sale to promote the software sale that I was trying to make. For example, I discuss below a very large software licensing and data hosting transaction with Bank of America that I was closely involved in and which we eventually executed in Q1 2011. I did not know at the time, but have since been informed, that in 2010 Autonomy resold, at a loss, more than $30 million of Dell hardware to a reseller called SHI whom in turn, on-sold the hardware to Bank of America. Had I known about this at the time, it might well have been useful information in our discussions with Bank of America. However, I did not know about it, and I never heard Dr Lynch or Mr Hussain refer to these hardware sales during our extensive negotiations with Bank of America.”

(9) Three illustrative transactions where marketing was no part of the purpose

The Claimants put forward three examples to illustrate and support their proposition that the hardware reselling strategy cannot really have been conceived of as a marketing programme because it is the essence of a loss leader marketing strategy that the purchaser should know who covered the loss, and Autonomy’s salesmen did not have that knowledge, and certainly did not deploy it as a ‘bargaining chip’.

The examples were (1) a transaction with BofA (2) a transaction with Citibank and (3) various transactions via Zones Inc (to H&R Block, Target and Oracle).

BofA

Mr Egan’s evidence in relation to the Bank of America (which was not directly challenged in cross-examination) was corroborated by the evidence of Mr Reagan Smith, who ran BofA’s procurement team for software purchases. Mr Reagan Smith did not give evidence before me, but his evidence in the US criminal proceedings was admitted into these proceedings by a hearsay notice issued by the Claimants. That evidence was to the effect that he was not aware of any hardware sales by Autonomy to BofA (or to SHI) and that, in the course of the negotiations with Autonomy, neither Mr Egan nor Mr Scott (who he said were the individuals principally involved for Autonomy) or Mr Hussain or Mr Krakoski (who had responsibility for sales to BofA) or anyone else raised the hardware sales as a selling point in any way.

More generally, the Claimants submitted that:

(1) The documentary record disclosed no evidence of a reference to a prior sale of hardware in any sales pitch made to any software company.

(2) If a marketing strategy had existed, senior management within Autonomy would have ensured that, if nothing else, the hardware deals were widely publicised within the software sales team so that they could be used to secure software sales.

Against this, Dr Lynch told me in cross-examination that Mr Egan did in fact know that Autonomy “had a history of selling hardware to Bank of America and that’s covered in some documents” , and referred in support to an email from Mr Egan to Mr Hussain dated 23 December 2009 in which Mr Egan expressly referred to the Morgan Stanley Dell reseller deal (for hardware) being done “on a lost [sic] leader basis by Autonomy” .

Dr Lynch also relied on his own evidence (in his supplemental witness statement) to the effect that he had “several private conversations” with another more senior director of BofA, Mr Simon Mackenzie-Smith (“Mr Mackenzie-Smith”, then head of BofA in London) when he:

This issue developed some importance, both as a microcosm of the larger disparity in the context of the hardware sales between Dr Lynch’s oral evidence and the documentation, and in terms of credibility.

The Claimants urged me to reject Dr Lynch’s evidence that he had sought to deploy the discounted hardware sales in support of the software sales: it was unsupported by documentation or by any witness evidence from Mr Mackenzie-Smith, had not been referred to in any communication in the period and had not been mentioned by Dr Lynch in his first witness statement. In his first witness statement, Dr Lynch had mentioned that he had spoken to Mr Mackenzie-Smith but had made no mention at all of any hardware deals, still less their use as leverage:

“ In Q4 2010, I spoke with high-level contacts at Bank of America (namely Simon MacKenzie-Smith, who I believe was the head of Bank of America in London, and Mr Thomas Fakhouri) to seek their assistance in encouraging the bank ’ s procurement team to prioritise the direct deal, in the hopes of closing the deal within the quarter ”.

In cross-examining Dr Lynch, Mr Rabinowitz deployed as his anvil the disparity between his first and second witness statements, and also (a) that Mr Egan had claimed that in late 2010, when he was negotiating the software deal with BofA’s Mr Reagan Smith, neither he nor (as far as he was aware) Mr Reagan Smith had any awareness that Autonomy had sold discounted hardware to Bank of America ^{^[188]} , and so obviously he had never thought to use that as leverage, and (b) more specifically, that not only did Mr Egan say that he did not know of the hardware sales to Bank of America, but also that he had “never heard Dr Lynch or Mr Hussain…refer to these hardware sales during our extensive negotiations with Bank of America.”

Mr Rabinowitz took Dr Lynch through a series of emails between Dr Lynch and Mr Mackenzie-Smith at this time which demonstrated that Dr Lynch was trying to enlist Mr Mackenzie-Smith to cut through some bureaucratic impediments which were delaying the software transaction (which involved a hosting restructuring) but in which there was no mention or reference to any parallel telephone conversations. Mr Rabinowitz repeatedly put to Dr Lynch that if there had been parallel telephone conversations between them in which Dr Lynch had sought to use the hardware sales as further leverage, there would surely have been some reference to them in that chain of email correspondence.

Dr Lynch displayed uncharacteristic irritation in answering these questions, and sought to dismiss the suggestion that he had no such discussions as “unrealistic”, on the basis that in a one-to-one conversation between people as senior as they were it was inherently likely that Dr Lynch would have sought to deploy such an argument, and correspondingly unlikely that either would have recorded the private discussion in any formal record or note.

In the absence of any evidence (oral or written) from Mr Mackenzie-Smith, I have had to reach a view on the basis of (a) my appreciation of Dr Lynch’s oral evidence, (b) the points made by the Claimants about the lack of any reference to these conversations in the first round of evidence and (c) the content and tone of the email exchanges.

I have concluded that in the general way of things it is more likely than not that Dr Lynch made some passing remarks about the previous relationship and discounted sales; but not in a way that caused Mr Mackenzie-Smith to treat it as a sufficient consideration to influence the decision or mention the matter to BofA’s negotiator Mr Reagan Smith.

In a sense, Dr Lynch’s version of the discussions in his second witness statement tells against his case that the hardware sales were a potent and successful marketing tool. The fact is that whatever the conversations were, they do not appear to have accelerated the proposed software sales to BofA: no deal was concluded with BofA in Q4 2010 (which is what Dr Lynch had been pressing for) and Autonomy, desperate for the revenue, had instead to restructure the deal, split it into smaller parts, and parcel it out and effect a sale to two VARs (DiscoverTech and Capax Discovery) from which it purported to recognise revenue, as I explain in the Schedule to this judgment on the impugned VAR transactions. Dr Lynch’s explanation did not answer the real point that the sales were negotiated by and between Mr Egan and Mr Reagan Smith, and the latter’s evidence, which I accept, was that he did not even know of the hardware sales.

In all these circumstances, I accept the Claimants’ submission and find that the hardware deals with BofA played no material part in relation to BofA’s agreement to enter into the software deal. The Claimants further submitted, and again I accept, that this seriously undermines any suggestion on the part of the Defendants that the purpose of those hardware sales was primarily marketing, to drive (in some way) separate sales of software.

Citibank

In the case of the Insight/Citibank transaction, which was a $10 million transaction for the supply of Dell hardware, the Claimants drew particular attention to commentary in a spreadsheet, prepared by Mr Sullivan and attached to an email dated 11 February 2011 from him to Mr Hussain which notes:

“Dell in competitive bid for business to be awarded in Feb. Purchase in March. Dell is incumbant [sic] and gives themselves 60-70% chance to win. We would need to put this through an existing reseller (Insight) who inventories so Citi may not know we are involved. Not clear if this revenue would all hit in Q1.”

The Claimants pointed out (and in cross-examination, Dr Lynch accepted) there is no indication anywhere in his comments or elsewhere that this was the sale of an appliance or a package that contained Autonomy software, or that Mr Sullivan had any interest in some related sale of software: his eyes appear firmly fixed on the revenue target and its timing.

What the Claimants suggested was particularly telling was that it appears that Mr Sullivan anticipated that Citibank might not even know of Autonomy’s involvement. Dr Lynch recognised this when cross-examined; and although he told me that he “wouldn’t be happy with that, unless we could tell Citi…I wouldn’t accept the order if I had known…” , the Claimants understandably pressed the point that whatever Dr Lynch might now choose to say, the fact remained that, at the time no-one, including Mr Hussain, suggested that the deal should not be accepted or concluded unless Autonomy’s involvement was disclosed to Citibank. I accept that.

Zones Inc

As to the various deals through Zones (a substantial Dell reseller and “IT solutions provider” ), the Claimants’ principal point was that, as they read it, each of the letter agreements which governed Autonomy’s relationship with Zones provided that its terms should be confidential and that no party should “disclose the terms hereof without the consent of the other party”.

Furthermore , the Claimants relied on each of the three deals with Zones as illustrating their case that Autonomy had no role apart from injecting or insinuating itself into arrangements between Zones and Dell so as to establish a contract from which it could derive recognised revenue. Thus, in each case and in the context of an existing relationship between Dell and Zones:

(1) The arrangement was that Zones would procure Dell hardware which Zones needed for supply to its own customers not (as would have been normal) from Dell but from Autonomy;

(2) The advantage offered to Zones for this process of what was described in an email dated 20 May 2010 from Mr Dean Bordeaux (“Mr Bordeaux”, an account executive with Dell) to Zones (including Mr Camden, of whom see below) as “injecting a new reseller between Dell and Zones” was a discount on the cost to it of the hardware: the discount resulted from Autonomy purchasing Dell hardware from Dell at its undiscounted price and then selling it on to Zones at a discounted price;

(3) Autonomy would account for the purchase from Dell as a sale giving rise to recognised revenue, but its involvement would be minimal in that the hardware in question would not be delivered physically to Autonomy and then by Autonomy to Zones or Zones’ customer, nor would it be loaded with Autonomy software: instead, it would be delivered by Dell to Zones or Zones’ customer;

(4) The first of the three Zones deals was a large deal of over $7 million with a tax preparation services company called H&R Block. As Dr Lynch accepted, the hardware was not going to be loaded with any Autonomy software: it was ‘pure hardware’ according to the Claimants’ definition. The evidence principally relied on by the Claimants was in the form of a transcript of the evidence in the US criminal proceedings given by the head of Zones’ sales organisation, Mr Dominic Camden (“Mr Camden”). This was introduced into evidence by hearsay notice served by the Claimants. The Claimants also relied on Mr Bordeaux’s contemporaneous emails.

Mr Camden, who did not provide evidence in these proceedings and whose hearsay evidence could not, therefore, be tested, stated in his evidence in the US criminal proceedings that whilst the discount was attractive, he had concerns about the introduction of Autonomy into this deal with a long-time customer (as was H&R Block), lest Zones should lose a client if the client found out that it could secure the solution it wanted for less money from Autonomy.

He explained that it was for that reason that he insisted on there being both non-solicitation and confidentiality clauses in the contract with Autonomy. He regarded that as very important. A clause was negotiated to address this and was included, as follows:

“The terms of this Letter shall be confidential. No party hereto shall disclose the terms hereof without the consent of the other party. For a period of one (1) year from the date Autonomy receives the most recent Customer PO [i.e. Purchase Order] hereunder, Autonomy shall not actively solicit H&R for purposes of reselling the Products directly to H&R Block. For avoidance of doubt, nothing herein shall prelude Autonomy from selling to H&R Block any Autonomy products, in any manner whatsoever.”

Mr Camden said in his evidence in the US criminal proceedings that he took from this that Autonomy would have no contact with H&R Block. This reflected and was reinforced by an assurance Mr Bordeaux had given in his email of 20 May 2010 quoted from earlier that:

“Autonomy has and will have no direct contact with Block but does need the attached simple agreement executed with Zones.”

That “simple agreement” was substantively like the agreement eventually signed. In the event, it seems that the confidentiality of Autonomy’s involvement was preserved: Mr John Meiers (“Mr Meiers”), who was in charge of H&R Block’s software procurement team in 2010, gave a witness statement in these proceedings (which was not challenged) in which he said that when in September 2010 he negotiated a $2 million software deal with Autonomy, he was unaware that, just three months earlier, Autonomy had been involved in the resale transaction. Mr Meiers added that to the best of his knowledge, no-one else at H&R Block involved in the software purchase was aware of Autonomy’s involvement either. The Defendants chose not to cross-examine Mr Meiers.

The Claimants posed the question why in such circumstances Autonomy wanted to be involved at all. Understandably, the Zones management team was puzzled at being offered the opportunity to sell its own longstanding customer cheap hardware, subsidised by Autonomy, with the customer kept in the dark about Autonomy’s role.

In this context, Mr Camden sent an internal email to the Zones team on 7 July 2010 offering a theory as to why Autonomy should be prepared to act in that way. He speculated that “Dell must be an Autonomy client” which received preferential discounting and pricing for business driven through Autonomy. His Zones colleague replied the same day, observing that this “will go in the books of unresolved mysteries”. The Claimants contended that:

“These contemporaneous exchanges merely confirm what is obvious; that it made no apparent commercial sense for Autonomy secretly to subsidise Dell’s sale to a hardware reseller and the reseller’s sale to its customer. These Zones’ employees could not have known that the Defendants had a very good reason for having Autonomy engage in this activity, namely facilitating the ability of Autonomy to report the revenues thereby achieved, without disclosing their source; and in this way, misleading the market .”

The same modus operandi was adopted to subsidise Dell’s sales to Zones for other Zones customers. Thus, a quotation dated June 2010 issued by Zones for the sale of hardware to Oracle, the large software company, contained no reference at all to Autonomy. The deal between Autonomy and Zones, again, was that Oracle should not be notified of Autonomy’s involvement in the deal.

The same was true of the third of the Zones deals, that in late 2010 for the sale of Dell hardware to Zones for resale to yet another Zones customer, Target Corporation (“Target”), a US chain of retail stores. The letter agreement between Autonomy and Zones dated 11 November 2010 contained, at clause 10, a confidentiality provision preventing disclosure of its terms to anyone, which would include Target itself.

The Claimants described and relied on the third Zones transaction as follows:

(1) In an email exchange on 11 November 2010 (the same date as the letter agreement) between representatives of Dell, Zones and Autonomy (Mr Sullivan), Mr John Niemier of Zones sought clarification that the way things would work was that Autonomy would provide a quote to Zones, and Zones would in turn deliver the final quote to Target. On 12 November 2010, Dell ’s Mr Randall Johnsen responded agreeing that “Zones is the only entity that should provide Target a quote for this transaction”: in other words, Autonomy was not to deal directly with Target.

(2) In parallel with this exchange of emails, there was an internal email discussion at Zones between Mr Niemier and Mr Camden. In an email dated 11 November 2010, Mr Niemier expressed confusion as to how the process would work. Mr Camden ’s response was clear: “Zones places PO with Autonomy, yes. Autonomy places order with Dell. Autonomy and Target don ’t touch”.

(3) When Mr Niemier asked who would issue the quote to Target, Mr Camden ’s response was “Zones”, in the same way as for H&R Block. He added, “HRB [i.e. H&R Block] never knew about Autonomy and neither should Target”.

(4) Mr Camden confirmed during his testimony in the US criminal proceedings that this email reflected his understanding of how the deals with Autonomy were set up.

(5) On 22 November 2010, Dell ’s Mr Johnsen contacted Mr Camden about the Target deal. He reminded Mr Camden, “not to use Autonomy name in messaging”, i.e. in the messaging Zones sent to Target about the hardware sale.

(6) Consistent with this, a Mr Tom Corley of Zones sent an email on 30 November 2010 to Mr Camden confirming that Autonomy had received the Zones purchase order for Target. Mr Corley noted that the delivery would be a

“ blind drop-ship from Dell directly to Target. So there should be nothing that says Autonomy on any packing slip ”.

(7) Mr Camden testified in the US criminal proceedings that a “blind drop-ship” meant that the hardware would be shipped directly from Dell to Target, rather than coming via Zones.

(8) The Claimants submitted that the significance of the lack of any reference to Autonomy, even on the packing slip received by Target, is plain: Autonomy was to get revenue, but no customer relationship.

The Claimants submitted that these Zones deals were an illustration of the true nature and purpose of the overall strategy: it was plain and obvious that there was no question of Autonomy using cheap hardware sales to H&R Block, Oracle or Target, as a means of driving software sales to those companies, in circumstances where the companies in question had no idea, and were contractually forbidden from knowing, that Autonomy was responsible for subsidising their purchases of hardware.

They submitted more generally that:

“ at no time during the Relevant Period was any attempt whatsoever made to identify or monitor the extent to which these hardware sales produced any marketing benefit to Autonomy. That, again, is consistent with the real purpose of the hardware reselling strategy being simply to, in effect, ‘ buy ’ (at a substantial cost) recognisable revenue that would be included in the revenue figures reported to the market without revealing the true source (or cost) of this additional revenue stream. It is inconceivable that Dr Lynch and Mr Hussain, intelligent individuals, could have considered that what they were doing was honest.”

The Defendants, however, submitted that:

(1) Even in the three Zones deals the only restriction as a matter of construction of the letter agreement was that Autonomy was required to keep the terms of the letter agreement confidential and did not extend to prohibiting Autonomy from saying that it was involved as the party supplying hardware to Zones, or that this was part of a marketing programme;

(2) There was no evidence that either of the Defendants was made aware of the contractual restriction;

(3) Dr Lynch stressed that (a) he had no involvement in the Zones/H&R Block transaction, and he was not aware at the time of any contractual term preventing Autonomy from disclosing its role in the transaction (and it was not suggested to him in cross-examination that he did) and he “wouldn’t have been happy with the lawyers if I’d known about that”; (b) the position was likewise as regards the sale via Zones to Target; and (c) he was also unaware of any sales of hardware via Zones to Oracle, and thus was not aware of any such restriction: but he added that he did not see how a term could have been agreed with Zones precluding Autonomy from soliciting business from Oracle given that Oracle were an existing customer;

(4) The Zones sales, even if they did not assist to reinforce a customer relationship, achieved the other objective of the strategy as (according to Dr Lynch) formulated at the Loudham Hall meeting of reinforcing relationships with the hardware suppliers themselves, and in particular with Dell (see above).

The Defendants also drew my attention to other parts of Mr Camden’s testimony in the US criminal proceedings in which he accepted that (a) there was a customer demand for “complete solutions” (where a seller of software would act as a “one-stop shop” and also provide the hardware) and (b) if one of Autonomy’s customers wanted a package or complete solution, it would make good sense to provide the hardware that those customers needed as well as the software. Mr Camden accepted also that Zones itself often planned, and it was perfectly standard business practice, to use a relationship (such as Zones’ with Target) established initially by sales of hardware to sell them other products like software; and that it made sense likewise for Autonomy to sell hardware to a customer with a view to using that relationship to introduce their own software products.

In any event, the Defendants submitted that the Claimants’ focus on the Zones deals flagged that they were the exception not the rule, and that for the large majority of hardware transactions the customer was aware of Autonomy’s involvement and sales were deployed to promote the customer relationship. Mr Sullivan acknowledged in his evidence in the US criminal proceedings that “absolutely” he was instructed to and did tell customers that the sales were “part of a marketing programme” , and he did not consider that to be wrong in any way.

Further, even though I would accept the Claimants’ more general point that most Autonomy software salesmen probably did not know of the hardware reselling strategy and did not therefore consciously use it as leverage for software sales (and see above as to my impression that the fact of the programme was known only to a small circle within Autonomy), the fact is that most of the hardware deals in question were undertaken by Mr Sullivan, who plainly did know all about the programme. It seems to me likely, consistently with Mr Sullivan’s evidence, that he would have felt some need to explain, in deals where he had some contact with the purchaser, what the rationale for offering a discount was; but all he said was that it was “part of a marketing programme”. That seems to me to be little if anything more than a statement of the patently obvious. There was, however, no suggestion in his evidence or any other evidence (apart from Dr Lynch’s discussions with Mr Mackenzie-Smith) that he ever said anything more.

In any event, the Defendants’ broader point was that there is evidence that both Mr Egan and Mr Sullivan did think that the hardware reselling strategy was a catalyst for software customer loyalty and software sales. Mr Egan’s evidence is of particular interest in the broader context of the utility of the hardware reselling strategy as a marketing strategy.

I have previously quoted his witness statement in relation to this (see paragraph 754 above). As usual in his case, Mr Egan’s evidence was carefully modulated, seeking to find interstices as best he could in the heavily lawyered agreed evidence he had given in the US criminal proceedings (from which he could not depart on pain of imprisonment) when it appeared or was presented as at odds with other evidence. But subject to the caveat that in his perception “a principal reason” for the hardware sales “was to generate income” , he accepted that there were other substantial reasons for it, and in particular, agreed with his cross-examiner’s recital of propositions dressed as questions and designed to elicit the answers that he gave that “often” hardware and software sales “were leveraged together” and the hardware reselling strategy “increased Autonomy’s standing with the big financial institutions that bought their software” (and see also paragraph 751 above).

The problem for the Defendants, as it seems to me, has always been that it is not sufficient for them to demonstrate that at least in some, perhaps many, cases, the fact that Autonomy was providing or subsidising the discount on hardware sold to customers, including software customers, encouraged goodwill and warm feelings about Autonomy which might in due course translate into future sales, including software sales. The case they have always had to meet is that even if that was a benefit it was never the real reason or driving purpose, as demonstrated by the fact that Autonomy made at least some substantial sales for which even that benefit would not be obtained, because the purchaser had no idea of Autonomy’s involvement: the Zones deals being the clearest examples, it being obvious in those transactions that it was no part of their purpose to promote or protect the prospect of software sales to Zones.

In Mr Hussain’s written closing submissions, it was stated that the above three deals “were the exception and not the rule.” The additional contention was advanced that “in any event, these sales achieved the objective of building relationships with the hardware suppliers themselves”.

The latter argument was given greater prominence by both Defendants as the case developed. It was, to my mind, paper-thin, even taking into account the possibility that Autonomy may have got discounts on its purchases of hardware for its own use (for archive hosting and the like). There was little or no evidence of any special relationship being developed, nor that such a relationship promoted software sales. (It may have assisted OEM ventures, but that is a different matter and would not justify the accounting treatment of the costs.) Autonomy’s supposed ‘partnership’ with EMC was extinguished with minimal notice. Dell was in it because it is not often that a third-party agrees to subsidise a company’s sales. But I was not persuaded that opportunism metamorphosed into a relationship of such real benefit to the software business as to be a substantial purpose of Autonomy’s hardware sales.

As to whether the three sets of transactions addressed above were exceptions or outliers, the express contractual prohibition on Autonomy being involved or asserting involvement at all was unusual. But it seems to me that the question in relation to the Zones transactions (and the others singled out by the Claimants described above) is ultimately whether or to what extent there can be discerned or established from them what the driving purpose and objective of the hardware sales really was, and what credence can be given to the Defendants’ assertions that the sales were driven by the interests of, and to enhance, Autonomy’s software business.

In my judgment:

(1) The Zones transactions, and those with BofA and Citibank (another example being the transaction with Progressive), give substantial general support to the Claimants’ case that the real purpose of the hardware reselling strategy was revenue generation: they demonstrate clearly that any marketing benefits in terms of Autonomy’s software customers might be desirable by-products but were certainly dispensable, and were not what was driving hardware reselling.

(2) It may be that Mr Egan’s qualification in responding “often, yes” to the proposition put to him in cross-examination in the US criminal proceedings that “hardware and software were leveraged together with these big customers” was intended to accommodate cases such as Zones where no such leverage was possible at all. But neither Mr Egan nor anyone else ever provided any evidence, apart from their say-so, of any co-ordination, let alone leveraging, beyond some attempt to aim at mutual customers (and the evidence for even that is scanty). I suspect that Mr Egan satisfied himself, in his interstitial efforts, that he could agree to the proposition, and seek thereby to put the strategy and his own involvement in a kinder light, on the basis that in certain cases the discount might generally be supposed to have oiled the wheels of the relationship with software users. But Mr Egan had no reason to believe that there was any greater nexus than that; and it seems to me likely (and I find) that he knew that whilst any benefit which might loosely be termed marketing was readily dispensable, recognisable revenue, being the real objective, was not.

(3) The real importance of the Zones and the other transactions, in which no marketing benefit was possible, was that they provided proof of what really mattered to Autonomy and the Defendants: revenue irrespective of benefit.

(10) The consistent pattern of concealing the hardware sales

The second limb of the Claimants’ hardware case (as pleaded in paragraph 54A of the RRAPoC) was that in its published information Autonomy actively concealed its ‘pure’ hardware sales. They relied on that not only as a facet of the improper purpose but also as further proof of it: disclosure would have undermined the purpose as alleged by the Claimants (being to drive up revenue by hardware sales and to present it as all won as part of the business of a successful ‘pure software company’); whereas if the purpose had been in any material part as the Defendants contended it was, there would have been no reason not to disclose it to the market.

As one of the Deloitte Reviewers (Mr Robertson) queried rhetorically when the issue of disclosure arose in respect of the Q3 2009 Report:

“What ’s the sensitivity about being more transparent on this score? If it ’s a strong strategic move for them, why wouldn ’t they want to explain this?”

There is no doubt, and the Defendants did not deny, that they and Autonomy’s management consistently and insistently declined to disclose any more than they were advised by Deloitte was mandatory under the Accountancy Standards and thus a pre-condition of approval of the Defendants’ accounts and financial statements.

I shall discuss later the Defendants’ reasons (based on “commercial sensitivities” ) for restricting disclosure and their defence that whilst they did no more, they did no less, than Deloitte advised was required and cannot be said to have been dishonest in following their advice. In this section, I summarise what (if any) disclosure relevant to hardware revenues was made in Autonomy’s Yearly and Quarterly Results in the Relevant Period. (I also indicate the Defendants’ explanations, as appropriate.)

Annual Reports: overview

As previously mentioned (see, for example, paragraph 654 above), Autonomy presented itself to the market generally, and to HP during the pre-acquisition discussions and due diligence, as a “pure software company”. The Defendants maintained that this was to distinguish itself from software companies deriving a significant part of their revenue from services and installation. The Claimants maintained that it was intended to and did connote a company whose revenues were almost exclusively derived from the sale of its IDOL software and related software (with some small amounts of ‘appliance sales’ as later described), and that is the basis on which they acquired it.

The Business Overview section of its 2009 Annual Report provides an example of the use of the phrase to distinguish Autonomy from other software companies. The following was stated under the heading “ Financial Model ”:

“Autonomy is one of the very rare examples of a pure software model. Many software companies have a large percentage of revenues that stem from professional services, because they have to do a lot of customisation work on the product for every single implementation. In contrast, Autonomy ships a standard product that requires little tailoring, with the necessary implementation work carried out by approved partners such as IBM Global Services, Accenture and others. This means that after the cost base has been covered, for every extra dollar of revenue that comes in significant benefits can fall straight through to the bottom line. What this offers is a business model with a proven record of strong operating leverage and that is expected to continue to deliver industry leading operating margins and revenue to cash conversion.”

Materially identical wording appeared in the 2010 Annual Report .

The Claimants accepted that it may be true that, in these passages, Autonomy was seeking to emphasise its distinctiveness from companies that made significant revenues from the provision of services. However, their case is that the reasonable reader - including HP - would also and more importantly have understood the phrase to distinguish Autonomy as being unusual in carrying on no other business, and thus in its financial position being based entirely on its software business and not on any significant revenues from sales of non-software products, particularly products, such as hardware, that were burdened with substantial associated costs. They submitted that this is what the Defendants must have known and intended in emphasising that aspect of Autonomy’s business. The reason for such emphasis and its importance was the same as that which the Defendants expressly gave for emphasising how little of its revenue came from services: in the case of software, “ after the cost base has been covered, for every extra dollar of revenue that comes in significant benefits can fall straight through to the bottom line ”, while services commanded much lower margins. The Claimants contended that that is an equally fundamental difference between software and hardware sales.

The 2009 and 2010 Annual Reports both referred to a hardware-based product described as an “ appliance ” in terms of hardware with Autonomy software pre-loaded on it. The explanation in the 2009 Annual Report was as follows:

“Appliance

Currently a small part of the business focused on quick time-to-value and high return. Where customers have an urgent need to deploy IDOL, either for regulatory or commercial imperatives, we are able to provide a complete solution installable on a turnkey basis to be used in a discrete part of the customer’s business. The value of these solutions is in the high end functions they offer in a complete package, and thus the margin profile is not dissimilar to our traditional license business.”

The equivalent wording in the 2010 Annual Report was more explicit about the fact that an appliance involved hardware:

“Appliance

This is currently a small part of Autonomy’s business, focused on quick time-to-value and high return. Where customers have an urgent need to deploy IDOL, either for regulatory or commercial imperatives, we are able to provide a pre-installed licence on appropriate hardware to start generating an immediate return. The value of these solutions is attributable almost entirely to the functions offered by the licence, so although there are some hardware costs involved, the margin profile is not widely dissimilar to our traditional licence business.”

Beyond this, there was no mention in the Annual Reports issued in the Relevant Period of Autonomy earning revenues on sales of hardwa re. Thus, the Financial Review section in the 2009 Annual Report stated that the increase in revenues in 2009 “ is a combination of strong organic growth and the successful integration of Interwoven ” (a company Autonomy had acquired in early 2009). Similarly, the Financial Review section in the 2010 Annual Report ascribed Autonomy ’s reported revenue growth between 2009 and 2010 entirely to the deployment by customers of Autonomy software.

Furthermore, the accumulated yearly costs of the hardware reselling strategy which might have revealed its extent, were not disclosed either, even though in 2009 there was an increase in the amounts allocated to Sales and Marketing Expenses of $35.6 million as compared to 2008, and the hardware sales were a major, if not the sole, cause of the increase.

Instead, this increase was addressed in Autonomy’s 2009 Annual Report as follows:

“Sales and marketing expenses totalled $170.8 million in 2009, up 26% from $135.2 million in 2008. The increase in sales and marketing expenses from 2008 to 2009 was primarily due to increased advertising, additional headcount and an increase in sales commissions due to an increase in sales and a change in the geographic and size-of-transaction mix, all of which also increased with the expansion of the group in 2009. As a percentage of revenues sales and marketing expense has fallen to 23% in 2009 from 27% in 2008.”

When cross-examined about this, Dr Lynch nonetheless sought to defend the passage in the 2009 Annual Report as “ a reasonable explanation ”. He identified especially the effect of Autonomy’s acquisition of Interwoven which, he said, had been an increase not only in advertising spend, but also of the numbers employed in marketing from around 238 to around 510 people. This is difficult to square with the fact that in 2009, $35.8 million of the costs of purchasing pure hardware were treated as sales and marketing expenses, whereas the total reported increase in sales and marketing expenses was only $35.6 million, leaving no room for the suggested Interwoven factor. When pressed again, Dr Lynch added that:

“Again, this is signed off by Deloitte, they consider it reasonable, and that’s good enough for me.”

In the 2010 Annual Report, the explanation given for the increase of some 20% in sales and marketing expenses made no mention anywhere of any loss-making hardware sales or the hardware reselling strategy. The explanation then given was that:

“The increase in sales and marketing expenses from 2009 to 2010 was primarily due to increased advertising, additional headcount and an increase in sales commissions due to an increase in sales and a change in the geographic and size-of-transaction mix, all of which increased with the expansion of the group in 2010…^{^[189]}”

The Quarterly Reports followed suit and contained no disclosure of the hardware reselling strategy either. On the contrary, the Claimants contended, they revealed varying efforts to camouflage the source of revenue.

Q3 2009 Quarterly Report

Thus, although of course the hardware revenues were included in the total, the way the results were presented in the Q3 2009 Quarterly Report did not distinguish their source. This was not apparent and could not be ascertained.

Four categories of revenue were presented under the heading “Supplemental Metrics”^{^[190]} : “ Product including hosted and OEM ” ($125 million), “ Service revenues ” ($9 million) and “ Deferred revenue release ” ($58 million) which equated (after rounding) to the total revenue figure of $191.6 million as shown under “ Financial Highlights ” and under “ Revenues ”.

Furthermore, the Q3 2009 Quarterly Report noted a reduction in adjusted gross margin (86% compared to 92% in the previous year) and attributed it to the “unexpected demand for our new product programs…We do not expect this to be a trend.”

The hardware reselling strategy, which was in fact the principal drag on gross margins since hardware sales were at a discount funded by Autonomy, throwing up a loss, was not mentioned in this Quarterly Report, just as it was not in the Annual Reports (see above), except that in one cryptic sentence later relied on by the Defendants the following statement (agreed with Deloitte) was made:

“During the quarter we saw some of our large customers promote Autonomy to strategic supplier status. This has led them to adopt a broader set of our solutions in a number of significant deals.”

As Mr Knight of Deloitte pointed out at the time (and see further paragraphs 1360 to 1364 below as to the discussions between Autonomy and Deloitte which led to this formulation), this could quite naturally be read as a reference to more IDOL sales: and it certainly did not constitute transparency about the hardware sales.

Q4 2009 Quarterly Report

The Q4 2009 Quarterly Report did not mention hardware sales either. The following points may be noted:

(1) Autonomy boasted increased revenues in the quarter and for the year, ascribing its performance to “a combination of strong organic growth and the successful integration of Interwoven” .^{^[191]}

(2) Dr Lynch wrote: “We continue to see our strongest growth in the new models of the software industry such as OEM and cloud computing…”

(3) Gross margins were marginally down on Q4 2008 (89% compared to 92% adjusted, 83% compared to 89% on IFRS) but had improved since Q3 2009. The Report stated: “As previously announced, the one-time additional costs in Q3 2009 from the IDOL SPE Quick Start program were not repeated in Q4.”

(4) The Report also stated:

“ Cost base returned to traditional model after Q3 2009 product launch costs, with fixed cost base modulated by seasonal market spend and revenue-tracking sales commission.”

(5) Amongst reported highlights were “Successful launch of IDOL SPE, Arcpliance, ICE, IDOL Social Media and Interwoven product range built on IDOL”.

(6) Increased R&D expenditure was ascribed to:

“new R&D efforts associated with the acquisition of Interwoven and the one-off spend in relation to the development of new products” .

(7) Autonomy again provided “supplemental metrics to assist in understanding and analysis of Autonomy’s business” . As in the Q3 2009 Quarterly Report, hardware revenue was incorporated homogeneously into “Product including hosted and OEM”^{^[192]} and was effectively invisible as a separate source.

Q1 2010 Quarterly Report

Autonomy’s Q1 2010 Quarterly Report made no mention of hardware sales, and in the “Supplemental Metrics” revenues from such sales were now included amorphously within the category ‘IDOL Product’ . The Claimants’ specific criticisms in addition referred to the explanation of the sudden emergence of a figure for “Inventory”.

As I have explained in paragraph 990 above, the decision to defer recognition of part of the revenues from hardware sales in the quarter brought with it the need to post the amount deferred as Inventory. Autonomy had not historically held noticeable amounts in Inventory; and the increase by some $9.75 million (to a total of $10.250 million) had to be explained in the Q1 2010 Quarterly Report.^{^[193]}

In summary:

(1) A trading statement issued by Autonomy on 16 April 2010 (five days before the results were released) included an explanation of Inventory as follows:

“In Q1 the company took advantage of discounted offers to purchase stock for the Arcpliance product in advance of Q2 sales, which affected the cash position. These sales have now been completed.”

(2) The Q1 2010 Quarterly Report explained the position as follows:

“Movements in cash flow during the first quarter of 2010 of note included:…Purchasing of inventory of $10 million for Q2 2010 sales, most of which have now been completed.”

(3) The Inventory was thus presented as comprised of stock acquired with a view to sale as part of Arcpliance products.

The Claimants ’ case was that this was all false: (a) the increased inventory was not due to stock purchases: it was the balance of goods sold and delivered but in respect of which revenue recognition had been deferred; (b) it was not “stock for the Arcpliance product”; and (c) it was not stock purchased to take “advantage of discounted offers” : it was as yet unrecognised revenue for hardware sold at a loss.

The statement that the $10 million of inventory related to purchases of Arcpliance for sale to meet unexpected demand was tested by questioning in the Q1 2010 Earnings Call; as I explain below (see paragraphs 1573 to 1588 below, the Claimants contended and I have broadly accepted) that to maintain the story Dr Lynch had to lie.

As previously, the figures for “ IDOL Product ” ($47 million), “ IDOL Cloud ” ($45 million), “ Service revenues ” ($11 million), “ Deferred revenue release ” ($62 million) and “ OEM derived revenues ” ($29 million) (presented in tabular form) amounted (after rounding) to the total revenues of $194.2 million appearing under “ Financial Highlights ” and under “ First quarter 2010 Highlights ”.

One further matter which I mention in passing to set it in its chronological context, but on which also I elaborate later, is that in their Report to the Audit Committee on the Q1 2010 Review, Deloitte did confront the issue of the presentation of costs associated with hardware sales, highlighting that:

(1) The revenues for the quarter included $12.2 million of hardware sales, most of which they observed had been made at an overall loss.

(2) Management ’s rationale for the hardware sales was noted as being that they were seeking “to develop a strategic relationship with Dell. The intention is that both Autonomy and Dell will market Dell hardware that incorporates Autonomy search software.”

(3) Autonomy had sought to allocate $3.8 million of the costs to Sales and Marketing expenses, as they had allocated the costs in the previous quarter, but Deloitte had made clear that this would not be acceptable. It was explained that they had sanctioned the previous allocation on the understanding that the deals in that earlier quarter had involved Autonomy “ purchasing hardware at a price which was considerably higher than they would normally have paid in order to become the preferred hardware reseller with EMC, Dell, SHI and HDS” but they had “expected these to be more one-off in nature” and had concluded that for Q1 2010 “it would be more appropriate to reflect all the costs of hardware sold in costs of goods sold”.

(4) Further, Deloitte noted that “we have included the $3.8 million as a classification adjustment…and would not expect to see such amounts in sales and marketing in subsequent quarters.”

(5) From Q2 2010 onwards, it was only the loss sustained on the hardware sales (normally about 10% of the costs of the hardware) which was allocated to Sales and Marketing expenses, with the balance being allocated to COGS.

Q2 2010 Quarterly Report

Once again, revenue from hardware sales was not mentioned in the Q2 2010 Quarterly Report. All that was said to explain an erosion in gross margins was:

“The small variation in gross margins in Q2 2010 was in line with our expectations due to the sales mix including appliances as discussed last quarter.”

It is fair to point out, however, and the Defendants naturally emphasised, that that form of words was approved by Deloitte, after an extended negotiation with management, as described in greater detail in paragraph 869 above.

The Supplemental Metrics as provided again presented all revenue as derived from IDOL (except, arguably, those described as “OEM derived revenues” ), with no clue that any such revenues might be derived from sales other than software sales (save the reference to “ appliances” in paragraph 1212 above).

Contrary to a suggestion made by Dr Lynch when cross-examined (albeit in a different context of the Q&A scripts prepared for the quarter), there was no mention of “strategic sales” either: that shorthand (the Claimants would say euphemism) for hardware sales made its appearance in the Q4 2010 Quarterly Report, but not this one (or indeed any other Quarterly Report in the Relevant Period).

Q3 2010 Quarterly Report, Earnings Call and Investor Bulletin

The Q3 2010 Quarterly Report did not mention hardware or “ Arcpliance”, “Appliances” or “Strategic Packages” . As an echo of the explanation given in the Q2 2010 Quarterly Report for the erosion of gross margins (see above) the recovery in gross margins was very briefly referred to as being “due largely to changes in the sales mix.” There was no mention of Arcpliance, appliance sales, strategic sales, or hardware.

Again, the Supplemental Metrics showed all revenue as IDOL-derived. The figures for “ IDOL Product ” ($59 million), “ IDOL Cloud ” ($47 million), “ Service revenues ” ($10 million), “ Deferred revenue release ” ($64 million) and “ OEM derived revenues ” ($31 million) (presented in a table) equated to the total revenues of $211 million appearing under “ Financial Highlights ” and the $210.6 million appearing under “ Revenues ”.

A new Autonomy publication in this quarter (Q3 2010) was an Investor Relations Bulletin (“the Q3 2010 Bulletin”).^{^[194]} In the Q3 2010 Bulletin, there was no express mention of appliance sales, and the only mention of Arcpliance was in explaining that deferred revenue declined because of “the sell-through of the remaining hardware” . Dr Lynch accepted in cross-examination that this was a reference to the sale of hardware as part of an appliance (Arcpliance) and not a sale of ‘pure hardware’ .

The Q3 2010 Investor Relations Bulletin also represented that Autonomy’s adjusted gross margin had:

“ snapped back to 87.6% from 86.3% in Q2 ’10, as the one off effects experienced in Q2 subsided. So despite there being around $6m of hardware revenue in the mix in Q3 ’10, the gross margin is now back in the usual 88-90% range ”.

This, as written, gave the impression that Autonomy had recognised $6 million in hardware revenue in Q3 2010 ^{^[195]} . In truth, as already stated, Autonomy had recognised $26.7 million in pure hardware revenue that quarter Dr Lynch sought to explain this away as “an error. It’s referring to the inventory that’s gone through and it shouldn’t be saying - it should say hardware inventory revenue, not hardware revenue.” But on that basis the suggestion that the erosion in gross margins was simply due to inventory sales is incorrect, and the hardware sales of some $26.7 million in the quarter were not alluded to let alone disclosed.

Dr Lynch sought to argue that he was not responsible for the inaccurate description because he had not reviewed the last draft of the Q3 2010 Bulletin because “By the time it was available I was already in the television studios, so it went out without my review.” The Claimants provided a detailed examination of the contemporaneous documentation showing that this was, at best, mistaken memory and that Dr Lynch cannot have been in the TV studios at the time; and in any event, he had seen previous drafts with the same error. It seems reasonably clear, and I find, that this was another example of the efforts made to conceal the fact of substantial “pure hardware” sales.

The Q3 2010 Earnings Call took place on 19 October 2010, after the publication of the Q3 2010 Bulletin and the Q3 2010 Quarterly Report. No mention was made of either appliances or Arcpliance or “strategic sales”. Mr Khan asked a question about the impact of “hardware sales” on deferred revenue, plainly referring to the appliance sales referred to in Q2 2010: but Mr Hussain (who dealt with the question) made no attempt to clarify what sales there had been or what their effect was: still less did he reveal anything about any hardware reselling.

As in respect of previous relevant quarters, I was not shown any document or exchange between Dr Lynch and Mr Hussain relating to these contemplated sales and Mr Hussain’s projections and modelling, that suggests that any of this had anything to do with marketing.

A further demonstration of not only the efforts made by the Defendants to conceal ‘pure hardware ’ sales but also their apparent success is apparent from an email sent by Autonomy ’s Mr Ganesh Vaidyanathan to Mr Yelland some time later in the story, on 22 March 2012, which referred to “hardware orders which we source from [Dell] and sell through to our customers” (emphasis added). It was suggested to Mr Yelland in cross-examination that he must have known from the email that Autonomy was selling pure hardware. Mr Yelland explained that, even at that stage, he did not understand ‘sell-through ’ in that way.

Support for Mr Yelland ’s position can be derived from the Q3 2010 Bulletin, which, as Dr Lynch accepted, used the expression “sell through” to refer to the sale of hardware as part of an appliance, not as a sale of ‘pure hardware ’ : (emphasis added):

“There were two principle effects that led to a modest decline in the deferred revenue balance in Q3 ’10, which fell to $167.7m (from $175.5m in Q2 ’10). Firstly, the sell-through of the remaining hardware related to Arcpliance, which if excluded would mean that deferred revenue would actually have risen sequentially given the related revenue that the hardware supports.”

That explanation plainly assumes that analysts had no idea that Autonomy had sold substantial amounts of pure hardware in this (or any other) quarter. The explanation appears to have been accepted, strengthening that assumption. Had it been or been thought to be common knowledge it seems most unlikely that Dr Lynch should have remained so determined to conceal it from them.

This is of relevance to a question which is bound to occur to observers of this long saga, and has troubled me: this is whether it was common knowledge in the sector, and amongst analysts and those paid to find out about this sort of thing, that Autonomy was buying and re-selling large amounts of hardware? I have concluded that although it was known to some (including, for example, Mr Morland and Mr Khan) that Autonomy was buying and selling hardware, the assumption, which it can be inferred was fuelled by the supposition that anything material would have been disclosed, was that these were immaterial.

Q4 2010 Quarterly Report, Earnings Call and Investor Relations Bulletin

Once again, in the Q4 2010 Quarterly Report, the Supplemental Metrics made no mention of hardware revenues. They were amorphously included in the figures for “ IDOL Product ” ($84 million); and Autonomy’s other revenues were included as “ IDOL Cloud ” ($51 million), “ IDOL OEM ” ($34 million), “ Deferred Revenue Release ” ($66 million) and “ Services revenue ” ($10-11 million) (now in narrative rather than tabular form). Total revenues thus comprised amounted (after rounding) to $244.5 million.

The key point to emerge in the Q4 2010 Quarterly Report and that quarter’s edition of the Investor Relations Bulletin (“the Q4 2010 Bulletin”) was that hardware was mentioned, and so was “ Arcpliance”; and the description “package solution” also figured for the first time. The Defendants’ position as to whether this did or did not amount to disclosure of hardware sales was somewhat variable.

It may be remembered that in the presentations in the preceding quarter (Q3 2010) no mention had been made of such sales, except for a fleeting reference to “the sell-through of the remaining hardware” purchased at a discount for future Arcpliance sales to explain a decline in deferred revenue. Dr Lynch was pressed in cross-examination to explain why no mention had been made of hardware sales which he had to accept amounted to some 14% of the total revenue in this quarter (Q4 2010). His response was that (a) it was not correct to say that no mention had been made, (b) on the contrary, language agreed with Deloitte had been added, at Deloitte’s request and in a form they approved, to disclose such sales.

As to (a) in paragraph 1230 above, when asked to identify where in the Q4 2010 Quarterly Report this was the case, Dr Lynch pointed to the following passage:

As to (b) in paragraph 1230 above, and when the passage above was then criticised as inadequate and misleading, Dr Lynch remonstrated that it had been “worked out with Deloitte” who were “completely alive to this issue” and had thought it “suitable to put in and we followed Deloitte ’s directions in these matters” . He also insisted, when told by Mr Rabinowitz to “ Never mind that, let ’s just look at what you said…”, that he did not himself think it misleading, though he did emphasise that Arcpliance was simply given as an example of a wider category.

Q4 2010 Earnings Call

The Q4 2010 Earnings Call took place on 1 February 2011. Yet again, the focus of the Claimants ’ submissions was on the Q&A script. Indeed, they made no complaint about what in the event was said on the call about hardware.

Q4 2010 Investor Relations Bulletin

A few days after the Earnings Call, Autonomy produced an Investor Relations Bulletin for Q4 2010, which was published on 7 February 2011 ( “the Q4 2010 Bulletin”). The Claimants focused much attention on the drafting stages of this also; but the essential point is that in each succeeding draft, as in the Q4 2010 Bulletin as eventually published, the “Organic Growth calculation” was described as intended to be a measure of “growth in the core IDOL business” , but included revenue generated by hardware sales at a loss without any distinction or explanation.

The Q4 2010 Bulletin contained references to “package solution sales…such as Arcpliance”, similar to those in the Q4 2010 Quarterly Report . If anything, greater emphasis was given to Arcpliance as the prime exemplar, though the elaboration sought also to explain that any hardware sales with a software solution in prospect (as well as with software installed) were included.

The Q4 2010 Bulletin included the following explanation of the calculation of organic growth (the emphasis is as in the original):

“This calculation arrives at growth in the core IDOL business, whether through up front license sales, on an appliance, in our private Cloud or through our IDOL OEM partners. We remain indifferent to the means by which a customer chooses to purchase the technology, because the share of the value in all of these models is so dramatically skewed towards the software component.

The value is in the software: Arcpliance is a good example, whereby Autonomy delivers its software pre-loaded onto a suitable piece of hardware in order to dramatically cut the time to value for the customer, who is often responding to crisis situations (early case assessment in eDiscovery, for instance). Where Arcpliance is concerned, the value of the IDOL software solution and the complex processing it performs is many times higher than the cost of the hardware on it is installed.

Therefore we do not think the approach of attempting to strip out the various components of a sale - e.g. Arcpliance internal hardware costs - makes sense. We would also point out that if one wanted to perform such a calculation one would also need to strip out the hardware element from the year ago period, in order to compare apples with apples, which in this case would actually increase the organic growth.”

In fact, on the Claimants ’ calculations, if hardware revenue had been excluded, organic “ growth ” would have been negative:

(1) The Q4 2010 Bulletin stated that Q4 2010 growth was 12%. Without hardware revenues, there would have been negative growth of 5%.

(2) Similarly, FY 2010 growth was stated to be 17% in the Q4 2010 Bulletin. Removing the hardware revenues would have reduced annual growth to 7%.

(3) Accordingly, on those figures, the last sentence of the passage quoted in paragraph 1236 above appeared to be untrue: in relation to this, Mr Rabinowitz accused Dr Lynch in cross-examination of a “bare-faced lie”.

When cross-examined on these comparisons, Dr Lynch had to accept the calculations, but he made three points:

(1) For the purpose of their calculations the Claimants had stripped out what they considered to be more “pure hardware sales” whereas, as already well apparent, his position is that many such sales were properly characterised as Arcpliance and/or “ appliance” or “ appliance-type” sales;

(2) More important than the organic growth figure for understanding the momentum of sales of IDOL was earnings growth which was not affected by the calculation (because, of course, the hardware sales were not profitable); and

(3) There had been “a little bit of picking of periods here” by the Claimants: he suggested that a comparison with the more relevant preacquisition period between H1 2010 and H1 2011 would yield a very different result.

What Dr Lynch ’s answers in cross-examination in the context of the Q4 2010 Bulletin served to emphasise was the inextricable connection, on the Defendants ’ case (at least as it was developed by Dr Lynch), between the purpose of the sale and its characterisation as an appliance: and as the purpose of all the hardware sales, on the Defendants ’ case, was to drive software sales, substantially all, if not all, its hardware sales were “ Appliance” or “ Appliance-type” sales. Put another way, Dr Lynch ’s ultimate answer to Mr Rabinowitz ’s accusation that the last sentence of the passage quoted in paragraph 1236 above was a “bare-faced lie” depended on what was comprised in the phrase “the hardware element”.

A problem for the Defendants in this way of developing their case is that at one and the same time they and management also represented that ‘ appliance sales ’ comprised only a very small proportion of total sales, say 1%. The argument that the nature and extent of hardware sales were revealed by the disclosure of the fact of appliance sales was untenable; and the mere revelation of the fact of such sales took the Defendants next to nowhere.

Mr Rabinowitz also suggested to Dr Lynch that the second of Dr Lynch ’s points summarised in paragraph 1238 above was inconsistent with a passage earlier in the Investor Bulletin itself stating that “In analysing organic growth Autonomy considers organic IDOL growth to be the most meaningful performance metric for understanding the momentum in the business”. Dr Lynch ’s answer to this was that the sentence was limited to stating that for the purpose of analysing organic growth the important factor was organic IDOL growth and did not affect the fact that, in his view, the most important metric in valuing Autonomy was not organic growth, but earnings growth (EPS) and/or free cash flow growth.

I agree that the metrics are different. I also agree, of course, that it is a tenable view that the more important valuation metric is earnings growth, and EPS. Further, though Dr Lynch accepted that growth driven by the sale of ‘ pure ’ hardware was not a meaningful metric for understanding the momentum of IDOL sales, and it seems to me it would follow that its inclusion would thus have skewed the presentation of IDOL organic growth, he did not accept that any of the sales were ‘ pure’ hardware sales. He argued that the Claimants ’ calculation, and its utility, depends on accepting their definitions. Again, at the heart of this was Dr Lynch ’s elastic definition of “the hardware element”. What was said entirely altered in significance according to the content of that phrase. In my view, all this was an exercise in obfuscation on the part of Dr Lynch. What he needed to do, and what this part of the Investor Bulletin did, was to count all hardware revenue in for one purpose, but remove only part of it for another, in order to maintain the picture of growth without material past adverse effect, and the prospect of future improvement, in gross margin.

Mr Miles took up in re-examination the last of Dr Lynch ’s points as summarised in paragraph 1238 above. Mr Miles put to Dr Lynch two different comparisons. These, as Dr Lynch had suggested in cross-examination, were based on information extracted from the interim results for the six months ended 30 June 2011.

The comparisons Mr Miles put were (a) between Q2 2010 and Q2 2011 and (b) between H1 2010 and H1 2011. Those comparisons, as Dr Lynch happily agreed, showed that if what the Claimants termed “pure hardware” was stripped out from both, the effect was actually to increase the organic growth rate from one period to the other. But this did not take Dr Lynch very far: the exclusion of "pure hardwar e” in both periods might show comparative organic growth, but at the price of reducing the revenue growth and ruining the whole picture which hardware sales were intended to paint.

In any event, even accepting (to test the argument) every part of Dr Lynch ’s three-part rebuttal summarised in paragraph 1238 above, the fact remains that if what was termed in the passage quoted “the hardware element” had been stripped out, Autonomy ’s revenue would have been materially reduced. That of itself would almost certainly have invited questions, which would have revealed the programme. Further, depending on exactly which sales were treated as the “hardware element” the comparison between the two periods selected by the Claimants would depend on the relative amounts of such sales in each period; but I suspect that, whatever the content given to the “hardware element” the claim that “organic growth” would have increased is a shaky one. And if, as at one moment in his cross-examination Dr Lynch suggested, all sales with a hardware content except Arcpliance sales were excluded, the statement made would have been false.

More generally, I cannot avoid the conclusion that the part of the Investor Bulletin with reference to organic growth quoted in paragraph 1236 above omitted any reference to hardware sales, other than Arcpliance or appliance sales, not because of the difficulty of comparing apples with apples, but as part of the continuing determination of the Defendants not to disclose to the market that a material part of Autonomy ’s revenue and IDOL “organic growth” was derived from such sales.

The Claimants ’ contentions in that respect are fortified by another passage in the same Bulletin which again presented the variability in gross margins as being due to appliance sales ( “ pre-packaged, turnkey solutions ”), with no reference being made to the substantial ‘ pure ’ hardware sales that Autonomy had made both in this and previous quarters. Under the heading “ Profitability analysis ”, the following passage appeared (emphasis added):

“Other factors to consider: There are a number of other factors that affect the gross margin. During the year Autonomy has also succeeded in addressing the urgent needs of a small number of customers with pre-packaged, turnkey solutions, constructed of services, hardware and software, of which Arcpliance is an example. The gross margin for such solutions is lower than for pure software. As a result, over the last few years the gross margin has fluctuated anywhere between around 85-92% without an easily discernible pattern. Autonomy has indicated that it expects the gross margin to remain within the range of 85-90% for the foreseeable future.”

Lastly in respect of Q4 2010, and to repeat my refrain from previous quarters: I was not shown any document or exchange between Dr Lynch and Mr Hussain relating to these contemplated sales and Mr Hussain’s projections and modelling that suggests that any of this had anything to do with marketing.

Q1 2011 Quarterly Report

I did not understand the Claimants to take any specific point or make any specific criticism of the Q1 2011 Quarterly Report (released on 21 April 2011) (nor, for that matter, of the Q1 2011 Q&A or the Earnings Call for that quarter), apart of course from the criticism that the hardware reselling strategy was once again not properly disclosed.

However, it is noticeable that the Q1 2011 Quarterly Report did not repeat the sentence which Dr Lynch stated Deloitte had insisted be included in the Q4 2010 Quarterly Report and (according to him) had regarded as sufficient disclosure of the hardware reselling strategy (see paragraph 1232 above), and (as in the Q3 2010 Quarterly Report) did not mention hardware or “ Arcpliance, Appliances or Strategic Package/Solutions”. By now, gross margins had recovered, and indeed were shown as having increased to 88%, with improvements forecasted. That, at least in part, reflected a reduction in loss-making hardware sales.

The figures for “ IDOL Product ” ($54.4 million), “ IDOL Cloud ” ($52.7 million), “ IDOL OEM ” ($37.2 million)^{^[196]} and “ Services revenue ” ($9 million) in the narrative descriptions and “ Deferred Revenue Release ” ($66.5 million) amounted (after rounding) to the total revenues of $220 million as stated under “ Revenue ”: the categories left no room for other sources of revenue.

Q2 2011 Quarterly/Half yearly Report and Earnings Call

There was no mention of hardware sales, or Arcpliance, or “strategic Package/solutions” in the Q2 2011 Quarterly Report either. The only reference was to ‘appliances’, and then only as follows:

“Autonomy saw expected improvements in gross margins in Q2 2011 compared to 2010 due to the sales mix including more appliances in prior years. Gross profits (IFRS) for H1 2011 were $388.3 million, up 16% from $334.0 million in H1 2010.”

Software sales had recovered well in Q2 2011, and with the increase in gross profit and gross margin, there was no longer any need for words to explain “low margin” sales. As elaborated later, it is clear from the Q&A scripts that had questions been asked about the reference to appliance sales in prior years, the answers would not have disclosed the fact of hardware sales, otherwise than as Arcpliance or “strategic” packages or solutions. In the event, no such question was asked.

Once more, the figures for “ IDOL Product ” ($68.5 million), “ IDOL Cloud ” ($64.3 million), “ IDOL OEM ” ($47.2 million), “ Deferred revenue release ” ($67.5 million) and “ Services ” ($8.8 million) (this time in a table) amounted to the total revenue figure of $256.3 million appearing at the foot of that table, as well as under “ Highlights ”.^{^[197]}

No mention was made of hardware sales in either the Q2 2011 Quarterly Report or in the Q2 2011 Earnings Call, although Autonomy recognised some $20.8 million in respect of ‘pure hardware’ in Q2 2011. Apart from that observation, no particular point needs to be noted in respect of either the report or the call: however, as I elaborate later, it is clear from the Q&A scripts prepared for the Q2 2011 Earnings Call that Dr Lynch and Mr Hussain remained determined that the market should not know about these ‘pure hardware’ sales.

Summary of the Claimants’ case as to the disclosure made in Quarterly Reports

In my judgment, this review of the Annual and Quarterly Reports in the Relevant Period demonstrates that the hardware sales were almost entirely concealed. Concealment required considerable invention.

When, as an inconvenient by-product of the deferment of revenue recognition, $10 million was required to be parked in ‘Inventory’, a story was invented that this was hardware inventory purchased in preparation for Arcpliance sales. The story was false. In subsequent quarters, any reference to ‘hardware’ was excused as being referable to appliance or Arcpliance sales: the fact and extent of what the Claimants have called ‘pure hardware’ sales were to that extent disguised and as to the rest concealed.

As will be examined in more detail later, the costs of hardware were allocated, not to ‘Cost of Sales’ or ‘COGS’ but as to large part as ‘Sales and marketing expenses’ , serving a dual function of (a) disguising the true costs of the hardware purchases and loss-making sales and (b) reducing the adverse impact on gross margin.

It also appears that, in line with the Claimants’ case:

(1) When gross margins were not impacted, or not impacted materially, by loss-making hardware sales, so that the gross margin percentage remained stable or had improved, no reference was included to anything except sales of software: that was the case, for example, in Q4 2009, Q1 2011 and Q2 2011.

(2) When gross margins were adversely impacted, a variety of explanations were put forward, none of which revealed the real reason (loss making hardware sales). Thus:

i. In Q3 2009 the reduction in gross margins was presented as being referable to “unexpected demand for new [products ] ” (SPE and Quick Start);

ii. In Q1 2010, the sudden increase in Inventory in consequence of the deferral of revenue from hardware sales was incorrectly ascribed to the purchase of hardware for Arcpliance;

iii. In Q2 2010, the deterioration in gross margins was ascribed to the quarter ’s “sales mix including Appliances” (appliance sales being at lower margins) without any reference to loss-making hardware sales;

iv. In Q3 2010, when gross profits were once more stable, the past was explained in terms of the “ sales mix” in the previous quarter;

v. In Q4 2010, at Deloitte ’s insistence, words to capture low margin and loss-making sales were included, but in very opaque terms ( “ package solutions, constructed of services, hardware and software, such as Arcpliance” );

vi. In Q1 2011 and Q2 2011, by which time gross margins first stabilised and then increased in percentage terms, no reference was made to hardware, or even to some broad description such as “ package solutions, constructed of services, hardware and software, such as Arcpliance”.

Overall, the impression given in every report was that all revenues were generated through IDOL software transactions, except those relating to services, which were separately accounted for. That further appeared to confirm the Claimants ’ case (as summarised above) that Autonomy presented itself to the market generally, and to HP during the pre-acquisition discussions and due diligence, as a ‘pure software company’ whose revenues were almost exclusively derived from the sale of its IDOL software and related software (and a small amount of ‘appliance sales’).

What was said, or not said, about the sales in the quarterly earnings calls was correspondingly disingenuous at best. I have determined that what Dr Lynch said at the Earnings Call for the quarter (Q1 2010) when both Defendants provided an explanation for the Inventory was patently false, and he elaborated on the falsity during the call which exacerbated the falsity. Allowing for his hearing problem, I was not persuaded that Dr Lynch was in some way at cross-purposes because volcanic ash had marooned him in California and he was on a mobile phone which he could not connect to his hearing aid: the transcript is clear and consistent and his responses did not seem to me to suggest that he misheard the relevant question.

Is the Defendants ’ avowed reliance on Deloitte a “trump card” ?

The fact remains that Deloitte approved the Quarterly Reports and Autonomy’s quarterly, half-yearly and annual accounts throughout the Relevant Period, and so too did the Audit Committee. Furthermore, at least one representative of Deloitte (I think invariably Mr Welham) sat in on each of the Earnings Calls (see further paragraphs 123 above, 1437 and 1519(6) below) in the course of which it is clear, at least in retrospect, that (at best) incomplete and on various occasions fundamentally inaccurate information was provided about the hardware reselling revenue’s contribution to overall revenues.

Neither Deloitte nor the Audit Committee expressed any material doubt as to the commercial propriety of the hardware reselling or the Defendants’ explanation of its purpose: on the contrary, both accepted the purpose given as a valid basis for accounting for and treating hardware reselling revenue as part and parcel of Autonomy’s core software business.

As was essential in order to achieve the true purpose of the hardware reselling strategy, both Deloitte and the Audit Committee were also persuaded, despite reservations they expressed, not to insist on separate disclosure of the hardware transactions, either in the ‘back-end’ of the accounts, or in any of Autonomy’s published information.

The fact that Mr Welham and all those concerned at Deloitte, including the audit engagement partners in overall charge of the Autonomy audit in the Relevant Period, Mr Knights (from 2005 to 2010) and Mr Mercer (from May 2010), considered the hardware sales in great detail and were persuaded (a) that their purpose was to protect and promote software sales; (b) that some proportion of the costs were properly to be accounted for as marketing expenses; and (c) that the accounting treatment, the Supplemental Metrics and the cryptic and limited information given could properly be approved, is obviously much relied on by the Defendants.

The Defendants accepted that the responsibility for the accounting treatment was ultimately that of the Directors. Directors put forward their account of the performance of the company of which they have management control: auditors audit that account. However, they insisted that unless materially misled, Deloitte’s acceptance of the accounting put forward by the directors, after careful audit, with full access to the accounting information, and with knowledge of the extent of the hardware reselling, demonstrates that, whatever else, they were not dishonest. Indeed, their case was that they believed the hardware reselling and the way it was accounted for to be entirely proper, and were confirmed in that belief by Deloitte.

However, if the Defendants did mislead Deloitte, that would of itself be powerful evidence, not only of the impropriety inherent in any deliberate misrepresentation, but also that they considered that disclosure would fatally undermine the programme (which could only be the case if its purpose was as the Claimants alleged).

The Claimants’ case is that the reason why Deloitte and the Audit Committee did not raise any red flags, accepted accounting treatment of the hardware reselling revenues and costs which undoubtedly obscured their nature and source, and never insisted on more transparency, was that both were fundamentally misled by an avowed purpose which was bogus and a false narrative which was developed by the Defendants with the assistance of (in particular) Mr Kanter and Mr Chamberlain and adjusted over time as events progressively raised more questions as to its legitimacy.

The Claimants’ case on the reasons for avoiding disclosure and the falsity of the Defendants’ presentation of the purpose of the hardware sales come together and are mutually supportive at this point. The presentation of the purpose of the hardware sales as being to drive software sales was necessary to persuade Deloitte that the hardware revenues did not have to be differentiated in the accounts nor did their source need to be disclosed. If the true nature and extent of the sales were to be disclosed, the programme would be revealed for what it was: a means of covering up disappointing software sales, and there would have been no point in it.

It was on that basis that Mr Rabinowitz submitted in his oral closing that:

“Deloitte is not the trump card that Dr Lynch would like it to be.”

The development of the narrative presented to Deloitte and the Audit Committee

It seems to me that the most convenient way of addressing and determining this key aspect of the hardware case is to track the development of the narrative by reference to the phases in the programme marked by:

(1) The background and development, and the presentation and effect, of the Strategic Deals Memorandum, which was principally focused on the arrangements between Autonomy and EMC but was formative of Deloitte’s perception of the hardware reselling strategy even after Dell replaced EMC as Autonomy’s hardware supplier for the programme.

(2) The reinforcement and adjustment of the justification of the hardware reselling strategy and its costs after the replacement of EMC by Dell (and especially from early 2010 onwards), principally in Quarterly Notes prepared by Mr Hussain and Mr Chamberlain for the Audit Committee (“Mr Hussain’s Quarterly Notes”). These Notes were always pre-circulated to Dr Lynch and Deloitte, and when presented to the Audit Committee were accompanied by Deloitte’s own quarterly reports on their reviews in each quarter of the accounts for that quarter (“Deloitte’s Quarterly Reports”), and its effect.

(3) The development of the “Linkage Analysis” in successive editions from Q2 2010 to persuade Deloitte and the Audit Committee that the hardware reselling strategy had discernible effect on and was of benefit to software sales.

Efforts to obtain “a helpful form of words” from EMC to “wave in front of Deloitte”

It will be recalled that in its original incarnation, when EMC was Autonomy’s hardware supplier, the Defendants presented the hardware reselling strategy as involving a ‘partnership’ with EMC for the development jointly of an ‘appliance’ and for marketing and related arrangements to be undertaken by EMC, funded in part out of the purchase price for the hardware purchased by Autonomy. This was bound to raise Deloitte’s expectation of confirmatory evidence from EMC.

Autonomy’s management spent much of September 2009 pressing Mr Sullivan to use his contacts in EMC to get them to agree to wording which would provide confirmation of the arrangements, and especially of some commitment on the part of EMC to begin the joint development of an appliance and to treat part of its receipts from the sale of its hardware as a “marketing incentive” (there being none in the relevant contracts).

The Claimants’ closing submissions contained a detailed chronological description, by reference to a number of email exchanges, of what they presented as the orchestration by Mr Hussain and Mr Chamberlain, in conjunction with Mr Sullivan, of efforts to extract from EMC “a helpful form of words...that could be waved in front of Deloitte.”

They also described how on 15 September 2009, Dr Lynch and Mr Hussain had agreed to pay Mr Sullivan a special bonus of $50,000 if he succeeded in this regard (I have quoted the email in question and the incentive arrangements in greater detail in paragraphs 1119 and 1127 above).

I do not think it is necessary for me to rehearse in as much detail as the Claimants provided the numerous exchanges between Mr Sullivan and Mr Hussain and/or Mr Chamberlain, which were characterised by the latter two formulating, re-formulating and pressing for the wording to which they wanted EMC to agree, and Mr Sullivan consistently having to suggest modifications and to warn (in effect, even if not in so many words) that EMC would not agree wording which did not reasonably accurately reflect his actual discussions with them.

It is, however, necessary to refer to the exchanges in summary to illustrate why it seems to me to be clear that what Dr Lynch in cross-examination sought to justify as the efforts of Mr Hussain and Mr Chamberlain to “get the right accounting…the accounting… reflecting the reality of the situation” (and see also paragraph 1342 below) ultimately exposed the fact that EMC did not regard itself as committed in any way, and was not willing to provide more than anodyne enthusiasm for the reselling programme which provided it (EMC) with undoubted benefit (a discount for its customers at Autonomy’s expense).

In summary:

(1) Initially, what Mr Hussain and Mr Chamberlain wanted (as is clear from the email of 15 September 2009) and urged Mr Sullivan to “extract” was EMC’s agreement (actual or tacit) to the presentation of the price paid by Autonomy for their hardware as comprised of three components, being “hardware cost” , “commission for resale” and a “marketing fee” , each with an allocated dollar amount. Mr Chamberlain gave as an illustrative example a $31 million hardware order, and the suggested description of $20 million as “hardware cost” , $6 million as a “commission for resale” and $5 million as a “marketing fee” . Mr Chamberlain explained to Mr Sullivan that this was “the sort of thing we need from EMC for internal cost allocation” . Mr Sullivan did not feel able to propose this since he did not think EMC would be “that specific on their cost” .

(2) Mr Sullivan’s proposal for blander wording sparked a counter-proposal from Mr Chamberlain which he presented as a “minor tweak” , but which the Claimants (justifiably, in my view) presented as a substantive change designed “to allow us to show $20m as COGS and $10m of marketing”.

(3) When Mr Sullivan balked at seeking the proposed wording from EMC, there followed a further exchange culminating in agreement on revised wording to be put to EMC providing “details of Autonomy’s Q3 marketing program” under which Autonomy would “purchase products from EMC at a price consisting of” (i) “EMC’s discounted price to Autonomy, which typically will represent a substantial discount off list price but will be determined on a deal by deal basis” and (ii) “plus a marketing incentive supplied to EMC to be determined on a deal by deal basis”. Mr Sullivan put this version to EMC in an email dated 18 September 2009 ending his email “I would appreciate it if you could confirm your understanding of the program.”

(4) Even though Mr Sullivan’s email involved no suggestion that EMC had assumed any obligation to provide marketing assistance to Autonomy, still less any suggestion of the agreed development of an appliance, EMC’s reply that afternoon was studiously non-committal, simply stating:

“This looks like a great programme and we are excited to participate in it.”

(5) All that appears to have been sent to Deloitte (addressed to Mr Knight and Mr Welham) by Mr Chamberlain (and only at a later date, in early October 2009) is the thread of (a) Mr Sullivan’s email and (b) EMC’s response on 18 September 2009. Mr Knight forwarded that thread to Mr Knights with the comment:

“Helpful but not enough to substantiate a $25m marketing element in my view. I have asked if EMC can quantify but I suspect that this is all we will get.”

(6) On 2 October 2009, by email to Mr Sullivan, copying Mr Hussain, Mr Chamberlain proposed new wording which invited EMC to confirm that (a) EMC would spend a material proportion of the difference between the Autonomy selling price and the EMC selling price, which was labelled “the premium” , together with a “distributor premium” , on “development of the EMC cells and working on training the sales force and joint marketing with Autonomy to further develop the EMC-Autonomy partnership” ; and also (b) that “the standard reseller margin is approximately 55%”.

(7) Mr Sullivan could not support this tack either. His response (by email to Mr Hussain and Mr Chamberlain dated 2 October 2009) was that he was “Not optimistic about this…” ^{^[198]}

(8) Mr Hussain immediately replied to both Mr Sullivan and Mr Chamberlain, “Give me an idea of what you could get”. The Claimants suggested also that Mr Hussain’s reply showed that “Mr Hussain simply wanted whatever form of words could be extracted from EMC to show Deloitte, whether or not it bore any relationship to the truth”.

(9) Mr Sullivan was not, however, to be pushed too far. As he put it in his witness statement:

“On occasion, I was asked by Mr Hussain or Mr Chamberlain to extract certain language from EMC and when the suggested language was not accurate, I offered modified language that was accurate.”

(10) Mr Sullivan’s further email response of 2 October 2009 to Mr Hussain and Mr Chamberlain (but not, it seems, Dr Lynch) was unequivocal:

“…They will not OK anything that says that what we paid them was for something other than for the product we purchased in this period. Nor will they say the money will be spent on marketing etc. They are extremely sensitive to these kinds of letters.

If I can get them to agree to anything it will need to be more general:

1. Maybe acknowledge that our marketing program is not limited to Q2. This would imply that we would do more of the same in Q4 and beyond.

2. They have already acknowledged that the fees paid were a marketing incentive…”

(11) Mr Sullivan could scarcely have been plainer; but still Mr Chamberlain persisted. On 5 October 2009 he emailed Mr Sullivan, copying Mr Hussain, stating:

“We have to get EMC to state that they will spend $’s over next few quarters on developing their cells…the key at this stage is to get something that will state that their [sic] will be a future investment by EMC”.

(12) Mr Chamberlain proposed new and briefer language including:

“EMC plan to continue to invest in the relationship through continued development of our cells and similar joint marketing with other customers”.

(13) Mr Sullivan could not accept this either. On 5 October 2009 he proposed and sent, with the reluctant but resigned approval of Mr Chamberlain and Mr Hussain, an email which thanked EMC for “participating in our key customer marketing program” and looked forward to the possibility of “the continuation of the program to include: development of an appliance bundle, mutual cross referrals, account introductions, reselling and other opportunities ” (as also quoted in paragraph 774(5) above). As the Claimants pointed out, that wording contained no suggestion that such matter had been agreed with EMC: rather, these were to be discussed in the future.

(14) Having received no reply, Mr Sullivan chased Mr Scannell by email dated 14 October 2009. The content of that email demonstrated that Mr Sullivan (a) recognised there was no prospect of the sort of confirmation that Mr Hussain and Mr Chamberlain had sought and (b) well understood that there never had been any commitment on the part of EMC either to spend money on marketing for Autonomy or to collaborate on a joint appliance. All that was left was to talk lamely of wanting:

“to follow up again to see if we could spend some time talking about other ways we could help each other. Amongst other things, we are going to launch some appliance products and will need hardware to bundle into the product.”

(15) Mr Scannell again sent no reply.

It seems to me that it was plain by then to Mr Hussain and Mr Chamberlain, as illustrated by their reluctant retreat from seeking to persuade Mr Sullivan to try yet again to extract wording from EMC showing some definite commitment to apply the ‘delta’ in pre-agreed ways, that some other justification for the allocation of a proportion of the hardware costs to sales and marketing had to be found and substantiated.

From then on, as it seems to me, their focus was no longer on obtaining from EMC any acknowledgement of any commitment. Instead, management turned to seek through Mr Sullivan EMC’s acknowledgement of some generally applicable but quantified parameter, whether a quantified “hardware amount” or “ the standard reseller margin” , with a view to persuading Deloitte that the difference between the amount to be regarded as the selling price and the amount actually paid to EMC by Autonomy (the ‘delta’) should be treated as referable to marketing incentives or development funding. This was what became known as the “residual approach.”

In the meantime, as the Claimants presented it, Mr Hussain had gone ahead with efforts to maximise the amount of hardware costs that could be allocated to sales and marketing expenses in Autonomy’s accounts. This process too had gone through a number of stages:

(1) On 29 September 2009, Mr Chamberlain sent Mr Hussain by email a summary of the Q3 hardware deals. He attached a spreadsheet showing that by that point Autonomy had purchased hardware from EMC for a total of $50.722 million and Autonomy had re-sold that hardware for $40.873 million, thereby incurring a loss of $9.859 million. Mr Chamberlain had proposed (as shown in the spreadsheet) to allocate $30.654 million of the $50.722 million spent to COGS and the remaining $20.068 million to sales and marketing.

(2) Mr Hussain was not satisfied. On 30 September 2009 he told Mr Chamberlain in an email that he needed “to get cogs down by at least $5 million so that it’s 50:50”. That necessitated decreasing the amount allocated to COGS by $5 million, and increasing the amount allocated to marketing by that amount, to achieve a roughly 50:50 split. No basis for the change in allocation was suggested. The Claimants therefore submitted that the allocation was driven simply by a desire to attribute as much as possible to marketing and reduce to a minimum the amount allocated to COGS.

(3) In response, Mr Chamberlain duly amended the spreadsheet, though he had to adopt a slightly more complex method to achieve the result. The $50.722 million was now allocated as follows: $26.567 million to COGS and the remaining $24.155 million to sales and marketing. This was achieved by reducing (artificially, according to the Claimants) the so-called “wholesale price” of the hardware. Whereas in the original spreadsheet the wholesale price ($30.654 million) was shown as 75% of the retail price ($40.873 million) showing a “Normal margin to reseller” of 25%, in the revised version, the wholesale price was reduced to $26.567 million (i.e. 65% of the retail price), with the “ Normal margin to reseller” increasing to 35%.

(4) The Claimant labelled this as “entirely artificial”, in that Autonomy was not at that stage aware of EMC’s real “wholesale price” , and the reduction of that “wholesale price” enabling an increase of the “Normal Margin to reseller” was arrived at without any input from or discussion with EMC, and to achieve only one end, being to maximise the amount of hardware costs allocated to marketing so that the effect on the bottom line of the hardware reselling costs would commensurately be reduced. It is difficult to disagree with that label: and I accept it.

(5) The evidence did not disclose whether Deloitte knew of these alterations or their basis.

Quite what Deloitte came to know of these unsuccessful efforts to extract serviceable wording from EMC to persuade them of a commitment on EMC’s part to apply the ‘delta’ to marketing and/or development is not clear. It was not disputed that Deloitte always had access to all emails and other internal documents: Autonomy appears to have operated an ‘open book’ policy. There was, however, no documentary trail or evidence demonstrating that Deloitte were sent or saw the emails internal to Autonomy referred to above. Nor, apart from the email from EMC of 18 September 2009 which was sent to Deloitte on 9 October 2009 (see paragraph 1278(5) above), was there evidence of Deloitte having been shown the emails to EMC or Mr Sullivan’s follow-up email dated 14 October 2009 when Mr Scannell did not reply, in which the best that Mr Sullivan felt able to achieve was an acknowledgment of some possibility of such an agreement in the future:

“I wanted to follow up again to see if we could spend some time talking about other ways we could help each other. Amongst other things, we are going to launch some appliance products and will need hardware to bundle into the product.”

In such circumstances, I have concluded that it is unlikely that Deloitte were aware of the exchanges within Autonomy, or (except for the email of 18 September 2009) those (which were few) between Autonomy and EMC.

As a result, it seems to me more likely than not, and I find, that, prior to the final drafts of the Strategic Deals Memorandum, Deloitte were (a) unaware that the metric central to the “residual approach” (EMC’s real “wholesale price” ) was a figure contrived by Autonomy and (b) still looking to Autonomy to supply the substantive evidential support of the collaboration and agreement with EMC which Autonomy’s management continued to cite as justification for the accounting treatment which they proposed.

Mr Welham’s unchallenged evidence in his witness statement was that on 7 October 2009 Mr Hussain met with members of the Deloitte team to explain the background to the EMC hardware transactions. A review working paper with the subject heading “EMC related revenue deals” , (which is actually dated at its top 5 October, but which refers to later events and was described by Ms Anderson of Deloitte, who prepared it, as an extract from a full memorandum), recorded what Deloitte had been told of the commercial rationale at that stage. This included the following:

(1) “Autonomy has decided to apply, in the case of large ongoing projects, a package approach to the demand for strategic selling at major institutions.”

(2) “This will replace the existing situation where a customer has the choice of purchasing software from Autonomy or another software provider and purchasing hardware in a separate transaction from EMC or another hardware provider.”

(3) “Autonomy will encourage customers to purchase EMC hardware and the customer will get a better deal by purchasing through Autonomy than they would receive directly with EMC.”

(4) “The benefit will be seen in the long term with these and other customers as they purchase additional cells to replace or expand their existing storage cells. As Autonomy has aligned themselves strategically with EMC, customers will purchase their cells from EMC via Autonomy. EMC will have configured their cells to make them completely compatible with Digital Safe. Customers will be able to choose Autonomy as their strategic supplier for hardware and software.”

(5) “Autonomy has considered the cost to them in respect of these deals” [the reference being to the deals with Citi, JPMC and Bloomberg] in three portions:

“Costs of goods sold: This is equal to the standard wholesale price…and is equal to 55% of the selling price

Marketing: Autonomy has repaid the reseller margin back to EMC. The standard reseller margin is 45% of the selling price. Autonomy considers this is a marketing cost that they need to incur…

Development: Autonomy has then paid an additional sum on top of the cost of goods and marketing cost which they consider to be a development cost…This amount represents a contribution to EMC’s development costs and is being expensed over the period during which Autonomy is expected to benefit. EMC has confirmed that they expect to spend these $’s over the next few quarters and Autonomy is expecting to generate additional benefit through further sales in that period.”

(6) A note from Mr Welham reads “But is this really a marketing cost. Remember IAS 38 has been amended to clarify this point…”

Behind the scenes at Deloitte, the attitude in September, and up to mid-October 2009, was quite plainly one of considerable scepticism. Mr Welham had also put comments in the margin on Ms Anderson’s Memo. These included:

(1) A comment next to the statement in that Memo that the “standard wholesale price of the goods” was 55% of the selling price, the question “How can we evidence this?” ;

(2) Three exclamation marks and the comment “That is a nice incentive” next to Ms Anderson’s note that Autonomy had “repaid the reseller margin back to EMC” and the standard reseller margin was 45% of the selling price;

(3) A comment next to Ms Anderson’s note that Autonomy had additionally paid what they were describing as a “development cost”, which read “This makes no sense. Why would Autonomy pay for EMC to develop their products, especially they are making no margin on such deals” ; and

(4) Next to Ms Anderson’s note that Autonomy accepted they could not capitalise what they asserted were development costs because they could not demonstrate probable future benefit, a comment “…so we are expecting this to be taken as a cost” .

On 7 October 2009 (and presumably after the meeting the same day referred to above), Mr Knight sent an email to Mr Hussain, Mr Chamberlain and Mr Knights and Mr Welham (and thus to the four finance persons principally involved). In addition to asking for further details of the three deals (with Citi, JPMC and Bloomberg), Deloitte requested documentation, and an explanatory memorandum, to be circulated to the Finance and “wider” audit team and also the Audit Committee, to justify what Autonomy proposed.

Deloitte asked that this explanatory memorandum should include:

(1) “a clear explanation of the commercial rationale for selling at a loss. This needs to explain the exact nature of the benefit you are getting from this deal at a loss and how it alters your relationship with JPMC, Citi etc and also the relationship with EMC [which] will need input from those who negotiated the deal” ;

(2) an explanation of “your rationale for showing the costs associated with the hardware purchase below the line rather than as a cost of sale” given that the contracts themselves “appear to suggest we are buying on and selling equipment as a straight sell through.”

Mr Knight also noted in the email that Mr Hussain had spoken “of an appliance being developed by EMC which will have Autonomy embedded into it” and he asked Autonomy to “please provide the documentation around this” and if to be accounted for as an intangible asset (as I assume must have been suggested at the meeting), an explanation of how its value could be identified, and how it could meet the IAS 38 definition (see below).

Development of the Strategic Deals Memorandum

The Strategic Deals Memorandum was developed as a response to Deloitte’s request for further information as above described. It was itself a development of a document originally file-named “EMC transactions - Audit memo” .

The first draft of the “EMC transactions - Audit memo” was prepared by Mr Hussain and the finance department on 1 October 2009. They amended it the same day, but not substantially. Its subject was presented as:

“a number of additional transactions with EMC whereby Autonomy and EMC sell to existing Autonomy customers product which is subsequently used with Autonomy software.

These transactions represent the commencement of a significant new partnership between EMC and Autonomy to develop EMC data cells such that they are readily compatible with Autonomy software and in particular the Digital Safe. The initial transactions have revolved around common customers of both entities but the long term goal is that customers with EMC storage cells will be referred to Autonomy for data storage software needs…

…

In recognition of the investment that EMC needs to make to ensure marketing focus and further development of the cells, Autonomy has paid a premium to EMC. This premium has two parts:

1. The normal reseller margin (35%) has been returned to EMC; and

2. An additional premium of 20-25% has been paid.

The purpose of the premium is to enable EMC to provide marketing dollars, encourage the EMC salesforce to jointly sell, develop EMC cells and provide joint marketing with Autonomy…

…

The cost to Autonomy is higher [than the price on the onward sale]… and comprises three elements: the cost of the goods and the two elements which make up the premium paid for marketing and development costs to be incurred by EMC.

The standard reseller margin for EMC distribution channels - to which Autonomy is entitled - is 35% so the cost of goods sold represents $23,804,684 (65% of customer selling price)…As previously noted the reseller profit margin of $12,817,907, normally retained by Autonomy, has been passed to EMC and as an additional incentive to deepen and widen the relationship a further $8,789,662 has been paid to EMC.”

The Claimants highlighted numerous falsities in that draft, including:

(1) In cross-examination, Dr Lynch accepted that the statement that the hardware sold was for use with Autonomy software was not accurate, at least in some circumstances, since the customer was under no obligation to use the hardware with Autonomy software and although in some instances the hardware would be used with Autonomy software, in others it would not.

(2) As and in the respects previously explained, Autonomy’s relationship with EMC was not as described.

(3) The arrangements in relation to the amounts paid by Autonomy to EMC were falsely described.

(4) So too were the statements as to the reasons for the “premium”. Autonomy had not agreed to pay EMC a “ premium ” in order for it to take any of the steps suggested in the draft memorandum, and EMC had not agreed to make any “ investment ”, nor had there been any discussion between Autonomy and EMC about Autonomy paying a “ premium ” for product development or marketing efforts.

(5) EMC had not agreed with Autonomy to incur any marketing and developments costs; nor had it provided any confirmation that it expected to expend $8.8 million over the period up to 30 June 2010.

On 2 October 2009, Mr Hussain circulated a draft to Dr Lynch’s PA for him to consider. Dr Lynch did not seek to correct any of the inaccuracies referred to above. Instead, he suggested a change of structure and a new introduction to extend it to cover and explain transactions with other hardware sellers (such as, at that time, Hitachi). The recast document (file-named “EMC accounting memo v3” ) was renamed “Review of strategic transactions with very large organisations - Audit memo (Q2 and Q3 2009).” It is not clear who drafted the amendments. Mr Hussain sent a copy to Dr Lynch’s PA later the same day (2 October 2009).

That recast document (v3) opened with the following (after which the same basic explanation as previously was given of the elements explaining the total costs):

“During the current downturn, to reduce costs, Autonomy has seen a major shift taking place with companies seeking to reduce the number of IT suppliers to 6 or 7 strategic ones. A significant part of these customer’s IT base is related to storage and archiving and, as part of this process, Autonomy has been asked to extend its role into becoming strategic suppliers to these companies.

…

Autonomy decided to apply a package approach to this demand for strategic selling at these major institutions which may mean that in certain instances Autonomy takes terms which may appear loss making initially but when considering the bigger picture are hugely profitable. Autonomy expects this approach to pay off in the longer term as it becomes a key part of the architecture for companies such as Citi, JPMC and Morgan Stanley…

…

Autonomy approached EMC and Hitachi for a strategic partnership to deliver the requirements in Q2… EMC have proven quicker…

…

In recognition of the investment that EMC needs to make to ensure marketing focus and further development of the cells, for certain transactions Autonomy has paid a premium to EMC. This premium has two parts - the normal reseller margin (45%) has been returned to EMC plus a premium of 20-25% has been agreed. The purpose of the premium is to enable EMC to provide marketing dollars, encourage the EMC salesforce to jointly sell, develop EMC cells and provide joint marketing with Autonomy.” [My underlining for emphasis]

The underlined phrases were new. The Claimants contended that each was inaccurate or false, for the following reasons:

(1) The first underlined phrase was false: there was no evidence that the hardware reselling strategy was a response to customer demand, and the presentation had previously been that it was an initiative by Autonomy.

(2) The second underlined phrase was inaccurate: (a) it was EMC, at Autonomy’s request, that found and delivered up to Autonomy customers wishing to buy hardware; (b) there was no “strategic partnership” between Autonomy and EMC; and (c) there was no “strategic partnership” between Autonomy and Hitachi either: only a “One-Time Reseller Authorisation Agreement”.

(3) As to the third underlined passage, the figure of 45% as the “normal reseller margin” contrasted with the percentage given as “the standard reseller margin” in the previous draft (of 35%). The Claimants submitted that (a) neither percentage was substantiated and (b) the change to the higher figure in the recast memorandum was designed simply to justify an increase in the amount of costs to be allocated to sales and marketing (Autonomy by now suggesting that only $20.1 million would be allocated to COGS, with $25.3 million to sales and marketing.)

A yet further (fourth) version was worked on and sent to Dr Lynch by Mr Hussain in the evening of 2 October 2009. This version (which Mr Hussain had envisaged to be the last) contained two additional passages which the Claimants considered to be worth flagging:

(1) A revised opening sentence of the third paragraph was amended to read “ Autonomy has decided to apply, in the case of large ongoing projects, a package approach to this demand for strategic selling to these major institutions ” (new words underlined). The Claimants submitted that this was inaccurate because Autonomy was selling EMC hardware on a freestanding basis and not as any part of an “ ongoing project ” or on a “ package ” basis.

(2) A further sentence was added stating: “ EMC is banking on Autonomy’s bid for strategic supplier in return for the promotion of their solution ”. According to the Claimants, this again was fictitious: EMC had not agreed to promote any “ solution ” nor was EMC “ banking ” on anything.

About half an hour later, Dr Lynch responded “Very clear” , having obviously read the re-revised draft.

Another (supposedly) final version of that document marked v5 but still, contrary to Dr Lynch’s express direction, subject headed “EMC accounting memo” , was circulated by Mr Hussain to Dr Lynch, Mr Chamberlain and Mr Stephan on 4 October 2009. This version was re-circulated at Dr Lynch’s insistence under the revised subject-heading “strategic deals and partnerships memo final” with a covering email from Mr Hussain stating “this is the final document for the files”. However, further drafts followed and the latest draft was sent to Mr Knight and Ms Anderson on 7 October 2009.

Prior to this, according to the Claimants, two things had happened:

(1) Some preliminary discussion had already taken place between Autonomy and Deloitte about the EMC relationship. This appears to be supported by an email that Deloitte’s Mr Rob Knight (not to be confused with Mr Richard Knights, also of Deloitte) sent to Mr Hussain and Mr Chamberlain on 7 October 2009. It is plain from the third paragraph of the email that, by this point, Mr Hussain had told Deloitte that an appliance was “ being developed by EMC which will have Autonomy embedded in it ”. But as already stated, this was exaggerated. EMC was plainly not committed and had repeatedly brushed off Autonomy’s efforts to obtain some sort of confirmation of intent (see paragraphs 1272 to 1284 above).

(2) Relatedly, steps were being taken to produce further documentation that the Claimants suggested was to be provided to Deloitte to support Autonomy’s accounting for these transactions. On 7 October 2009, Mr Hussain sent an email to Ms Julie Dolan (Autonomy Senior Corporate Counsel), copied to Mr Kanter, with the subject “ draft partnership agreement ”, in the following terms:

“Can you draft up a one pager with the following:

- between emc and autn

- partnership to work on developing an appliance that combines Autonomy compliance software (digital safe, introspect) with EMC hardware

- timescale 6 months

- both parties to respect confidentiality of information

- neither party has obligation to perform etc

need it immediately”

The Claimants submitted that this was plainly an effort on the part of Mr Hussain (involving also Mr Kanter) to paper the record to support what he had told the auditors, who could be expected to wish to see some sort of written confirmation that EMC had committed to spend the significant profit that it was making from these deals - at Autonomy’s expense - on a variety of marketing and development initiatives.

The Claimants further submitted that it was plainly a contrivance given that (a) EMC had not entered into a partnership with Autonomy, and had not agreed to develop an appliance with Autonomy, (b) no such partnership agreement was even under active discussion (as is apparent from Mr Sullivan’s email to Mr Scannell of 5 October 2009), (c) at most, this was all something that might be discussed in the future, and had no such urgency as was imported by the last sentence of the email, and (d) there is nothing to suggest that this draft partnership agreement was in fact ever shared with EMC: the Claimants suggested that this was because it was recognised that EMC would not sign even a non-binding statement of intent.

However, there is no evidence either that any such draft Partnership Agreement was prepared; still less that such a draft was shown to Deloitte. Quite what its purpose was remains unclear, and though I accept that the Claimants’ explanation of Mr Hussain’s sudden call for a draft is probably the most likely one in all the circumstances, he seems to have thought better of it.

No mention was made of it in the recast document, renamed and referred to from now on as the Strategic Deals Memorandum or SDM, the objective of which continued to be to overcome Deloitte’s scepticism as to the justification for (a) reselling at a loss and (b) allocating a substantial proportion of the costs as Sales and Marketing expenses. The drafting process continued in tandem with the faltering and ultimately unsuccessful efforts by Autonomy’s finance department (described above) to extract some confirmation from EMC which would satisfy Deloitte on both points and Autonomy’s assertion as to the development of an EMC/Autonomy appliance. The document went through a number of drafts, usually being exchanged between Mr Hussain, Mr Chamberlain and Mr Sullivan, but not, until the final draft, with Dr Lynch.

Exchanges evidencing the drafting process show, to my mind, (a) Deloitte’s focus turning increasingly (as might be expected, since it was an accounting not merely a narrative matter) to the issue of cost allocation and whether there was any justification for a proportion being allocated as a sales and marketing expense or R&D costs; and in consequence (b) Mr Hussain continuing to try (i) to get from EMC (via Mr Sullivan) some details, especially as to the “standard reseller margin” to substantiate the figure of 45% given to Deloitte, and (ii) to plump up the description of the “strategic supplier relationship” with EMC in such a way as to persuade Deloitte that in surrendering to EMC the difference or “non-reseller margin” Autonomy was paying sales and marketing expenses or R&D costs.

An example of the latter was Mr Hussain’s email to Mr Chamberlain of 12 October 2009, which was forwarded to Deloitte later that day, and which included the following:

“The customer relationships are very hard to achieve and Autonomy has accepted that for its part paying for the marketing, sales and r&d effort of the h/w vendors is of major long term benefit. EMC and Autonomy are developing an appliance for the future … In terms of the accounting we have provided evidence of the reseller margin, so the remainder of the cost is accounted as sales, marketing and r&d - we have allocated to sales and marketing. The non reseller margin monies are being used to incentivise the emc, hds, acs salesforce, provide discounts to the customer, provide funds for the development of the appliances, to hold marketing programs … We would strongly argue that a large proportion of the monies are being used for the development of the appliance which has a significant future value but we have taken a very prudent view and have expensed the total amounts.”

The problem continued to be, however, that there was nothing emanating from EMC to substantiate any of this, which is really what Deloitte had asked for in order to justify the accounting treatment Autonomy’s finance department were pressing for. Later that day (12 October 2009), Mr Knight responded to Mr Hussain and Mr Chamberlain as follows:

“Can we get anything from EMC quantifying the hardware amount?

Is there a set marketing programme or any further information available which can help us to quantify the marketing element - eg a joint marketing plan or similar? Evidence which helps us understand the marketing side in some more depth would be very helpful.

Again on the appliance development, is there anything to further substantiate the amount of development you are funding as opposed to relying on this being a balancing figure.”^{^[199]}

Thus, as at 12 October 2009, it was plain that Deloitte’s attitude to the proposed accountancy treatment of the hardware sales costs/losses still remained sceptical. The email from Mr Scannell at EMC of 18 September 2009, which Mr Chamberlain had forwarded to Deloitte on 9 October 2009 (see paragraph 1278(5) above) had not changed this substantively. In particular, and as Mr Chamberlain had anticipated, as already recorded above, it was not regarded by Deloitte as “enough to substantiate a $25m marketing element”.

Mr Hussain was well aware that Deloitte were, at that stage, unpersuaded. Knowing by now (that is, by 12 October) that evidence of any bilateral agreement would not eventuate, he pressed Mr Sullivan instead to get the basic information necessary to substantiate the figure for the “standard reseller margin” which was the foundation of the “residual approach”:

(1) On 12 October 2009, in an email chain between Mr Hussain and Mr Sullivan, copying in Mr Chamberlain, Mr Hussain asked Mr Sullivan to contact EMC:

“Having issues with the auditors.

Can you get someone to send us an email stating what their standard reseller margin (can be for s/w or appliance sales) is between 40% and 50%?”

(2) Mr Sullivan responded on the same day with:

“I will try, but it is holiday today in Mass. So it will be tough. Also, I have learned that EMC guards the details of their reseller program closely. I believe the answer is that standard margins are lower than what you indicate, but on a one-off basis, EMC will approve deeper discounts for larger or more strategic accounts similar to those that we did in Q3.”

(3) Later, on 12 October 2009 at 9.38pm, Mr Hussain sent an email to Mr Chamberlain with the subject “strategic appliance sales” stating:

“Steve,

Auditors already have our draft memo which explains the strategic supplier status we have established with the likes of Citi, jpmc, Morgan Stanley etc by selling both software and hardware for compliance application. In addition I have explained the nature of the relationship with the suppliers emc, hds and acs to the auditors. The driving force for the sales has come from our joint customers. The customer relationships are very hard to achieve and Autonomy has accepted that for its part paying for the marketing, sales and r&d effort of the h/w vendors is of major long term benefit. EMC and Autonomy are developing an appliance for the future and we are engaged in negotiations with HDS for installing our software on their hardware for our data centres.

In terms of the accounting we have provided evidence of the reseller margin, so the remainder of the cost is accounted as sales, marketing and r&d-we have allocated to sales and marketing. The non reseller margin monies are being used to incentivise the emc, hds, acs salesforce, provide discounts to the customer, provide funds for the development of the appliances, to hold marketing programs (e.g. we attended a major EMC marketing event for JPMC last week which we would not have without the marketing dollars). We would strongly argue that a large proportion of the monies are being used for the development of the appliance which has significant future value but we have taken a very prudent view and have expensed the total amounts. The sales and marketing spend has resulted in new accounts being developed- for example more deals have been identified already.

Please let me know what additional information the auditors need but we are being prudent in our accounting. Add what you need that is relevant otherwise feel free to share this email with Rob and Lee.”

(4) Mr Chamberlain sent the above email on to Mr Knight and Mr Welham of Deloitte within 5 minutes of receiving the email from Mr Hussain under cover of an email stating:

“See below from Sushovan.

I have provided an analysis of standard reseller rates as well as evidence from EMC regarding the fact that we are paying for hardware and marketing incentive. Below is further explanation of what we are getting for our marketing $’s. There is an argument that some of this is future development costs but we did not feel we met the IAS38 definitions and so have expensed those balances.”

(5) Mr Knight responded shortly thereafter copying in Mr Hussain stating:

“Steve/Sushovan,

Can we get anything from EMC quantifying the hardware amount?

Is there a set marketing programme or any further information which can help us quantify the marketing element-eg a joint marketing plan or similar? Evidence which helps us understand the marketing side in some more depth would be very helpful.

Again on the appliance development, is there anything further to substantiate the amount of development you are funding as opposed to relying on this being a balancing figure…”

(6) On 13 October 2009, at 9.31am, Mr Hussain sent an email to Mr Sullivan (which I have no reason to think was seen by Deloitte) in the following terms:

“This is what I’d like from emc

Dear Mike

I confirm that the standard reseller margin is around 35% but for large strategic deals this can go as high as 45%”.

(7) Mr Hussain then forwarded that email to Dr Lynch at the same time stating:

“We have spoken a number of times-this is in essence what he is trying to get.”

(8) Later on 13 October 2009, Mr Hussain informed Dr Lynch that Deloitte may wish to speak to him in relation to the matter. To prepare for this, Mr Hussain sent Dr Lynch an email (cc’d to Mr Chamberlain) timed at 1.24pm, which largely recited the themes that Mr Hussain had developed in his email of 12 October 2009 to Mr Chamberlain (see paragraph 1305 above). The Claimants contended that this email was not only “replete with the same falsehoods” but also that it was designed to give the impression that Dr Lynch had no knowledge of the strategy whilst nevertheless reminding him of the storyline:

“The auditors would like some more colour on the strategic relationships with EMC and HDS in order to understand the accounting for the costs associated with the strategic deals signed with Citi, MS, JPMC and Bloomberg this quarter. I have spoken to the auditors several times and have submitted a paper that described the background to the sales made.

The relationship with EMC and HDS are very important to Autonomy. These are companies with multi billion dollar sales and with significant relationships with global companies. The trend with these global companies is to move towards limiting the number of strategic suppliers and Autonomy is positioning itself as the key vendor of choice for compliance and related software and appliances.

Based on the success of the last six months, Autonomy is paying EMC and HDS for marketing and also for r&d in developing the next generation integrated appliance. The marketing is very targeted because there are relatively few companies both of us would want to market to. Already we have taken part in a large marketing event organised by EMC for JPMC-all the senior IT and business decision makers were at this event-and Autonomy took part in it. These marketing events are focussed and Autonomy’s payment to EMC allows participation.

Autonomy is also paying for EMC and Hitachi salesforce to be incentivised and to introduce Autonomy to the key identified accounts. Finally there is expenditure related to the development of the appliance. Already Autonomy uses EMC hardware with compliance software in its data centres and EMC have OEMed Autonomy into a number of its strategic products. The appliance is the next stage in the development of the relationship.

So overall the payment to EMC and HDS consists of a refund of the reseller margin (COGS) and the remainder is funding sales and marketing (events, advertising, seminars, commissions and r&d (development of the appliance).

The auditors may call you later today to discuss.”

(9) Mr Hussain followed this up some 10 minutes later with the following email (also cc’d to Mr Chamberlain) under subject heading “reseller margins for h/w”:

“The auditors may ask you a question that reseller margins for h/w are not 45%. Points I have made are:

1. Its not pure hardware but part of an appliance sale

2. The level of sales were large and strategic so reseller margins would be higher.

3. The margins for pure h/w are actually much bigger.”

I am not persuaded by the Claimants’ suggestion that the purpose of these emails was to give the impression that Dr Lynch had no knowledge of the strategy. In my view, their purpose was to ‘prep’ Dr Lynch, and update him as to the then position. However, given Dr Lynch’s previous involvement, the emails do seem to me to betray Mr Hussain’s recognition and concern that the presentation to Deloitte had become so embellished that its detailed recitation was necessary to cover any changes from that previously agreed or understood. Even if its toes were in the truth, its exaggerations had become significant (to the point of falsity), and its full recitation was necessary lest Dr Lynch speak out of turn.

So far as I am aware from the evidence, there was no record of the discussion between Deloitte and Dr Lynch for which it appears Mr Hussain was ‘prepping’ Dr Lynch; but it seems likely that there was one ^{^[200]} , and that it was the catalyst for a turn in events which to my mind substantially shaped Deloitte’s move from scepticism to acceptance.

That turn of events was that, on the evening of 13 October 2009, Mr Knights himself became involved in the re-drafting of the “ Strategic Deals Memorandum”, the document intended to answer questions raised by his own firm.

Mr Knights’ own involvement in drafting the Strategic Deals Memorandum

Although he was careful to direct that the final version as sent to Deloitte should come from Autonomy, and when sending it to others in Deloitte on 14 October 2009 described it as “the client produced memo which I have been through with the CEO and CFO…” , it was Mr Knights who was responsible for the penultimate (or possibly pre-penultimate) version of the Strategic Deals Memorandum , and for instigating the work which resulted in its finalised form the next day.

It is apparent from an email dated 13 October 2009 (timed at 10.11pm) sent from the email address of Claire Knights to Mr Hussain (copied to Mr Chamberlain) with the subject “draft of your EMC paper-from RK” and attaching a file entitled “Autonomy EMC transactions”. The body of the email set out:

“S&S

“After a glass or two of red wine and a plate ful [sic] of Mrs K medieval pasta I’ve had a stab at writing the Autonomy paper on EMC-

This needs to come form [sic] you to us.

I need it to square the position on COGS allocation-i’ve still not seen anything from EMC.

If it is useless please re-write but hopefully it points in the right direction. As per our discussion on Mike L’s ideas the paper goes through this analysis - you need to beef it up - the one key area is setting out quite what “sales and marketing” type stuff or further product development stuff EMC might provide you.

Please improve this and together with the EMC email I’m hoping this will move to where it needs to be.

I do need to run through this with Mike as well tomorrow.

As mentioned above this was rattled out pretty quickly and fortified with a few liveners so as a modest bookkeeper it would benefit from the cutting edge of you software gurus….!!

R”.

The basic structure of the arguments to get “where it needs to be” remained in the final version, but as the above email indicated, there were matters which, when drafting his late-night version, Mr Knights identified as needing further support or elaboration. In particular, there were five important parts relating to the allocation of the hardware costs (which Mr Knights in a later mail described as “the nub of this” ) left uncompleted in Mr Knights’ 13 October evening draft, which needed (and on which Mr Knights expressly invited) explanation and further detail:

(1) After a reference to the potential “value add opportunities” that could arise from “this relationship with EMC” , a sentence followed which read “It is possible that if successful it may be possible to move towards an “Intel Inside” type of arrangement with EMC hardware” : Mr Knights has in square brackets noted: [“Sush this is probably simplistic…are there any other upsides??”].

(2) After a sentence which read “The hardware component of this transaction was used to supply our customers [Bloomberg/Citi ???] with equipment for their existing data warehousing and storage functions” , Mr Knights, has in addition to the words square bracketed above, put in square brackets: “[Sush/Steve….please improve]”.

(3) After a sentence which read “The allocation of $45m between hardware and other marketing services is not established in the purchase order or invoice from EMC”, Mr Knights has in square brackets asked: “[can you comment here on how the $45 m was arrived at ???]”.

(4) After a sentence which read “The Autonomy rationale for this transaction was to combine the delivery of hardware to key customers (thereby beginning to develop the recognition of being an application player) together with making an investment in the growing relationship with EMC” , Mr Knights has noted in square brackets: “ [I need help explaining what EMC and you think you will be getting. Trade sales/customer meetings ??? This is the part that Mike Lynch was alluding to this morning - so can you put some ideas in here]”.

(5) After a reference to “…the standard reseller margins that [EMC] would expect to see in the sale of its hardware through a third party” Mr Knights has in brackets noted: (see appendix XX. Insert confirmation) ; and also he has interpolated the words in square brackets in the following: “This demonstrates that in normal situations a range of [37.5% - 50%...check I’ve not seen this yet] is the standard reseller margin”.

It appears that first thing the next morning (14 October 2009), Mr Knights met with Mr Hussain to go through his draft and the questions raised. At 9.30am, Mr Knights emailed Dr Lynch and Mr Hussain (under subject heading “EMC - can we have a brief chat today?” and commencing “Dear Mike”) recounting that he and Mr Hussain had been going over the Q3 position. Remarking that he had “a sense that the engine is running a little hot in certain areas of the Q3 numbers”, he stated that he would appreciate Dr Lynch’s thoughts on the EMC transactions and “the strategic long term importance of the deal” so as to assist in determining “the nub of this” which he identified as being “how to appropriately reflect the costs of this transaction.”

He identified three potential “landing areas”:

“-$45 million to be shown as a cost of sale (and therefore shown in the gross margin)

-$36m to be shown as a cost of sale (on say assumed net margin of nil) with $9m classified as sales and marketing expense, or

-A different split. If however this is arrived at it does need to be fully supported. To date we’ve only seen a purchase order for the $45m albeit I understand there has been some further correspondence with EMC to determine what was actually included in this transaction.”

The further correspondence that Mr Knights was referring to comprised:

(1) An email dated 13 October 2009 from a Mr Mussulli at EMC (which was in due course appended to the Strategic Deals Memorandum) stating (and see also paragraph 1324 above):

“Mike, per your request, our typical pricing for entry level partners in our Velocity programme is 36% off Clarion and 56% off on DMX. If you have any additional questions please feel free to call me at the number below…”^{^[201]}

(2) An email from Mr Sullivan with the subject “Re pricing info” with an explanation (which I assume came from Mr Mussulli) explaining the split between Clarion and DMX stating:

“Fortunately-this breaks out along customer lines:

JPMC =100% DMX

Bloomberg =100% Clarion

CS = 100% DMX

Citi = approx.: 35% Clarion; 65% DMX…”

Dr Lynch forwarded that email to Mr Hussain, who replied immediately to Dr Lynch by email:

“…We have firm evidence of the product cost of the hardware so the residual is clearly marketing which we have explained to Richard as being for:

- Seminars

- Meetings

- Customer events

- Incentives to the emc salesforce to bring us deals

- Development of the appliance”

[My emphasis]

It appears from Mr Welham’s witness statement (which was in this respect not challenged) that at some point in the morning (I assume after the exchanges between Mr Hussain and Dr Lynch) Mr Welham and Mr Knights had a telephone call with Dr Lynch. The purpose of this (as described by Mr Welham) was to hear directly from Dr Lynch, as Autonomy’s CEO,

“as to the rationale for what was a material transaction for the quarter involving a new strategy, that is reselling third party hardware at a loss in order, we understood, to support a broader marketing objective.”

Mr Welham went on to say that during the call, Dr Lynch confirmed to them that this was the rationale.

There is no suggestion in Mr Welham’s evidence that he or Mr Knights discussed with Dr Lynch the other questions that Mr Knights had raised, which went to the proper accounting treatment and allocation of the costs of the purchase from EMC (as distinct from the narrative of the rationale for the hardware reselling strategy). However, the Claimants put to Dr Lynch in cross examination that Mr Knights (in particular, in an email to colleagues in Deloitte enclosing the final draft of the “Strategic Deals Memorandum” further described in paragraph 1324 below) stated that he had been through the memorandum with the CEO and CFO in relation to hardware sales.

Dr Lynch was asked repeatedly about this; and his invariable response was that he did not think he had done this. He sought to maintain throughout that his discussions with Deloitte were limited to dealing with questions as to “why we wanted to do this” and explaining “what we were trying to achieve” and that “his role and recollection in dealing with Mr Knights was to talk about commercial rationale”.

For reasons which I elaborate in paragraphs 1341 to 1348 below (when assessing what Dr Lynch knew of the decision to account for part of COGS as marketing expenses) I have concluded that it is more likely than not that Dr Lynch had been through all aspects of the memorandum, including the part relating to the costs allocation between COGS and sales and marketing.

After changes had been made in the draft memorandum to fill in the gaps left by Mr Knights in his draft the previous evening, and in particular to reflect the email exchanges with EMC (which were not forwarded to Mr Knights until after he had sent out his 13 October 2009 draft), the final version of the “Strategic Deals Memorandum” was sent to Deloitte on 14 October 2009 at 10.19am in an attachment “Strategic deals memo 14^th Oct” (copied to Mr Hussain) stating “here is our paper updated for additional items received from EMC last night”.

Showing my own underlining of passages especially emphasised by the Claimants, the final version of the Strategic Deals Memorandum, which also included in its Appendix 1 all the email correspondence from EMC quoted in paragraphs 1278 and 1317 above, stated in full:

“Background

As part of the Directors’ continual assessment of the strategic opportunities for the business the Executive management recognized an opportunity to develop an application based sales and marketing initiative. The background to this position was the recognition that there was likely to be a continuing rationalization of IT suppliers to major financial institution customers resulting from both the impact of the global credit crisis and the continuing evolution of hardware/software applications to major multi-national organisations.

In terms of specifics, after a meeting between Guy Chiarello (global CIO) of JPMC and Mike Lynch in Q2, Autonomy was asked to step into a strategic supplier role for JPMC. This has resulted in 2 deals-one in Q2 ($6m) and one in Q3 ($11m) following on from the $10m deal in Q4’08. There have also been a number of senior level conversations between Autonomy and Morgan Stanley again resulting in two sales in Q2 ($7m) and two more in Q3 ($4m) which followed the $18m deal in Q3’08. Similarly significant discussions have taken place between Citi and Autonomy resulting in multiple sales ($22m) in Q2 and Q3 following on from the $20m plus deals in 2008. Discussions have taken place between Sushovan Hussain and John Goaynes (CIO of Deutsche Bank) but a deal has yet to be consummated although a major $10m plus deal is expected in Q4. Finally major companies such as Eli Lilly, Kellog Brown and Root (KBR) and Pfizer have chosen Autonomy as strategic vendor for compliance.

Autonomy has decided to apply, in the case of large ongoing projects, a package approach to this demand for strategic selling at these major institutions. This may mean that, for certain individual components, Autonomy takes terms which may appear less attractive initially but, when considering the bigger picture, are significantly profitable. Autonomy expects this approach to allow it to become a key part of the architecture for companies such as Citi, JPMC, Deutsche Bank, Eli Lilly and Morgan Stanley going forward.

It was noted that IBM, EMC, HP amongst others were increasingly going to market with a combined hardware/software application offerings in the compliance space. For example, EMC acquired a company called Kazeon solely for the purpose of creating an appliance for the compliance space. Another company Clearwell Systems has partnered with EMC and IBM to offer an appliance for the ediscovery space.

Autonomy recognized the importance of being able to demonstrate to substantial multi-national organisations that they could deliver a combined application solution. In order to put together such a solution it was necessary to find an appropriate hardware supplier that could provide:

-Global reach

-Highest quality product reputation

-Industry accepted product recognition

-A management team/culture that was similar to Autonomy and who could recognize the value in a trade association

-Networking and major customer relationships that could be exploited by Autonomy

-The need for reciprocal benefitting from Autonomy’s relationships with existing major customer relationships.

Our management team had previously attempted to establish through a working relationship with Hitachi Data Systems on a basis set out above but found that the Hitachi business culture and speed/flexibility was not compatible with Autonomy. Through Mike Sullivan, worldwide head of Zantaz, our executive management team had a strong connection with EMC and we have begun to develop a relationship with this business to enhance the Autonomy presence in the application space. The strategic partnership with EMC has been led by Bill Scannell (head of worldwide sales at EMC) showing the level of importance afforded to the relationship. As part of the strategic relationship EMC extended the Autonomy OEM agreements by 3 years and extended to new products and Autonomy extended the use of EMC products within its data centres. In addition EMC is spending monies developing an appliance with Autonomy software pre loaded and immediately operational on its hardware. Hitachi have been slower but are actively considering replacing Fast with Autonomy as an OEM as part of the strategic relationship and also creating an appliance which Autonomy would host in its data centres for Hitachi and Autonomy customers. EMC have proven quicker in being able to deliver in the timescale required by the customers (Citi, JPMC, Bloomberg) although Hitachi have started (albeit more slowly) with Morgan Stanley.

Additionally, we consider that there will be further value add opportunities that will arise from developing this relationship further with EMC. We are working on the possibility to move toward an “Intel Inside” type of arrangement with EMC hardware. i.e. the creation of appliances whereby archiving, ediscovery and compliance solutions are offered as a one stop solution to key strategic customers. The key element of this strategy is that by investing significant dollars in the relationship today, Autonomy will “own” the customer for many years yielding multi-million dollars of revenue from each customer.

Q3 Transaction

During Q3 Autonomy entered into a $45m purchase of hardware and additional sales and marketing support. The hardware element represents the purchase of hardware that was sold on to the above mentioned customers. The sales and marketing incentive reflects the payment of $’s for the investment in the relationship and future development of cells as well as joint marketing initiatives.

This transaction was appropriately approved and authorized by Executive management in accordance with the standard business procedures. The hardware component of this transaction was used to supply Autonomy’s own customers (Citi, JPMC, Morgan Stanley and Bloomberg) with equipment for their existing data warehousing and storage functions. In addition Autonomy has sold software for the applications over the past few quarters to Citi, JPMC and Morgan Stanley. Our intention is to use this opportunity to aggressively further exploit software opportunities with these types of organisations.

The revenue recognized on these hardware sales in the Q is $36m. The transactions with each of these customers was appropriately structured so that Autonomy acted as principal to these transactions. The key accounting consideration is the recognition of the $45m of costs, negotiated at arms length by the executive management teams of both companies. [sic] with particular reference to the allocation of costs between COGS and Sales and Marketing. The allocation of $45m between product and other marketing services is not established in the purchase order or invoice from EMC and HDS but has been identified from confirmatory documentation from EMC (refer to email 1 in Appendix 1).

The Autonomy rationale for the transaction was to combine the delivery of hardware to key customers (thereby beginning to develop the recognition of being an application player) together with making an investment in the growing relationship with EMC.

Having established the product cost at between 63% and 43% directly from EMC (depending on product type), the residual cost is for sales, marketing and development efforts. There are significant sales and marketing activities and a series of plans including seminars, trade show stands, customer specific events, sponsorships and incentive payments to the EMC salesforce to market the Autonomy sales. The number of customers and type of customer is very targeted and does not require general advertising. In addition, EMC and Autonomy are obtaining joint meetings with customers and the payment by Autonomy to EMC incentivizes the EMC salesforce to obtain these meetings.

To determine the appropriate allocation of costs between COGS and Sales and marketing the executive management have received confirmation from EMC of the standard reseller margins that it would expect to see in the sale of hardware through a third party. This confirmation was received from Mike Mussulli-Regional Partner Manager (refer to email 2 in appendix 1). The standard reseller discount is 36% for Clarion and 57% for DMX. This has been applied to the sales made by Autonomy this quarter and is computed in the attached spreadsheet (EMC summary Q3 2009 Final.xls.)

The table below-extracted from the spreadsheet-shows the relevant cost of sales based on the standard discounts:

Customer	Product	Discount		COGS
Citi (65%)	DMX	57%	5,193,026.18
Citi	Clarion	36%	4,161,852.82
Bloomberg	DMX	57%	3,065,520.60
JPMC	DMX	57%	4,692,922.39
			17,113,321.98

On this basis Autonomy has allocated $17.1m cost to COGS-representing the standard cost of hardware- with the remaining costs being split between sales and marketing expense and research and development. The sales and marketing amount represents the return of the Autonomy profit of the transaction which it has been agreed with EMC will be reinvested in the relationship in the manner set out above.

The additional payment relates to payment to EMC for the development of the appliance referred to above. Management did not feel that this payment met the definitions of IAS 38 for capitalization and hence has expensed this payment during the quarter.

Final cost allocation

COGS 17,113,322

Sales and marketing incentive 19,509,269

R&D costs 8,860,079

45,482,670

Having reviewed this allocation the directors have concluded that this represents the fair and appropriate split of the costs of this transaction with EMC.”

The alterations from the version sent by Mr Knights the previous evening (so far as not editorial or stylistic) have echoes of Mr Hussain’s earlier briefing note to Dr Lynch (see paragraph 1308 above), and earlier memos he had sent Deloitte, re-ordered and re-formulated to fit into Mr Knights’ revised structure.

Did the Strategic Deals Memorandum spin a false narrative?

The Claimants denounced the final version of the “ Strategic Deals Memorandum” as “replete with falsehoods” (see above). By reference in particular to the underlined passages in the full quotation at paragraph 1324 above, the Claimants submitted that the main thrust of the memorandum was thus to frame Autonomy’s supposed relationship with EMC to be all about the delivery of an appliance (or application solution), and that this was false. They encapsulated their case as to this falsity in their written closing submissions as follows:

“ Whatever semantics Dr Lynch may wish to employ, the basic indisputable fact is that the EMC sales were of pure, standalone hardware - not appliances, not applications, not solutions, and not packages. The sales were of EMC hardware, sold unmodified to customers selected by EMC. Autonomy’s sole contribution was to interpose itself into an existing relationship and to buy the hardware at one price and resell the same hardware to the EMC customer at a lower price.”

In addition, the Claimants submitted that the final version of the memorandum repeated falsities that appeared in the previous versions, giving as an example, the statement in it, for which Dr Lynch had not been able to identify any documentary material in support, that there were:

“ significant sales and marketing activities and series of plans including seminars, trade show stands, customer specific events, sponsorships and incentive payments to the EMC salesforce to market the Autonomy sales. ”

In his oral closing submissions, Mr Rabinowitz dismissed as “spin” the submission in Dr Lynch’s written closing submissions, that this:

“fairly reflected the twin drivers for the hardware sales, namely (i) the need for strategic package sales to major financial institutions and (ii) the need to find an appropriate hardware supplier with which Autonomy could work on an appliance.”

Mr Rabinowitz submitted that this was not what the “ Strategic Deals Memorandum” actually said:

“…rather what it said in the memo is that EMC was in fact developing an appliance with Autonomy software loaded on to it and that it was spending money on this. That….was completely untrue, because EMC had not agreed to develop such an appliance and nor was it spending any money doing so.”

As so often in this case, this presentation of falsity is too black and white: even the passages in the “ Strategic Deals Memorandum” highlighted by the Claimants which predominantly related to the accounting treatment might be said to have had a toe in the truth, and to be more in the nature of exaggerations rather than falsities. Thus, for example:

(1) Both Mr Sullivan’s evidence in the US criminal proceedings and (after much circumlocution) Mr Goodfellow’s evidence^{^[202]} in cross-examination in these proceedings confirmed that EMC was (in the summer of 2009) refining its hardware product at the request of Autonomy in order to make it more suitable, efficient and cost effective for use with Digital Safe and with a view to Digital Safe being embedded in it. Dr Lynch told me in cross-examination that he too had understood and seen emails confirming “that they were configuring hardware for us.”

(2) Mr Sullivan expressly confirmed in cross-examination in the US criminal proceedings that he had further discussions later in 2009 about (a) “building potentially an appliance, so something like potentially shipping a Digital Safe or actually e-Discovery [and] bundling it on their hardware and shipping it…” ; (b) a Digital Safe appliance using EMC hardware and “split cell technology”; and also about (c) getting into packaged servers, as well as “bringing each other into each other’s accounts, linking up our sales teams…”

(3) More generally, Mr Sullivan agreed that Autonomy’s relationship with EMC “was not just hardware reselling, it was more a strategic general discussion of possibilities to work together…”

(4) He also regarded the purpose of the hardware purchases and the development of the relationship as being to “drive some revenue to Autonomy”, which the Claimants seized on as showing that the driver was revenue by selling on the hardware at a loss, but which Mr Sullivan clarified in his evidence in the US criminal proceedings meant getting some of EMC’s hardware customers to buy Autonomy software.

(5) There were also email exchanges, in June 2009, albeit internally within Autonomy, which referred to telephone calls between Autonomy and EMC which culminated in EMC agreeing to configure and provide hard drives which suited Digital Safe. For example, an email dated 26 June 2009 from Mr Wang to Dr Menell referred to Autonomy having:

“had a breakthrough with EMC today in having them agree to sell us essentially hard drives without their fancy software which is irrelevant for Digital Safe…”

(6) This “custom configuration” was also referred to in emails from Mr Sullivan. This inevitably required investment by EMC: as Dr Lynch said when it was suggested that EMC had not agreed to spend money, “…you can’t do it without spending money”.

(7) Mr Hussain’s written submissions also drew attention to the evidence in an email chain dated 23 to 25 November 2009 (under the subject heading “Re: Appliance” ) of further technical discussions between Autonomy and EMC with the view to the harmonisation of EMC hardware with Autonomy products and the development of an Appliance.

(8) These later discussions involved the relevant departments of each company and continued in subsequent emails that year and the next, including an email chain in February 2010 in which Mr McLaughlin of EMC spoke encouragingly of the integration of Autonomy software into EMC products as a “win-win”.

(9) In those circumstances, Mr Hussain made the further point that:

“the very worst that could be said of the Strategic Deals Memo is that it suggested that steps were being taken in October which were in fact taken the very next month…it is highly improbable that [Autonomy] manufactured the story of EMC’s commitment to an appliance programme to achieve a spurious accounting treatment only for the lie to come true in the next quarter.”

It might also be said that the “ Strategic Deals Memorandum” was also in terms aspirational, casting the relationship with EMC as offering future opportunities. For example,

(1) It stated:

“Additionally, we consider that there will be further value add opportunities that will arise from developing this relationship further with EMC. We are working on the possibility to move towards an “Intel Inside” type arrangement with EMC hardware, i.e. the creation of appliances whereby archiving, ediscovery and compliance solutions are offered as a one stop solution to key strategic customers. The key element of this strategy is that by investing significant dollars in the relationship today, Autonomy will “own” the customer for many years yielding multi-million dollars of revenue from each customer.”

(2) Similarly, with specific reference to the Q3 2009 transactions themselves, the memorandum emphasised both the rationale of the immediate deal and the future prospects:

“The Autonomy rationale for the transaction was to combine the delivery of hardware to key customers (thereby beginning to develop the recognition of being an application player) together with making an investment in the growing relationship with EMC.”

Further, and in contrast to earlier drafts, some of the limitations on the evidence available to substantiate the accounting treatment of part of the costs as marketing expenses, and capitalising a proportion, were expressly recognised (even if the positives may be said to have been accentuated). It was recorded:

(1) As to allocating part of the costs to marketing, that:

“The allocation of $45m between product and other marketing services is not established in the purchase order or invoice from EMC and HDS but has been identified from confirmatory documentation from EMC (refer to email 1 in Appendix 1)” and

(2) As to capitalisation of what Mr Hussain had wanted to characterise as development costs, that:

“Management did not feel that this payment met the definitions of IAS 38 for capitalization and hence has expensed this payment during the quarter.”

Nevertheless, all this said, and though (like the hardware reselling strategy itself) some of its constituent elements appeared rational and could be justified, and others were expressly aspirational and not ostensibly false in being so, I have concluded that the “ Strategic Deals Memorandum” was, in the round, seriously misleading.

The fundamental realities which the Strategic Deals Memorandum disguised were that:

(1) As Mr Hussain acknowledged , Autonomy and EMC never did form a “strategic partnership”.

(2) There is no evidence that their discussions for the development of an appliance moved beyond the discussion stage. The Strategic Deals Memorandum papered over Autonomy’s repeatedly unsuccessful efforts to obtain any actual agreement or commitment from EMC (beyond agreeing to sell its hardware through Autonomy to customers (largely) in common). Mr Sullivan’s vague description of the relationship (see paragraph 1330(3) above) as comprising “more a strategic general discussion of possibilities to work together…” was really the extent of it, save of course for the purchase and reselling itself.

(3) There was some evidence that EMC spent some money on reconfiguring its hardware to enable easier and more effective use of IDOL software, but no evidence of any definite programme or commitment on the part of EMC.

(4) The extent of the sales, marketing and development activities in fact taking place at the time was ( as I find) overstated considerably^{^[203]} ; and any plans for the future were general in nature and not supported by documentary evidence or any confirmation of commitment by EMC.

(5) The emails from EMC attached to the Strategic Deals Memorandum did not confirm the allocation of $45 million between product and other marketing services, nor even did they confirm that the sale price covered both hardware costs and marketing and support: the most EMC was prepared to do was speak in general terms which, especially in the context of previous exchanges (not referred to in or attached to the memorandum), demonstrated a refusal rather than agreement to give any such confirmation.

More particularly, the “ Strategic Deals Memorandum” covered up the lack of any evidence such as Deloitte had originally asked for “to substantiate a $25m marketing element” (see paragraphs 1278(5) and 1307 above). It advanced and relied on a proxy (the “residual method”) to persuade Deloitte to approve an accounting treatment which perpetuated an impression of there being commitments on the part of EMC to spend the ‘delta’ on marketing for Autonomy and the joint development of an appliance for which management knew there was no substance, and when in fact EMC would have denied any, as Mr Sullivan had explicitly warned in his email of 2 October 2009 (also quoted in paragraph 1278(10) above):

In my judgment, the picture presented in the Strategic Deals Memorandum of a “strategic partnership” with EMC involving joint marketing, joint development of an appliance and cost-sharing of both via a “premium” paid by Autonomy was false. EMC never committed to any of that. Had it done so, the document would almost certainly have been unnecessary.

The Defendants argued that these were matters which principally, and certainly most directly, affected the issue of cost allocation, and that it was essentially an accounting judgement as to what part, if any, of the costs could be allocated as marketing expenses. However, for so long as the principal means of achieving the overall purpose could plausibly be presented as being the pursuit and formation of a partnership, Autonomy did not need to show any more measurable result to justify treating the programme as calculated to protect and promote its core software business: a partnership to develop an appliance in which to embed Autonomy software could fairly easily be (and was) presented as both a legitimate objective and a legitimate part of the software business. If Autonomy could only rely on the alleged benefits of the hardware reselling in terms of generating new software business to justify the programme, it was bound sooner rather than later to provide some evidence of real linkage.

Furthermore, whilst the issue of cost allocation would usually be a matter of judgment, I do not accept that it was so in the particular context in which the allocation of the costs of the hardware purchases and loss-making sales was determined in this case. That allocation was part of the means of concealing the hardware reselling and attenuating, and thereby disguising, its real effect. Unexplained costs would be likely to be queried by analysts; and any material erosion in gross margin would be a concern to the market and undermine the success of the strategy. In this case, what was involved was not a judgement; it was a necessary part of the scheme.

The false presentation was, in my judgment, driven by three complementary imperatives:

(1) One was that the programme should appear to be justified by the prospect or actuality of a partnership for the sale of Autonomy software without the need to show any other measurable benefit.

(2) A second was that the effect on key metrics of losses incurred in the hardware reselling strategy should be attenuated.

(3) The third was to hide or disguise the programme so as not to expose it to enquiry and/or undermine its utility as a surreptitious means of maintaining the appearance of continuing organic growth by including hardware revenues in forecasts and drawing on them as and when required.

In my judgment, this was not honest judgment but dishonest expedient.

Dr Lynch ’s knowledge of falsity of the Strategic Deals Memorandum

As will already be apparent, Dr Lynch did not dispute that he was aware and indeed involved in the presentation of the purpose of the hardware reselling strategy and EMC’s participation in it. For the reasons given above, I have concluded that that presentation was misleading. But even if I am wrong about that there can be no real doubt that the presentation of the nature of the arrangement with EMC was misleading in any event.

Dr Lynch did dispute that he was involved in or aware of the efforts in exchanges with EMC to find some written acknowledgement to justify the allocation of a large proportion of the costs to marketing expenses; and he insisted that he was not aware of any impropriety in this context. As far as he was concerned, he told me, “what Autonomy was trying to do was get the right accounting…the accounting…reflecting the reality of the situation” (as quoted previously in paragraph 1277 above). He added, when further pressed in cross-examination (with particular reference to the emails sent to EMC):

“What is happening at this time - and I’m not involved but I’ve seen the emails - is that a combination of Mr Chamberlain, Mr Sullivan, EMC and ultimately Deloitte are trying to get the right answer. It is a complex situation and lots of people are involved in that and they work through it and they ultimately come up with an answer that they’re all agreed on and that’s what I’m relying on when I go forward.”

He emphasised also that, as CEO, he could not be involved in the details of this sort of matter. He relied on what he was told by his lieutenants (hence needing prepping by Mr Hussain before meeting Deloitte about the Strategic Deals Memorandum , for example). He said this in cross-examination:

“So just to keep some perspective here, I’m running a FTSE 100, I spend most of my time not dealing with this sort of thing. I’m not on top of it. I have a finance department which does all of this that has very good people in it…”

Some support for Dr Lynch’s oral evidence that though involved in the explanation of the genesis and rationale of the programme he was not involved in the issue as to the accounting allocation of its costs (“the COGS issue”), is provided by the emails from Mr Hussain prepping Dr Lynch as referred to in paragraph 1308 above. Thus, as recorded in paragraph 1318 above when Mr Knights emailed Dr Lynch asking to discuss the issue as to allocation of the costs of the EMC purchases and resales on 14 October 2009, Dr Lynch immediately forwarded the email to Mr Hussain, which prompted Mr Hussain’s second email of 14 October 2009 (referred to in paragraph 1318 above). That email was plainly intended feed Dr Lynch with what to say. As to its content, Dr Lynch was pressed in cross-examination to accept that he knew it was “pure fiction” ; he rejected this and told me:

“A. Well, at this time I am relying on emails and information from other people, but, no, I understand seminars, meetings and customer events were happening. The incentives to the sales force is sort of self-evident. And EMC - I’d seen emails that EMC confirmed that they were configuring hardware for us. So I’m reasonably confident that that is accurate.

Q. Had you attended any of these seminars, meetings, customer events?

A. No, they were in the US, because, if you remember, the arrangement was for New York; I ’m based in Cambridge in the UK.”

Nevertheless, though Dr Lynch’s answers were persuasively delivered, I have concluded that I cannot accept that Dr Lynch was not involved in the discussions relating to the COGS issue. It seems to me that the issue as to the rationale of the programme and the issue of the accounting treatment of its costs were and are intertwined, despite Dr Lynch’s efforts to cordon off the one from the other.

I have reached this conclusion principally on the following grounds:

(1) Although undoubtedly Dr Lynch was busy , and he was not an accountant, he had time and accounting awareness enough to know that the loss-making hardware reselling strategy had a serious potential disadvantage in terms of its adverse effect on Autonomy’s ‘bottom line’, which if it could properly be attenuated should be so.

(2) As, on his own evidence, the hardware reselling strategy was his concept, I think it unlikely that he would not have been interested in its accounting treatment, even if its rationale was as he depicted: and if the rationale was a pretext, then all the more reason to be engaged to disguise it as much as possible.

(3) Furthermore, it is not disputed that, unless advised that it was a requirement to disclose, Dr Lynch was very keen to ensure that the hardware reselling strategy was not disclosed: and it was always obvious that material increases in COGS would be likely to lead to inquiry which would in turn reveal the nature and extent of the programme. This was a concern which it seems to me would have been likely to have prompted his interest and involvement.

(4) Mr Knights seems to have thought that Dr Lynch was involved in devising ways to deal with the COGS issue: it will be recalled that in his email (using his wife’s email address) of 13 October 2009 to Mr Hussain and Mr Chamberlain (see paragraph 1313 above) Mr Knights had referred to “Mike L’s ideas” in relation to the “key area” of “quite what ‘sales and marketing’ type of stuff or further development stuff EMC might provide you.”

(5) Albeit not quite contemporaneous, there is documentary evidence, in the form of an email dated 18 July 2010 from Mr Hussain to the Deloitte audit team (copied to Mr Chamberlain), recording Dr Lynch’s “strong views” on the issue of COGS in the context of hardware sales; and although the evidence is from a later date, there is no reason to suppose that these strong views only crystallised in 2010, rather than when first relevant.

(6) More generally, the conclusion seems to me to follow from my conclusion that the allocation of costs to ‘Sales and Marketing expenses’ was all part of the strategy; and Dr Lynch was well aware of that.

My conclusion has been reinforced by my other conclusion that Dr Lynch was a knowing party to the exaggeration of the amounts properly referable to the launch and marketing of SPE, which were treated and accounted for as sales and marketing expenses as another stratagem to cover up the substantial costs in Q3 2009 of the hardware reselling strategy.

In summary, therefore, I have concluded and find that both Defendants were involved, not only in elaborating the extent of the relationship with EMC, but also more particularly in the false presentation to Deloitte of the COGS issue in order to attenuate (by moving part of COGS to sales and marketing) the adverse effect on Autonomy’s gross profit and gross margin figures resulting from a very considerable expansion of the hardware reselling strategy.

Assessment of the effect of the “Strategic Deals Memorandum”

The question then is how Mr Knights came not only to accept, but also to help present, the accounting treatment based on the “residual approach” (see paragraph 1280 above), and later to require no clear disclosure.

This is not easy to fathom, given that it seems to me to be clear that Mr Knight, Mr Knights and Mr Welham all fairly quickly identified that Mr Hussain and the finance department had exaggerated the extent of Autonomy’s relationship with EMC. That is evident from Mr Knight’s email of 9 October 2009 to Mr Knights and Mr Welham after receipt of what, by then, was the best that Autonomy appeared to be able to offer by way of substantiation, which was EMC’s email of 18 September 2009 (and see paragraphs 1278 and 1282 above)^{^[204]} . Further, of course, the fact that neither Mr Knights nor any other person who was in the Deloitte audit team apart from Mr Welham, gave evidence in these proceedings, has made the task of unravelling why Mr Knights, and subsequently Mr Mercer, accepted the unacceptable even more difficult.

There may have been a combination of factors, including:

(1) Until tested sceptically against email exchanges demonstrating a very different objective, the overall rationale of the hardware reselling strategy appeared broadly credible. Except that he saw and considered the final emails between Autonomy and EMC as referred to above, there is no evidence that Mr Knights saw the emails from Mr Hussain and Dr Lynch focused entirely on “revenue revenue revenue” ( see, for example, paragraphs 906 to 911 above) or other internal Autonomy emails conveying the same focus. He was thus not confronted with material to excite further scepticism; and it seems plain that his mindset was always to seek to assist his client to overcome apparent inconsistencies, rather than sceptically to test the reasons for them. That mindset appears to have distracted him from a more sceptical review of the true reasons for the resort to the “residual approach”^{^[205]} and management’s disproportionate determination (relative to the comparatively small amount of costs to which the accounting treatment related) to mitigate the adverse effect of the programme and reduce its visibility.

(2) The process of collaborative engagement with Mr Hussain and Mr Chamberlain in working up a document which they were to present as their own caused him to become parti pris. Instead of testing Deloitte’s objectives, Mr Knights had become too close to question them: in his own word, “wordsmithing” replaced sceptical review. Caught up in presenting on Autonomy’s behalf the narrative of the overall purpose of the programme, he failed to assess objectively the significance of the continuing lack of any of the evidence of commitment on the part of EMC which his own audit team had required, and of Autonomy’s resort to a proxy in the form of the residual approach. He convinced himself of what he had helped to write, and his auditing team followed suit.

(3) Last, but I suspect not least, it seems more than likely that Mr Knights felt considerable pressure to retain for Deloitte’s office in Cambridge (of which he was head and which seems to have committed most of its resources to this one client) a valuable account in respect of a FTSE 100 company which was a leading light in a high-profile sector.

Whatever the reason, ultimately the more important point for the purposes of this case is the fact of his involvement, and the effect of it. In my judgment, there can be no real doubt that Mr Knights ceased to be a sceptical auditor and became an inventive wordsmith, who was even prepared to disguise his own contribution, which included using (a) his wife’s email to send his draft; (b) emphasising in his covering email attaching it the need for it “to come from you to us”; and (c) artfully describing the final version as “the client produced memo” when sending it to the Reviewers the next day.

Although (as will be seen) in future quarters Deloitte would only accept a lower proportion of the costs of the programme being accounted for as Sales and Marketing expenses, Mr Knights’ contribution to and support for the narrative in the “ Strategic Deals Memorandum” had an important influence in overcoming Deloitte’s original scepticism, and in leading them to treat the issue of cost allocation as purely one of accounting judgment, rather than as one which invited questions as to the use and propriety of the programme in the manner in which it came to be implemented.

More generally, the “ Strategic Deals Memorandum” , reinforced by Mr Hussain’s Quarterly Notes (see below), in substance shaped Deloitte’s view of the hardware reselling strategy from then on until at least Q2 2010 when the arrival of Mr Mercer and the growing obviousness that the hardware sales had become “business as usual” prompted Autonomy to provide the further support of the Linkage Analysis.

Mr Knights’ conversion and support seems to me likely also to have had some influence on the attitude of the reviewing agencies (“the Reviewers”) whose involvement was stipulated by Deloitte as part of its internal oversight and control of complex audits. The Reviewers were its Professional Standards Reviewer (“PSR”, at that time in Q3 2009, Ms Lisa Bennett)^{^[206]} , Engagement Quality Assurance Review (“EQAR”, at that time, in 2009, Mr Stuart Henderson^{^[207]} ), and Independent Review Partner (“IRP”, in 2009 for Autonomy, Mr Robertson^{^[208]} ). Deloitte’s final approval depended on their sign off.

Again, the determination of what persuaded them is compounded by the fact that like Mr Knights, none of the Reviewers gave evidence, and there seem to me to be gaps in the documentary evidence provided (despite its immensity). However, what is apparent is that:

(1) Mr Knights gave them very limited time for review and response: after sending them the final version at 11:40 on 14 October 2009 he not only chased for a response that day but then chased again at 13:54.

(2) Although Mr Knights stated that he was happy to discuss with each of them individually, or collectively and Mr Welham’s evidence was that their practice was to perform a collective process together, there is no record of any such collective decision, perhaps because of time constraints.

(3) There was a one-line email (timed at 14:23 on 14 October 2009 and apparently in response to Mr Knights’ chaser at 13:54) from the then PSR, Ms Bennett, stating “I am OK with it” (which supports the inference that there was no collective process).

(4) There is no record (or at least I have no record of having been shown one, and I have found none in the Trial Bundle), of any written response from the EQAR or the IRP.

(5) It seems that by 16:30 Mr Knights had received their response(s), since he emailed Dr Lynch and Mr Hussain at that time on 14 October 2009 stating that Deloitte had considered the “cost analysis on the transactions we have been discussing” and were “satisfied that your paper appropriately sets out the accounting”, but subject to three caveats (which, since Mr Knights had not previously mentioned them, I assume reflected further consideration with the EQAR and/or the IRP).

Again, the more important matter than trying to discern in their absence what persuaded them is that the Reviewers signed off on the basis of the three caveats and a further cautionary note that the board should consider what further disclosure might be appropriate.

The caveats appear to me to signify that the approval was somewhat hesitant, and that the Reviewers and Deloitte would require more convincing detail about the allocation of costs in future transactions. Their central or common theme was to treat the Q3 transactions as something of a one-off and to leave open the prospect of more rigorous review in the event of further hardware deals. In detail they were that:

(1) The split between COGS and marketing expenses had been considered for this “initial transaction” and “would not necessarily recur in subsequent deals” ;

(2) The allocation for further transactions “would need to be assessed on its own merits” ;

(3) The expectation would be that any future deals “would not necessarily include the same level of marketing costs” and

“importantly it will be necessary to have a more itemised breakdown of component parts of any purchase between Hardware and other items.”

Mr Knights set out these caveats when reporting back to Dr Lynch and Mr Hussain by email timed at 16:31 the same day (14 October 2009). He ended the email with the following, which I take also to have reflected part of the basis on which the Reviewers acquiesced:

The issue of disclosure in Q3 2009

That leads on to the issue of disclosure, and the way that Mr Knights was persuaded, perhaps against his initial and better judgement, to agree to minimal disclosure and only tinkering changes in the Q3 2009 Accounts. This, like the “ Strategic Deals Memorandum”, seems to me to have set a pattern, with Deloitte repeatedly raising the issue of further disclosure but then either deferring to management as being a matter of commercial judgement, or settling for bland or cryptic sentences which were only decipherable by those in the know already and fell a very long way short of the transparency about the “quantum and nature of these sales” which Deloitte’s Report has suggested should be “considered” for Q3 2009 and had warned could well be “necessary” in the full or half-year accounts.

Although I have set out previously in chronological sequence what was said about hardware in the Yearly and Quarterly Reports throughout the Relevant Period (see paragraphs 1187 to 1261 above), an elaboration of the way Deloitte’s initial concerns were dealt with in Q3 2009 seems to me to illustrate how Deloitte was led to agree to so little disclosure then and thereafter. Thus:

(1) When provided by Mr Kanter (on 15 October 2009) with his suggestion for the way the new Supplemental Metrics might be presented^{^[209]} , Mr Knights’ response expressed concerns including the absence of anything to show the hardware sales: he put this in the form of a question “…I’m not sure where hardware sales would sit in your table?”. But he softened his other concerns with a final sentence:

“It might be that with some word smithing it can be achieved.”

(2) He amplified his concerns in red type on a revised draft “Updated Press Release” which he circulated within Deloitte (to Messrs Welham and Knight and Ms Anderson) on 15 October 2009 which included the following:

i. A comment next to a bullet point in the draft highlighting “Strong organic IDOL growth of 15%” which read “check this calculation excludes impact of hardware sales” ;

ii. Comments next to a description of revenues for the quarter having “totalled $191.6 million, up 51% from $127.1 million for the third quarter of 2008 due to strong organic growth” which (a) queried whether “due to” should be amended to “including” and (b) stated “but hardware sales are not organic” .

(3) In a follow-up email to Mr Hussain on 16 October 2009, setting out Deloitte’s role and responsibility as regards the “front-end of the accounts”,^{^[210]} Mr Knights noted at the end (the underlining is mine):

“Can we have a detailed breakdown on how the figures are compiled. My biggest concern will be that hardware sales were neither IDOL based or organic !!

Let’s see the analysis and work out how to sensibly disclose.”

(4) Yet none of these concerns was reflected in an updated draft press release sent to the Audit Committee prior to its meeting the next day (16 October 2009). That draft, which Mr Welham also sent to Mr Robertson and Mr Henderson (cc Ms Anderson) to consider in time before anticipated release on 19/20 October, set out under “Supplemental Metrics (not reviewed)” the following, none of which gave any hint that the source of a proportion of the revenue was hardware sales:

“Supplemental Metrics (not reviewed)

Autonomy is supplying supplemental metrics to assist in the understanding and analysis of Autonomy’s business

Software sales including hosted and OEM……………$125m

Service and support revenues……………..……………$9m

Deferred revenue release (primarily maintenance).. .$58m

IDOL OEM derived revenues…………………………..$24m

IDOL Organic Growth………………………………….$15%”

(5) In response, Mr Henderson (in an email to both Mr Welham and Mr Knights and copied to Mr Robertson, sent some 20 minutes after Mr Welham’s email) did not focus on the Supplemental Metrics but expressed considerable and more general disquiet:

“As anticipated I am deeply concerned by the total lack of reference to the fact that nearly 20% of their Q3 revenues representing a major strategic change in the nature of their business attracts no comment….They don’t even seem to mention the customers to whom these highly material hardware sales have been made. I will take a fair amount of convincing this is appropriate.”

(6) Given the first phrase, it seems likely that Mr Henderson had voiced this concern earlier, or at any rate it would not have come as a surprise. Mr Knights replied immediately to say that the matter would be discussed at the Audit Committee Meeting later, and he would then revert. Mr Knights followed up again some 30 minutes later in an email to Mr Henderson, Mr Robertson and Mr Welham, it seems likely after a discussion with Dr Lynch or perhaps Mr Hussain and/or Mr Chamberlain, stating:

“Wording being now put into Mike ’ s quote….

during the quarter we saw some of our customers promote Autonomy to strategic supplier status. This led them to adopt a broader set of our solutions in a number of significant deals.

We should remember that the 36 is split into 3 deals of around 9-11m I think - Moving the battle ship [sic] slowly-this is bound to go round and change a few times…”

(7) As Mr Knight pointed out after Mr Knights had forwarded the same email to him, a problem with the proposed wording was that it “talks about the strategic supplier status but doesn’t talk about the nature of the deals. It could be read as they have bought more IDOL”. Mr Knight then suggested that:

“If they do not want to talk about hardware then the solution could be to get them to remove comments about organic growth and idol growth.”

(8) Later that day (at 17:03), Mr Robertson also replied by email to Messrs Welham, Henderson and Knights (cc Ms Anderson) making two points:

i. With reference to the proposed statement in the gross margin section that “The unexpected demand for our new product programme had a small depressing effect on gross margins”, he asked, “Do we think this explains things sufficiently?” and suggested that the drop in margin was considerable and not done justice by the description;

ii. In the same connection, he also posed the question at the heart of things:

“What’s the sensitivity about being more transparent on this score? If it’s a strong strategic move for them, why wouldn’t they want to explain this? I’d have thought the analysts will be bound to ask a lot of questions about it given the results look quite different this Q (usual big increase in revenue but comparatively small increase in profit)”;

iii. With reference to the statement in the draft “We do not expect this to be a trend” , he made the point that he had the impression “from our various conversations over the last few days that they were planning on doing more of this” and posed the question also at the heart of things: “Can they really make this statement?”

iv. He ended with the comment that he would be interested to hear how the discussions at the Audit Committee “have moved on the transparency around these transactions.”

These comments went to the root of the issues which have arisen since. Yet apart from prompting further apparently inconclusive discussion at the Audit Committee about the description “IDOL organic growth” (which was not minuted) and the debate I have mentioned previously as to how that might be altered, there is nothing in the evidence before me to suggest that the full discussion about the need and rationale for greater transparency ever took place.

It seems that the issue of transparency became subsumed, and indeed lost, in a discussion (without any evident further involvement of Deloitte’s review partners) about the Supplemental Metrics and the description of organic growth.

(1) The nature of the discussion is illustrated by an email from Mr Chamberlain to Mr Hussain and Mr Kanter on 16 October 2009 stating that:

“The key issue with the auditors seems to be around the use of the word IDOL.

Strong organic growth and strong organic IDOL growth - to them the former includes hardware sales and the latter does not

Product including hosted and OEM and IDOL product including hosted and OEM-same point

IDOL organic growth and organic growth

Whilst we may find a way to get there on organic growth through allocations they are going to really struggle with including hardware within the description of IDOL product or IDOL organic growth.

Battle lines have been drawn.”

(2) The “battle” was not fierce. Its upshot was that (a) Autonomy (with the intervention of Dr Lynch) agreed to remove the prefix “IDOL” when describing the various categories of revenue in the “Supplemental Metrics” (so that, for example, “IDOL Organic growth” became “Organic growth”) and (b) after further exchanges, Autonomy overcame Deloitte’s concerns about including hardware revenues within “Organic growth” with the argument that all revenue growth should be treated as organic unless its source was an acquisition.

(3) Mr Welham stated in his witness statement (without further elaboration) that:

“Ultimately, we agreed that growth in hardware sales was organic, in that it did not derive from the acquisition by Autonomy of a pre-existing business.”

(4) In the final tie-through version of the Q3 press release checked by Deloitte on 19 October 2009 both the statement “Strong organic growth of 15%” and the statement “Revenues for the third quarter of 2009 totalled $191.6 million, up 51% from $127.1 million for the third quarter of 2008 including strong organic growth” which Mr Knights had expressed concerns about in his email of 16 October 2009 (see paragraph 1361 above) were ticked off by Deloitte.

(5) As previously noted in paragraph 1259(2)(i) above, and apparently heedless of Mr Robertson’s query as quoted in paragraph 1361(8)(iii) above, in the final version of the Q3 2009 Report an erosion in gross margin was ascribed to unexpected demand for SPE and Quick Start, which was not expected to be “a trend” and did not mention the hardware reselling strategy, which was in fact the principal drag on gross margins and was expected to be a “trend”; and the only reference which the initiated might have recognised related to hardware was the cryptic sentence also already quoted that:

(6) There is no sign of the concerns expressed by Mr Henderson and Mr Robertson having been addressed; as far as the evidence before me goes, they appear to have sunk without trace.

(7) Thus, Autonomy’s management had got its way with Deloitte, overcoming Mr Knights’ earlier concerns and the Reviewers’ objections, for the price of a false explanation, a cryptic sentence and the removal of the prefix “IDOL” before the description of the categories of revenue. (Even that retreat was soon made good: in subsequent quarters and half and full year accounts, the prefix was restored.)

As my earlier chronological review of the Annual and Quarterly Reports (see paragraphs 1187 to 1261 above) demonstrates, none contained any further disclosure of the hardware reselling strategy, except (and as only the initiated could have surmised) (a) an even more cryptic explanation for an erosion in gross margins in Q2 2010 as being “due largely to changes in the sales mix” and (b) a very opaque reference in Q4 2010 to “ package solutions, constructed of services, hardware and software, such as Arcpliance” .

Quarterly Notes prepared by Mr Hussain/Mr Chamberlain

Although Dr Lynch had been insistent that what became the Strategic Deals Memorandum should not be restricted to justifying the programme with EMC, its principal focus had been on the Autonomy/EMC relationship as at Q3 2009. As the hardware reselling strategy continued in subsequent quarters its results and justification required confirmation and refreshment. The Claimants contended that the false and misleading description of the hardware reselling strategy in the Strategic Deals Memorandum was perpetuated, reinforced and as necessary modified throughout the Relevant Period in Quarterly Notes (which were marked as having been “Prepared by: Sushovan Hussain, Steve Chamberlain” ).

These “Quarterly Notes” contained an analysis of the quarter’s results which were circulated to the Audit Committee (and also to Deloitte), usually a day before their quarterly meeting, as part of the pack of material which included (a) such a memorandum (b) Deloitte’s report on its review of the financial statements and (c) a draft press release for the quarter.

The Claimants alleged that in these Quarterly Notes, Mr Hussain consistently misled Deloitte and the Audit Committee about the nature of Autonomy’s relationships with its hardware suppliers, and the purpose and extent of the hardware reselling strategy; and that Dr Lynch, who in every quarter was sent a draft in advance, knew that Deloitte and the Audit Committee were being misled in this way, but said nothing.

The Claimants alleged further that the fact that Mr Hussain and Dr Lynch perceived that they needed to mislead the Audit Committee and Deloitte in this way gives rise to the inference that they knew that (a) the accounting description and treatment of the hardware sales they had adopted could not be justified except on the basis of a false description of their context, nature and extent and (b) if the Audit Committee and/or Deloitte had been told the truth, they would have objected to the hardware reselling strategy and/or the accounting treatment of its costs, and required its disclosure to the market.

In other words, the burden of the Claimants’ case in this regard is that the Defendants’ resort to the misrepresentation to Deloitte and the Audit Committee of the extent of the hardware sales and the nature of Autonomy’s relationships with hardware suppliers and the hardware reselling strategy (a) betrays the real purpose and impropriety of the sales and the programme, (b) exposes as futile their claim to have relied on the approval of their published information by Deloitte and the Audit Committee, and (c) damns conclusively the Defendants’ relentless efforts to ensure that the extent of the sales and the nature of the programme were never disclosed to the market.

The Claimants’ written closing submissions analysed in some detail both Mr Hussain’s Quarterly Notes (referred to in the Claimants’ RRAPoC as “memoranda” ), and Deloitte’s own quarterly reports which were likewise provided to the Audit Committee every quarter. They also cross-examined Dr Lynch on these documents, which he acknowledged he would usually review cursorily but not consider in detail. He made the point also that he did not attend meetings of the Audit Committee; and could not provide any evidence of any discussions.

It will be necessary to consider various of Mr Hussain’s Quarterly Notes and Deloitte’s Quarterly Reviews in some detail. But as an overview, (a) the Quarterly Notes served to reinforce, elaborate and, as circumstances changed, refresh what had been said in the Strategic Deals Memorandum (b) Deloitte’s own Reviews appear to have borrowed in relevant part from the Quarterly Notes but tended to be more restrained in their description of the purpose and effect of the hardware sales and (c) the Audit Committee were largely unquestioning of what they were told, and content to rely on Deloitte.

As to (c), the evidence showed that the Audit Committee, having satisfied themselves that the segmental analysis confirmed that Autonomy had only one operating segment, being its IDOL-based software business, would have been troubled only if the hardware sales appeared to be of such nature, extent and materiality as to undermine that conclusion and amount to a separate business.

As it was, that never became a real concern, and a flavour of their attitude emerges from the evidence of Mr Robert Webb, who was not a member of the Audit Committee but as Non-Executive Chairman of Autonomy from May 2009 until its acquisition by HP in 2011, occasionally attended its meetings. Although he could not recall the context , and it may not have been at the Audit Committee meeting to consider the Q3 2009 Reports itself ^{^[211]} , Mr Webb related in his witness statement, on which he was not cross-examined, that:

“I do not recall the discussions around the disclosure of hardware being particularly heated. There was a discussion and the conclusion was to account for hardware sales in whatever way Deloitte said the company was to account for it.”

How the hardware reselling strategy was presented in the Quarterly Notes

The focus of the Quarterly Notes and Quarterly Reviews in dealing with the hardware reselling strategy altered over time. In my view, the following themes and changes of emphasis emerge:

(1) In Q3 2009, the embellishment of the depiction in the Strategic Deals Memorandum of the relationship between Autonomy and EMC, principally with a view to justifying the accounting treatment of the costs of the hardware sales in Q3 2009 .

(2) From Q3 2010, and especially from Q1 2010 after EMC had withdrawn, the recasting of the justification of the hardware reselling strategy towards its “customer-facing” aspects, and “strategic package sales” to meet all their IT needs, with a view to justifying the treatment of hardware sales as an integral part of the software business and continuing to justify accounting for part of the costs of the sales as sales and marketing expenses.

(3) The introduction in Q2 2010 of the notion that such “strategic package sales” had already been “ flagged ” to the market, in response to increasing concern from Deloitte about the absence of disclosure, with a view to deflecting any requirement for further disclosure.

(4) In tandem with the Linkage Analysis, increasing focus from Q2 2010 onwards on the line that hardware sales had generated significant new software business for Autonomy, and that there was a “strong linkage” between hardware sales and “highly profitable software sales” , with a view to addressing increasing concern on the part of Deloitte that the hardware reselling strategy had become “business as usual”.

(5) The disguise of the various stratagems which Autonomy came to adopt to reduce the effect of loss-making sales on accounting metrics, and thereby to support efforts to prevent discovery and disclosure of the extent and nature of the hardware reselling strategy.

These themes are elaborated and illustrated below by reference to individual Quarterly Notes and Reviews.

Mr Hussain’s Q3 2009 Quarterly Note

The final version of Mr Hussain’s Quarterly Note for Q3 2009 was sent by Mr Kanter to each of the members of the Audit Committee (namely, at that time, Mr Richard Perle, Mr John McMonigall and Mr Barry Ariko) late in the evening (9:54 pm) on 15 October 2009 before the meeting next day. Those to whom Mr Hussain’s Quarterly Note was circulated included Mr Hussain and Mr Chamberlain. (As I describe below an earlier draft, in a different form, was sent to Dr Lynch and Mr Hussain by Mr Chamberlain at 16:23 on the evening of 14 October 2009.)

As became characteristic, Mr Hussain’s Q3 2009 Quarterly Note embroidered on the “ Strategic Deals Memorandum” in a rather expansive way. Four facets of this embroidery were particularly striking in this quarter.

First, what had been described in the “ Strategic Deals Memorandum” as “continuing rationalisation of IT suppliers to financial institutions” was embellished considerably into an assertion that:

“… These organisations are restricting their key suppliers to 6 or 7 companies comprising the usual suspects, Cisco, Microsoft etc. Because of the strategic nature of Autonomy they have asked us to assume the last of these slots. In doing so, however, this has pushed EMC, a major supplier of storage, out. It should be noted EMC’s business with these organisations is very large. In order to allow all parties to accept this outcome the companies have asked Autonomy and EMC to partner closely together. This close tie has given Autonomy the scale the banks require and has given EMC the security to acquiesce to the arrangement. Obtaining this strategic supplier status we believe will be very valuable to Autonomy in the coming years.”

As to this:

(1) The Claimants submitted that almost every part of this was false, and none of it was supported by the contemporaneous documentation.

(2) Thus, they said, it was not the case that EMC had been “ pushed out ” by Autonomy; nor was Autonomy appointed as a panel supplier to any of these organisations. On the contrary, the hardware sales, through EMC, had come about by Autonomy approaching EMC: it was EMC that was bringing the hardware deals to Autonomy, and it was EMC that was selling the hardware into its customer base.

(3) The passage was introduced after Dr Lynch had been provided with the proposed last draft (which did not include it). When cross-examined, Dr Lynch denied having had more than a “quick look” at the draft, but the Claimants submitted and I accept that the most likely explanation of the change is input from him, especially since the contribution reflects previous amendments he had suggested to pre-final versions of the Strategic Deals Memorandum .

(4) Dr Lynch sought to support Mr Hussain’s description: but he had no documentary evidence to substantiate any of it. Dr Lynch maintained that the companies had indeed indicated that there was not room for both EMC and Autonomy in their supplier lists, and encouraged them to collaborate. He instanced especially a conversation with a Mr Chiarello, whom he described as JPMC’s CIO, and another person whom he described as Mr Chiarello’s right-hand man, a Mr Feinstein, who he said “were dealing with Autonomy and EMC, sometimes even together.” His evidence was that:

“…there was a discussion at the time that the way that it was played by JPMC was that we were both vying for the final place and that if we could put together this kind of an arrangement then it would be workable for both of us. That was what was presented.”

(5) There was no documentary record of this. Even though I would accept Dr Lynch’s dismissal as “just not realistic” the suggestion that this sort of conversation would have been minuted (and it is also fair to note that the draft sent to him did not include this passage), he did indicate that he thought there would definitely be emails between him and “people at JPMC” ; but none was ever produced.

Second, what had been described in the “ Strategic Deals Memorandum” as “a package approach to this demand for strategic selling” and “combined hardware/software application offerings in the compliance space” was elaborated into “large sales in Q3 (comprising software and hardware utilising the software in an appliance model)” and sales of “software and hardware as part of their compliance solution”. The Claimants submitted that this reference to an “appliance model” was false. No such “appliance model” had been developed, and the sales were of ‘pure hardware’.

Third, what had been referred to in the “ Strategic Deals Memorandum” as “the strategic partnership with EMC… led by Bill Scannell” was elaborated into:

“The partnership with EMC is allowing the development of joint products and marketing of EMC/Autonomy solutions to the customer. This has served to give EMC confidence despite no longer owning the relationship. These arrangements have been brokered at the highest level by the CTOs and we view these sales as part of a bigger strategic picture.”

As to this:

(1) The Claimants dismissed the passage as also untrue in every material particular, contending that (i) there was no partnership with EMC; (ii) there was no development of joint products with EMC; (iii) there was no joint marketing initiative between EMC and Autonomy; (iv) EMC still “owned” its relationship with the customers to whom the hardware was being sold; and (v) any arrangements between Autonomy and EMC had been made between Mr Sullivan and Mr Scannell, and no part had been played by Dr Menell.

(2) When cross-examined on it, Dr Lynch disagreed that there was no partnership with EMC and stuck to his line that the process of developing a joint product with EMC had “already happened at one level by this time and we then go on to work on other things with them” . He also defended the description of “strategic package sales” as extending to sales of hardware to established customers for Autonomy software with a view to supporting software already acquired, or further software to be supplied whether as part of the hardware sale or later. But all this was a semantic exercise, with Dr Lynch‘s justification amounting in reality to insistence on his own dictionary. The plain import of what was said was against him.

Fourth, the explanation in the “ Strategic Deals Memorandum” of the increase in “Operating costs” including costs of sales was expanded in Mr Hussain’s Q3 2009 Note as follows:

“ There were 2 large movements. Firstly, in sales and marketing we spent around $20m on sharing marketing costs with EMC and extra marketing on our new product launch (Structured Probabilistic Engine). EMC are using the monies in highly targeted joint marketing programmes with companies such as JPMC and Citi and in also jointly developing further appliances for future sales. Secondly we capitalised $11m of R&D under IAS 38 as a direct result of significant development effort on the new product release (SPE)…”

The Claimants alleged that both aspects of these explanations were false. More particularly:

(1) The aspect relating to R&D and SPE has been dealt with separately.

(2) The aspect relating to statements as to there being a “highly targeted joint marketing programme” with EMC and as to joint development of further appliances for future sales was exaggerated: as previously explained, there was no such “targeted joint marketing programme” ; EMC had not committed, and was not required, to use the money it received for the hardware in any particular way; and any development plans for joint Autonomy/EMC appliances were at best inchoate.

In my judgment, the intended effect of Mr Hussain’s Q3 2009 Quarterly Note was to confirm and amplify for the particular attention of the Audit Committee the message that the hardware sales did not indicate that Autonomy was engaged in a new and separate (and loss-making) business of hardware reselling: that they were part and parcel of the existing software business and could properly be accounted for as such.

Deloitte ’s Q3 2009 Review

On 16 October 2009, Deloitte issued their Q3 2009 report to the Audit Committee. Prior to its finalisation, Deloitte had sent a copy in draft to Mr Hussain and Mr Chamberlain. Mr Hussain forwarded it on to Dr Lynch stating: “ See the positioning re emc - any comments? ”

The wording of that report is a further refined version of the “ Strategic Deals Memorandum” which may have borrowed also from Mr Hussain’s Q3 2009 Quarterly Note. It displays the same misleading depiction of the extent of the relationship between Autonomy and EMC, with the same implicit suggestion of commitment on the part of EMC to invest the ‘delta’ in agreed ways (though Deloitte was more circumspect in avoiding any reference to a ‘partnership’ than Mr Hussain had been). Its references to the prospective development of an appliance described the cost/loss as “Autonomy’s upfront investment in working jointly with EMC to develop this proposition.”

T he opening paragraph stated as follows:

“During the quarter, the executive management identified a new and significant longer term market opportunity for Autonomy to develop in the provision of appliance related solutions to leading multi-national financial institutions. In order to begin to establish the Group’s presence in this space Autonomy management identified the need to develop a close working relationship with a major hardware company. EMC were approached by the executive management team and a significant hardware, marketing and development purchase entered into with this organisation. The purpose of this transaction with EMC was:

- to provide hardware to Autonomy for it to deliver to its existing customers,

- to provide ongoing joint sales and marketing support to promote further sales to this emerging market and

- to begin to develop an appliance based hardware and software configuration whereby Autonomy software might be fully integrated into EMC hardware for products aimed at the appliance sector and major financial institutions.”

This was at best exaggerated: the arrangements made with EMC were not concerned with developing “ appliance related solutions ”; there was no “ hardware, marketing and development purchase ” entered into with EMC; EMC did not “ provide hardware to Autonomy for it to deliver to its existing customers ”; there was no “ joint sales and marketing ” arrangement with EMC; and there was no agreement, or even discussions, between EMC and Autonomy regarding the development of “ an appliance based hardware and software configuration ”.

Deloitte’s draft report went on to state:

“The procurement of goods, marketing services and future development costs have been allocated between cost of goods (within gross margin) and sales and marketing costs (which fall to be treated as operating costs). … The marketing cost is being used to incentivise the EMC salesforce; provide discounts to the customer; provide funds for the development of the appliances; and to attend marketing events.

Management’s rationale behind entering into these loss making contracts is that Autonomy is seeking to develop a strategic relationship with EMC whereby in future an appliance will be marketed which combines Autonomy’s software with EMC’s hardware. The $9 million cost over and above the $36 million recovered through the sales reflects Autonomy’s upfront investment in working jointly with EMC to develop this proposition. Management has considered whether these costs should be capitalised but has concluded that they do not meet the necessary asset criteria and accordingly has expensed them as incurred.”

Again, this was based on what Deloitte had been told by Dr Lynch and Mr Hussain. Although Deloitte steered away more carefully than had Mr Hussain from asserting a partnership or any commitment on the part of EMC, I accept the Claimants’ contentions that (a) the scope and extent of Autonomy’s then existing relationship with EMC was exaggerated, and (b) there was next to no substance in the assertion that part of the amounts paid by Autonomy went towards attendance at marketing events. Except for the emails attached to the Strategic Deals Memorandum, these calculations were arrived at unilaterally by Autonomy, without any input from or discussion with EMC. The only prices of which Autonomy in fact had knowledge were (i) the price they had agreed to pay EMC and (ii) the price at which it was on-selling to the customer.

At that time, none of the members of the Audit Committee had any accountancy qualification; and it is unsurprising that they accepted Deloitte’s approval of the Q3 2009 Report (subject to the caveats referred to above) without (so far as the evidence goes) any question, except as to disclosure.

How Dr Lynch dealt with the Audit Committee ’s questions about disclosure (Q3 2009)

Before resuming my description of how Mr Hussain’s Quarterly Notes and Deloitte’s Quarterly Reviews shaped the attitude of the Audit Committee over the course of the Relevant Period, it is convenient to interpose an episode which arose immediately after the Q3 2009 Audit Committee meeting relating to the issue of disclosure.

An important detail in Deloitte’s Q3 2009 review was a passage as follows:

“These hardware sales did not include any IDOL software component and reflect Autonomy’s early targeting of the emerging market of appliance solutions. The Board should consider how best to communicate this new opportunity to the shareholders as these revenues are not driven from the IDOL technology of the Group.”

[Emphasis supplied by me]

On 21 October 2009, and following up earlier emails from him and his colleagues about the lack of any mention in draft minutes (prepared by Mr Kanter) of discussions they all recollected having had at the Audit Committee meeting about (in Mr Ariko’s words) “ how to best represent and incorporate the EMC servers we are reselling into our financial results”, Mr Ariko sent an email to Dr Lynch regarding “the revenue we’re putting on our books from the reselling of these storage servers”. He stated that he agreed that “our Strategic position with many of our large customers is a good thing” and noted also that “because of that reselling another companies products does present us with incremental revenue and margin contributions.”

Mr Ariko’s only real concern was with disclosure. Perhaps prompted by a passage in Deloitte’s Q3 2009 Review, he wanted to ensure that the effect of this new revenue stream, and the performance of Autonomy’s software business (which he described as “our most strategic products” ), were fairly presented both for internal purposes and also from the point of view of the market . He was explicit:

“…I think we have a potential problem with how we discuss the revenue we’re putting on our books from the reselling of these storage services. We need to agree how best to present those numbers to the Street and how best to review the performance of the Company without those revenue/profit numbers included to best understand the state of our business regarding our most strategic products. I believe this needs to be an extensive discussion at the next Board meeting and I think we should review our [current] press release so that it adequately reflects the effects of this new business line.”

This concern was shared by all his colleagues on the Audit Committee, and expressed clearly after it transpired that the Press Release was issued notwithstanding these concerns. All three members wrote expressing concern about the lack of discussion about and disclosure of the hardware reselling strategy in the Q3 Press announcement:

(1) On 19 October 2009 at 4.18pm, in response to having been sent for approval the draft minutes of the Audit Committee meeting held on 16 October 2009, which did not record any discussion about hardware sales and revenues, Mr McMonigall sent an email to his fellow Committee members, and also to Mr Kanter, copying Mr Hussain and others at Autonomy, which stated:

“Fine with me. But how did we deal with the hardware GM issue in the announcement?” .

(2) Mr Ariko sent an email less than an hour later stating:

“I think the minutes need to reflect that we had a discussion about how to best represent and incorporate the EMC servers we are reselling into our financial results”.

(3) The third member, Mr Perle, later that evening sent an email agreeing with Mr Ariko and stating that he:

“share[d] John’s interest in knowing what the press release says”.

As has already been seen, the Press Release had been issued earlier that day. It made no mention of the “new business line”. Although one of the functions of the Audit Committee was to approve the quarterly results and press release, it appears that they, and these concerns, were by-passed. The responsibility for this is not clear. There is no record of Mr Kanter having sent to Deloitte his draft minutes, which the members of the Audit Committee all thought missed out this vital discussion. I did not hear from any of those members; or from Mr Kanter; or from Mr Knights, who did attend the meeting. I have had no opportunity to assess properly what may be competing recollections.

But, to my mind, who was responsible matters less than the undoubted confirmation that the episode provides of (a) the natural reaction at the time of informed but objective persons, asked to consider the rationale for the hardware sales, that in any event they should be disclosed; (b) the determination on the part of management that they should not be so; (c) the inevitable suspicion surrounding (i) the omission from the minutes of the discussion all three members of the Audit Committee recalled and (ii) the apparent failure (including by Mr Knights, if the discussion took place) to factor in the Audit Committee’s views on a matter which was part of their remit. Both (a) and (b) above were further demonstrated by two further emails.

Mr Hussain’s reaction to Mr Ariko’s email was to send an email to Mr Kanter (only) which provides an insight into his outlook on the issue of disclosure:

“We made a statement about the strategic sales in MRL’s quote - we cannot give too much detail in the press release as it’s commercially sensitive. In the press release we said “during the quarter we saw some of our large customers promote autonomy to strategic supplier status. This has led them to adopt a broader set of solutions in a number of significant sales.””

Dr Lynch responded directly to Mr Ariko on 23 October 2009. The Claimants fastened on his response as showing clearly that he was “perfectly content to mislead Mr Ariko”. His response was prefaced by an assurance of complete agreement that the matter needed to be monitored closely, and a reference to Mr Knights having “explained his treatment of this to me and Andy and Sushovan advised that they felt the positioning was correct but also added that like you something we needed to keep under review”. The remainder and the Claimants’ basis for criticising particular parts of it are summarised below:

(1) Dr Lynch stated that:

“ The market is already aware we sell hardware, something we have done for 5 years or so and indeed we mentioned it as being relevant to q3 on the conference call and in Q3 shareholder meetings and indeed that this had been more pronounced this quarter. This has been mentioned by financial analysts in their coverage of the quarter.”

(2) The Claimants contended that the market had not been informed that Autonomy was selling pure hardware, and that the transcript for the Q3 2009 earnings call does not contain a single reference to the hardware sales. They paraphrased Dr Lynch’s references to the situation in Q3 being “quite complex involving hardware, our software to customers and a partnership development” and to “ the move … towards an appliance model rather than usual hardware re-sell ” and a “ new strategic product offering ” as intended to maintain a fiction of a partnership with EMC for the development of an appliance which did not accurately reflect the reality of the position with EMC.

(3) Dr Lynch stated that Autonomy “would not want to become resellers of unrelated hardware which is not about furthering our software sales”.

The Claimants contended that was the reverse of the truth because according to the Claimants’ case that was precisely the strategy that had been implemented and the hardware sales during Q3 2009 were a means of producing recognisable revenue, rather than having anything to do with “ furthering … software sales ”.

(4) Dr Lynch assured Mr Ariko that Autonomy was “issuing a new press release next week on these matters relating to hardware”.

(5) According to the Claimants, this was incorrect. As Dr Lynch well knew, the press release in question had nothing to do with pure hardware sales: it concerned the launch of the Arcpliance appliance: and they cited a press release issued on 28 October 2009 (“ Autonomy Announces New Archiving Appliance ”).

(6) Dr Lynch concluded:

“I think its a good subject to discuss at the board meeting although perhaps we need to monitor the next couple of quarters to see if this is a one off or a trend before reacting”.

(7) The Claimants said that this was misleading because by that time, it was already clear that these hardware sales were not intended to be a one-off (as Dr Lynch acknowledged in his own witness statement by the words “ EMC was just one of the hardware providers we planned on working with as the hardware strategy developed ”). Rather, the Claimants continued, the sales had become an important means of generating revenue for Autonomy and indeed, as explained below, a matter of days later, Dr Lynch and Mr Hussain were setting an ambitious hardware revenue target for Q4 2009.

I accept that Dr Lynch’s email was misleading. Tracking the sub-paragraphs above:

(1) As to (1), Dr Lynch’s assertion that the market was already aware that Autonomy sold hardware was true only in the most literal sense; it skated over the fact that the very considerable increase in the volume of sales had been kept from the market, and the sales were only ever mentioned in the context of sales of appliances or Arcpliance. Further, and more concretely, it is correct that there was no mention of hardware sales in the Q3 earnings call; and though Dr Lynch may have confused the call with his own public statement on the Q3 results, which did contain the Delphic reference to Autonomy having, in consequence of certain large customers having promoted it to “strategic supplier status”, adopted “a broader set of our solutions in a number of significant deals”, the inaccuracy and evasiveness are very striking.

(2) As to (2), Dr Lynch’s references to partnership coupled with references to a “a new strategic product offering” had no real grounding in the truth.

(3) As to (3), I have determined that whatever its purpose when first formulated, by the time of this exchange the hardware reselling strategy’s purpose was the generation of revenue to make good shortfalls in software sales without disclosure of their extent or true source.

(4) As to (4) and (5), the Claimants were correct in their contention that Dr Lynch must have been referring to a press release dated 27 October 2009 which announced Autonomy’s Arcpliance product, which had nothing to do with the hardware sales to which Mr Ariko was referring.

(5) As to (6) and Dr Lynch’s concluding statement to the effect that board discussion of hardware sales might be better postponed because “perhaps we need to monitor the next couple of quarters to see if this is a one off or a trend before reacting” smacks of evasion and sits uneasily with Dr Lynch’s presentation of the hardware reselling strategy as a continuing strategy, involving not only EMC but also other hardware sellers.

Again, this is plainly at odds with the picture of open and constructive engagement painted by the Defendants as described above. The impression that Dr Lynch interpreted Deloitte’s advice that the Accounting Standards did not require separate disclosure of revenue from the hardware reselling strategy as justifying him taking active steps to disguise or distract attention from it is reinforced by what he said at the Earnings Call for Q3 2009 (which took place on 20 October 2009).

Mr Hussain ’s Q4 2009 Quarterly Note

A draft of Mr Hussain’s Q4 2009 Quarterly Note was, as usual, pre-circulated to Dr Lynch for his comments. The final version was dated 29 January 2010 and circulated to the Audit Committee in advance of its meeting on 1 February 2010. The Note did not mention EMC or appliance sales, but contained a passage on “Strategic Sales” , as follows:

“Strategic sales - only one new strategic sale for $1m this quarter (Bank of America) and 2 were delivered from deals concluded last quarter (Morgan Stanley for $6m and Credit Suisse for $4m). The total is $11m or 4.9% of total sales. The Morgan Stanley sale is particularly strategic as we made a further $12m of software sales to Morgan Stanley in the quarter. The Bank of America sale is also strategic in that we are in the midst of a large 7-figure sale of software in Q1’10.”

The Claimants submitted that this was misleading because:

(1) As Dr Lynch accepted, the “ $12m of software sales to Morgan Stanley ” was in fact the restructuring of the hosting arrangement with Morgan Stanley. That transaction was entirely unrelated to the hardware sales that had been made to Morgan Stanley. There was no linkage between the two. Thus, it was misleading to state that the hardware sale to Morgan Stanley was “ strategic ” to the software sale.

(2) The same was true in relation to the BofA hardware sale, which was made to the reseller SHI. Again, the hardware deal was not “ strategic ” to the software deal referenced, which was a hosting arrangement that was negotiated entirely independently.

In cross-examination, Dr Lynch sought to justify their description on the basis that both were nevertheless “part of our strategic solutions problem” . He explained this as follows:

Q. How do you say there was a relationship between a restructuring of the hosting deals and a sale of hardware to Morgan Stanley which was going to go on site?

A. Because they’re being done under the Morgan Stanley arrangement from 30 June 2009 which is an arrangement that is to do with hardware and also mentions Digital Safe, and then the other aspect of this is that Morgan Stanley owns some of the hardware that we host, and then the last aspect of this is that Morgan Stanley also has some safes which we’re not hosting.

…

Q. I suggest to you that the way in which this has been presented to the audit committee, suggesting some kind of close link, or link between these transactions, was just misleading, wasn’t it?

A. No, that’s incorrect. And Deloitte were well aware of the level of linkage in each deal and there are working papers that discuss their understanding of that in detail…”

In relation to the BofA deal, he explained the linkage as being:

“..that they are part of our strategic program…which by selling these large banks the hardware, we are able to fulfil the strategic need. Now, there are other sales if you’re correct about these ones, where actually the software is running on the hardware and there are other ones which are the appliance-type sales. So there’s a mixture going on here but the basic principle is that we are doing one-stop shopping for the banks.”

The effect of this was to cover all situations. In his oral evidence before me Dr Lynch extended his asserted meaning of “strategic sales” to apply to any sale of hardware to a software customer of Autonomy, existing or prospective. As also appears from the passage quoted in paragraph 1406 above he also told me that Deloitte were “well aware” of the “level of linkage”. However, what Deloitte had been told appears from a passage in Deloitte’s own Report to the Audit Committee on the Q4 2009 Audit:

“These hardware sales did not include any IDOL software component and reflect Autonomy’s continued targeting of the emerging market of appliance solutions. However, we do note that during Q4 2009, Autonomy sold and have also sold significant software to Bank of America (who are the end-user in the SHI deal) in recent quarters. As consistent with the hardware sales reported to the Audit Committee in our Q3 report, these sales have been made at an overall loss with the costs being allocated between costs of sales and marketing expenses. Management’s rationale for entering into these loss-making contracts is that Autonomy is seeking to develop a strategic relationship with EMC whereby in future an appliance will be marketed which combines Autonomy’s software with EMC’s hardware. It should be noted that the sale to SHI international was not connected to this strategy with EMC and involved a sale of 1,000 laptops to Bank of America via SHI International which Autonomy had purchased from Dell. The intention here is that both Autonomy and Dell will market Dell hardware that incorporates Autonomy search software.”

There are a number of problems with Dr Lynch’s evidence in this regard:

(1) In respect of each transaction , Deloitte appear to have proceeded on a much more specific basis than Dr Lynch’s very broad extension of the concept of a strategic sale in cross-examination.

(2) But in neither could the basis on which they proceeded be supported.

(3) By the beginning of Q4 2009 EMC had withdrawn from the hardware reselling strategy. There was no “strategic relationship” or agreement for marketing of a joint appliance. Mr Welham’s unchallenged evidence was that Deloitte had not been told this during the course of the 2009 year-end audit.^{^[212]} Nor had the Audit Committee.

(4) Any arrangements for the development of an appliance with Dell were inchoate and uncertain.

I am persuaded and find that Deloitte were misled and the Audit Committee likewise.

Mr Hussain’s Q4 2009 Note was also misleading in stating as follows under “ Operating costs ”: “ … the costs associated with the strategic sales (joint marketing and cost of sales) were down … ”. However, no “ joint marketing ” had been agreed, let alone undertaken, with any of the hardware manufacturers.

Mr Hussain’s Q1 2010 Quarterly Note

By Q1 2010, Deloitte knew that Dell had taken over from EMC as the primary hardware supplier. Some recasting of the justification for hardware sales was required. The Claimants’ allegations in respect of Mr Hussain’s Q1 2010 Quarterly Note (the Claimants called it a ‘paper’) related particularly to the alleged depiction of a sale of hardware to BofA as part of a “package sale” resulting in a software purchase by BofA from Autonomy, both in Q1 2010.

These allegations were particularly directed against Dr Lynch, as having been responsible for changing a draft of Mr Hussain’s Q1 2010 Note in the manner shown by the text and deletions below:

“Strategic Package ic sales - the total was $7m (3.5% of total sales) of lower margin business including a strategic sale to the Bank of America for $145m$9m of ~~. This should be seen in the light of a separate $9m major strategic sale of~~ compliance software and 5m of low margin business ~~also in the quarter~~ (we flagged this sale in the Q4 Audit Pack). We also sold $41m to Fannie Mae ~~and at the same time sold~~ $3m of compliance software and $1m low margin ~~to Freddie Mac~~.”

Thus, whereas the original draft had described the software and hardware sales as separate (indeed, that very word was used in the pre-circulation version in relation to the BofA software sale), the effect of what I accept were changes made by Dr Lynch ^{^[213]} was to suggest that they constituted part of the same “ package sale ”. These changes, which recast the justification for the hardware reselling strategy in conjunction with Dell as a means of enabling Autonomy to deliver its software through “ package ” sales of both software and hardware, were all adopted by Mr Hussain in the final version.

As amended, the wording conveyed an inaccurate impression of composite sales; but, read on its own, the justification was no longer based on the quest for a joint appliance solution. Yet it is clear that Deloitte understood the strategic rationale still to be to develop a strategic relationship with Dell, as can be seen from their own Report to the Audit Committee on the Q1 2010 Review dated 20 Apr il 2010 in which, in the section on “Hardware” , it was stated that:

“Management’s rationale for entering into these loss making contracts is that Autonomy is seeking to develop a strategic relationship with Dell. The intention is that both Autonomy and Dell will market Dell hardware that incorporates Autonomy search software”.

It is not clear to me from the evidence what the source or basis was for this gloss on what was stated in Mr Hussain’s Q1 2010 Note. The gloss was wrong: there was no strategic relationship between Autonomy and Dell, and any intention of marketing a joint appliance was vague and inchoate. Indeed, Mr Chamberlain expressly acknowledged this in an email to Mr Welham and Mr Knights (cc Ms Harris and Mr Hussain) dated 19 April 2010 seeking to justify the allocation of 62.4% of the costs of sales of Hitachi hardware to Sales and Marketing on the basis that “HDS are developing appliances and providing similar levels of support to EMC” , but an allocation of the reduced figure of 37.6% in the case of sales of Dell hardware because “there is no development effort for Dell…”

In such circumstances, it does not seem to me that Deloitte was misled in this regard by Mr Hussain’s Note; nor by Mr Chamberlain. That does not alter the fact, however, that Deloitte’s gloss on the justification was put forward to the Audit Committee without correction by Mr Hussain or Mr Chamberlain (or anyone else at Autonomy). It seems to me that Autonomy was content not to correct a confusion which was to its advantage in offering the best prospect of securing the result it wanted, and more especially, the allocation of part of its costs to Sales and Marketing.

That said, and as Deloitte’s Report explained, even on the basis of part of the costs being referable to b uildin g a “strategic relationship” Deloitte was not prepared to agree to the allocation percentages presented by Autonomy ^{^[214]} . This was reflected in their report:

“ Given the period that has elapsed since these initial deals were transacted and the fact that we expected these to be more one-off in nature, we conclude that it would be more appropriate to reflect all of the costs of hardware sold in cost of goods sold.

We understand that management has allocated the $3.8 million to sales and marketing based on the previous analysis prepared for the EMC sales in Q3 2009 which demonstrated that Autonomy were purchasing hardware at a price which was considerably higher than they would normally pay in order to gain a strategic partnership and become the preferred hardware reseller with EMC, Dell, SHI and HDS.

Based on the limited information available, we have included the $3.8 million as a classification adjustment in Appendix 1 and would not expect to see such amounts in sales and marketing in subsequent quarters.”

In summary, Mr Hussain’s Q1 2010 Note:

(1) paved the way for a shift in the justification of the hardware reselling strategy away from the development of an appliance to the customer-facing justification of a “one-stop shop” ;

(2) reinforced the presentation of the hardware reselling strategy as nevertheless intended to promote Autonomy’s software core business;

(3) was interpreted by Deloitte, when taken together with other representations, to justify the accounting treatment of hardware revenues as part of the overall revenues of the group without differentiation.

Mr Hussain’s Q2 2010 Note

The shift in the justification for hardware reselling which had been commenced in Q1 2010 was confirmed in Q2 2010. Dr Lynch accepted in cross-examination that there was a shift:

“Q. It’s right, isn’t it, from Q2 2010 onwards, rather than trying to position the hardware sales as being part of some strategic relationship with a hardware manufacturer, including the production of appliances, what one finds going forward is a somewhat different justification being advanced?

A. Yes, it changes. What we’re doing changes. It starts to move away from the appliance-type situation more to the strategic package sales.

Q. ….it’s not about appliance sales any more, it’s about how this positions Autonomy well in order to make valuable software sales not loaded on to the hardware but just separately to some of the customers, correct?

A. Yes, I think it ’s shifted.”

That was consistent also with an earlier passage in his cross-examination:

“I would say at the beginning of the strategy it was more heavily weighted towards the appliance than the hardware providers, but a few months later, probably around Q2 2010, we were seeing more of Cloud take-off rather than appliance and that became less important and at that point we had more interest in the customer-facing aspect of it…”

In their submissions as to the alleged deception of Deloitte and the Audit Committee in Q2 2010 the Claimants focused on three points in Mr Hussain’s Q2 2010 Note dated 20 July 2010 (a draft of which was, as usual, pre-circulated to Dr Lynch):

(1) The first was the statement under the heading “Strategic package sales” that:

“…these types of sales were flagged in the Q1 results presentation so the market is aware of them ”.

(2) The second was that:

“ These lower margin sales have generated significant new software business for us (over $80m of sales have been associated with these sales over the past few quarters). ”

(3) The third was the statement under a heading “Gross margins” that:

“We have charged the cost of the lower margin sales to the cost of sales line even though we had agreed with our suppliers that the 50% of the cost could be used for marketing purposes.”

O n the first point, the Claimants submitted that the statement that “strategic package sales” had been “flagged” in the Q1 results presentation was not true. There had been a reference in the Q1 2010 Earnings Call to having purchased hardware inventory to take “advantage of discounted stock offers to purchase stock in advance of Q2 sales” and the PowerPoint presentation used for that call referred to “Pre-investment in inventory for Q2’10 sales”. But in ensuing discussion when asked what “Arcpliance or appliance sales and hardware sales would have been in a normal quarter”, Dr Lynch had replied “…It would be a fraction of that number” .

The Claimants placed the blame for the untrue statement on Dr Lynch because it did not appear in the draft that was sent to Dr Lynch and was inserted only after that. When asked about this, Dr Lynch told me that he did not know whether the addition had been made at his request; but he submitted that there was no basis for the Claimants’ suggestion that the text was inserted at his suggestion, or even that he saw it. But no other explanation was suggested as to how the amendment to the pre-circulated draft came to be made.

Dr Lynch was obviously discomfited when confronted with this in cross-examination, both as regards the inclusion of the statement and as to what it stated. As to whether the false statement was his suggestion, I have been struck by how often Dr Lynch used the words “flag” or “flagged”: 11 times in his first witness statement, for example. Taking account also the fact that the amendment was introduced sequentially following the draft having been sent to him for his consideration, and the lack of any other explanation for it, on balance, I think it more likely than not, and I find, that he did cause the insertion of those words, even if he cannot now remember that he did.

He admitted that the statement was “not as clear as I would like it to be”. When pressed on whether it was misleading, he first suggested that the intended reference was probably to the Q3 2009 Earnings Call, but then added that there had been more information given in Q1 2010 as well in respect of “the Arcpliance solutions” which he described as a “subset” of “strategic package sales”. But this was glib, and I formed the impression that Dr Lynch was grasping at straws. None of his efforts to explain really addressed the point: they did not justify the statement that volume sales of “strategic” hardware had been undertaken.

Mr Bloomer, asked about the same statement appearing in the Q3 2010 Audit Committee presentation, indicated that the text would not necessarily have concerned him, because “ at the time, hardware sales were not a particularly big issue in Deloitte's report ”. He added when asked whether, if he had become aware that the strategic package sales which he was told had been flagged had not been so, that would have concerned him, that:

“It would have been interesting rather than concerning…hardware was not seen as a big topic actually to flag. ”

This was, as I see it, because Mr Bloomer, and the other members of the Audit Committee took the view that unless the hardware sales were of such an amount as to (a) undermine, or potentially undermine, the judgement that Autonomy had a single operating segment (its core IDOL-based software business) and (b) be material, hardware reselling was not a problem unless of course Deloitte said it was: and they never did.

The statement in the Note (see (2) in paragraph 1422 above) that “ over $80m of sales have been associated with these sales over the past few quarters ”, was based on Mr Stephan’s analysis (see below), which became known as “the Linkage Analysis”. This showed $78m of software sales to Autonomy’s hardware customers in the period from Q2 2009 onwards. As I develop in the context of my analysis of the dispute relating to the “linkage analysis”, ‘linkage’ between events or transactions is an elastic expression which could denote anything from temporal coincidence to causation; but the Defendants submitted that the wording “ associated with ” did not suggest a direct, close, causal link. While the version sent to Dr Lynch contained the text, the number it contained was different ($40m rather than $80m). In any event, as Dr Lynch said, the document would have gone to Deloitte, who would have checked it and flagged to the Audit Committee if it contained anything misleading; and they never did so.

Behind the statement in Mr Hussain’s Q2 2010 Note (see (3) in paragraph 1422 above) that Autonomy had “charged the cost of the lower margin sales to the cost of sales line even though we had agreed with our suppliers that the 50% of the cost could be used for marketing purposes” lay a debate between Autonomy’s finance department (and, in particular, Mr Chamberlain) and Deloitte (with both Mr Knights and Mr Mercer involved) on the abiding issue of costs allocation.

In the previous quarter (Q1 2010), Deloitte had pushed back and after much wrangling obtained an agreement to limit the allocation to Sales and Marketing to an immaterial $3.8 million, on the basis of “a classification adjustment in Appendix 1” (see paragraph 1418 above). It is clear, as also intimated previously, that Deloitte were very uncomfortable with any allocation at all; and the identification of the relatively small allocation permitted as a “disclosure deficiency” (which appears to have been Mr Welham’s idea) was intended (in Mr Welham’s words in his witness statement) “to flag to management that we would not expect such classification going forward.” Deloitte gave the same explanation and admonition to the Audit Committee.

Even so, Mr Chamberlain pressed in Q2 2010 for a 50% split between COGS and Sales and Marketing. In a Memo to ‘The Files’ cc Mr Hussain dated 16 July 2010 and headed “ Product - cost allocation” , Mr Chamberlain sought to justify this as follows:

“During Q2 2010 Autonomy has continued its practice of procuring hardware at a perceived loss on behalf of its most strategic customers. As has been previously explained the purpose of these deals is to become the single source supplier for all of the major banks in relation to its data management activities…

…

…the total cost represents a “loss” on the hardware sales. However, these sales should not be considered in isolation. A proportion of the payment represents an investment in the customer relationship and has helped enormously in the procurement of significant software sales. In the last 4 quarters alone we have signed licence deals that have generated almost $80 million in revenues with a further $40-60 million in backlog of estimated hosting fees that will be generated for the initial term of the agreements.

The cost allocation of 50% to COGS and 50% to sales and marketing is consistent with the quotations provided to the hardware vendors. The sales and marketing payment is effectively a commission payment to the hardware vendors for allowing us to secure the sole supplier relationship with the banks and has been a very successful enabler in closing the larger, much more valuable, software deals.

Every purchase quotation provided to the vendors contains the following clause:

“The purchase order amount above includes payment for the equipment/services listed in the attached quote, in addition to payments for the joint marketing support referenced in the Agreement. For Autonomy purposes, the value of these transactions has been apportioned as follows: equipment/services 50% and marketing support 50%”.

The allocation reflects two things. 1) normal cost allocation to COGS such that hardware margin is consistent with margins earned by large hardware vendors; 2) balance of payment being allocated to S&M.”

I would make the following observations on this:

(1) In the teeth of Deloitte’s obvious discomfort and express warning that they did not consider allocation of the costs of hardware sales to Sales and marketing rather than COGS to be justified as a proper expense, and were only prepared to permit the allocation of a much smaller amount ($3.8 million) than Autonomy had pressed for as an exceptional “disclosure deficiency” which they would not expect to sanction in future quarters, Mr Chamberlain’s tenacity is notable and indicative of quite how important was the allocation of costs issue to the finance department for the very reason which Deloitte’s PSR had identified in Q1 2010 - preserving gross margin.

(2) Mr Chamberlain presented the “commission payments” as if the hardware suppliers had agreed that this is what they represented: whereas the purchase orders themselves simply indicated how Autonomy was intending internally to account for them ( “For Autonomy’s purposes” ).

(3) Mr Chamberlain’s assertion that Autonomy’s funding of discounts on hardware had “helped enormously in the procurement of significant software sales” , with a further suggestion of the generation thereby of “almost $80 million in revenues with a further $40-60 million in backlog of estimated hosting fees”, had no support beyond the Linkage Analysis (which from now on became the principal basis of justifying the desired allocation of Hardware sales costs), but which the Defendants have had to accept amounted to little more than what its author Mr Stephan referred to as a “matching up names” exercise to show a correlation between buyers of hardware and buyers of software and sometimes a broad temporal correlation, which, since all purchasers were existing customers for Autonomy software, demonstrated little.

(4) The basis on which Mr Chamberlain felt able to determine and represent that the “normal cost allocation to COGS such that hardware margin is consistent with margins earned by large hardware vendors” is unclear, just as it had been in Q3 2009.

(5) Although Mr Hussain’s Q2 2010 Note stated that Autonomy had agreed to forego the benefit in terms of cost allocation of what was presented as an agreement “with our suppliers that the 50% of the cost could be used for marketing purposes” (see paragraphs 1422 and 1430 above), adding that Deloitte concurred “with this prudent approach” , behind the scenes Mr Chamberlain and the Defendants continued their efforts to persuade Deloitte to agree to some allocation of costs to Sales and marketing. The importance of this to them all was again evident from an email dated 18 July 2010 from Mr Hussain to Mr Mercer, Mr Knights and Mr Welham, in which he made clear that “The main issue to get across the line on is COGS. Steve will discuss with Lee” and noted “Also MRL has strong views on this in terms of the business rationale”.

In the event, Deloitte were not persuaded to accept the proportions proposed by Mr Chamberlain. However, rather extraordinarily in general terms, but by now characteristically in the context of the relationship that had developed between Autonomy and Deloitte, they were persuaded to permit the allocation of losses to sales and marketing on the basis of “a solution” devised by Mr Knights and recorded in an email from him to Mr Hussain, Mr Mercer, and Mr Welham (cc Mr Chamberlain) late in the evening on 18 July 2010.

The email is quoted later as a further illustration of the extent to which Mr Knights had adopted the role as supportive and constructive adviser rather than questioning and sceptical auditor, and was quite prepared to “workaround” his own compliance people. It is sufficient to note for the present that the upshot was that once again Deloitte approved the allocation of some $4 million to Sales and Marketing, which enabled Autonomy to present its gross margin figure as 88% for the six months ended 30 June 2010 compared to 89% for the first six months of 2009, and the small variation in the comparative Q2 2009 and Q2 2010 quarterly gross margin (89% as compared with 86%) as (quoting from the Q2 2010 Quarterly Report released on 22 July 2010):

“…in line with our expectations due to the sales mix including appliances as discussed last quarter.”

In summary, Mr Hussain’s Q2 2010 Note, Deloitte’s own Report to the Audit Committee on the Q2 2010 Review, and the work done behind the scenes for both:

(1) completed and reinforced the ‘shift’ in the justification for the hardware reselling strategy;

(2) reassured the Audit Committee that the hardware reselling strategy did not destabilise the previous analysis that Autonomy was a single segment business with revenue from sources which required no differentiation; but to do that;

(3) contained a demonstrably false assertion that “strategic” hardware sales had previously been “flagged” to the market which I have concluded was inserted at the suggestion of Dr Lynch;

(4) confirmed yet again the Defendants’ determination to say as little as possible about the hardware reselling strategy, and nothing as to its nature and extent (apart from the falsity mentioned above);

(5) demonstrated Mr Knights’ role as presenter of ‘solutions’ even in the teeth of concern from his team and the Reviewers, and the warning from one of the latter that it appeared likely that Autonomy’s objective in posting as much as Deloitte would endorse of the costs to ‘Sales and Marketing’ instead of COGS was to preserve gross margin.

Whether Deloitte were actually deceived by the Q2 2010 Note, however, is a more problematic question. As previously noted, Deloitte had full and detailed information about the hardware sales, except perhaps what was revealed in email exchanges between Autonomy and hardware suppliers. It seems to me to be difficult to conclude that Mr Knights was misled by the representations adumbrated in paragraph 1422 above; and, for example, Mr Welham, who attended all the Earnings Calls, must surely have known that the hardware sales had not been “flagged” to the market.

Obviously, it would have been helpful, in this context especially, to have had evidence which could be tested from one or more of Mr Knights, Mr Mercer, Mr Knight and the Reviewers. On what has been put before me, the uncomfortable truth appears to me to be that by this time, at least Mr Knights and thereafter, Mr Mercer, no longer had the sceptical approach required of an auditor: neither in his turn retained the detachment and independence of mind that was required to call out Autonomy. Instead, they were content to take at face value representations which sceptical reflection would have revealed to be inconsistent with what they or the market had been told in the recent past. Autonomy was an important client; and I should imagine (though I had no direct evidence about this) of very great importance to the Cambridge office which Mr Knights had managed.

That raises further questions as to (a) whether this undermines the Claimants’ submission that had Deloitte and/or the Audit Committee been told the truth, they would have objected to the programme and/or its accounting treatment and required its disclosure to the market (see paragraph 1368 above) and (b) whether this establishes, or at least strongly supports, the Defendants’ case that the propriety of what they were doing had been confirmed by Deloitte and the claim of dishonesty should fail accordingly.

As to (a) above, in my view, even if Deloitte continued to rely on what they knew to be unreliable or false, it seems to me that it was the story spun by Autonomy which continued to be the basis on which Deloitte themselves reported to the Audit Committee and on which both Deloitte and the Audit Committee approved the accounts in Q2 2010, as before. Even if Deloitte had no proper basis for relying on what they were told in the respects identified above, these proceedings contain no claim against Deloitte, and in these proceedings, it is the fact that the misrepresentations were made, and not any issue as to Deloitte’s reliance, which is in issue. In other words, the fact that Deloitte acted does not excuse the way the matter was presented to them in the Q2 2010 Note (see paragraph 1422 above) or justify an accounting treatment on a false basis.

Further as to (b) in paragraph 1439 above, and in any event, although the Claimants have presented the question as being whether reliance on Deloitte is a trump card for the Defendants, the real question is whether the fact that Deloitte and the Audit Committee approved both the accounts and the narrative demonstrates honesty on the part of the Defendants: and the truth as I find it is that what was stated in the Q2 2010 Note was false to their knowledge.

Mr Hussain’s Q3 2010 Note

The Claimants’ allegations in respect of statements made in Mr Hussain’s Q3 2010 Note that the strategic package sales had been “flagged” in the Q1 and Q2 results presentations, mirrored those concerning his Q2 2010 Note and discussed above. The best Dr Lynch could offer to support the statement in Q3 2010 that the hardware sales had been ‘flagged’ in Q2 2010 was to say that in the Q2 2010 Earnings Call “we’d talked about Arcpliance, which is a type of strategic package sale”.

However, not only was there no record of any such reference to Arcpliance in the transcript of the Q2 2010 results presentation, but Dr Lynch had told one of the analysts, in response to a question as to the effect on gross margin of the sale of Arcpliance inventory, that what he described as “the appliance model” would “account for about 1%...” and was an “exceptional situation” . There was thus a fundamental inconsistency in Dr Lynch’s position: even if ‘appliance’ sales had been flagged, his minimisation of them destroyed any argument that thereby the very considerable hardware sales had been flagged also.

The Claimants relied on this as a further demonstration that Dr Lynch was determined not to disclose the extent and nature of the hardware reselling, and would resort to very incomplete answers if that is what he felt was required. I agree and find accordingly.

The Claimants alleged that Mr Hussain also misrepresented the position in his Q3 2010 Note in claiming that “lower margin sales have generated significant new software business for us”. They tied this in with statements in Deloitte’s report that management’s linkage analysis showed a “strong linkage between the loss making hardware sales and subsequent highly profitable software sales.” They made the further point, which was not contradicted by the Defendants, that the Audit Committee was not provided with the Linkage Analysis in that quarter or any subsequent quarter, and were reliant on the summary provided by Deloitte and Mr Hussain. The Claimants alleged that Mr Bloomer and the Audit Committee had been misled into a misunderstanding that “the linkage analysis showed that hardware sales to a particular customer had driven software sales to that customer”.

The Claimants made two further allegations in respect of Mr Hussain’s Q3 2010 Note. The first concerned the bonus payments to Mr Sullivan. Mr Bloomer told me in cross-examination that he was not aware either of the payments, or that they were calibrated according to the amount of revenue Mr Sullivan generated from hardware sales. But the Claimants did not suggest that this made any particular difference.

The second further allegation relating to the Q3 2010 Note concerned the costs allocation issue. The Note went on to state , under “ Gross margins ”, that:

“ We have charged the cost of the lower margin sales to the cost of sales line even though we had agreed with our suppliers that the 50% of the cost would be used for marketing purposes ”.

The Claimants alleged that this too was false: no such agreement had been reached with any hardware supplier; and not all of the cost of the pure hardware sales was included in “ the cost of sales line ”. Mr Bloomer stated that he thought that this passage may have “ related to prior periods ”, but the Claimants alleged that it was not true in relation to earlier periods either.

I agree that the Q3 2010 Note misstated the position. As previously noted, Mr Hussain habitually and incorrectly presented the justification for allocating part of the costs of the hardware sales to Sales and Marketing expenses as lying in a bilateral agreement, rather than in the (obviously unilateral proxy) “residual method” which I have described previously as a proxy in the absence of any such agreement.

Deloitte’s own report on their Q3 2010 review to the Audit Committee likewise re-hashed the false line that “these types of sales were flagged in the Q1 and Q2 results presentation so the market is aware of them.” Deloitte also stated in their report that management’s analysis showed a “strong linkage between the loss making hardware sales and subsequent highly profitable software sales”.

I took from Mr Bloomer’s evidence when cross-examined that, on what they were told, his mindset and that of his committee continued in this quarter to be that on the basis that (a) the hardware sales did not cause Autonomy to be a “hardware seller” in the sense of running a separate business not incidental to the IDOL software business so as to upset the segmental analysis, and had also been “flagged” and (b) the issue of costs allocation involved amounts that were “tiny”, “just immaterial” , there was no real issue or problem.

More generally, in my view, both Deloitte and the Audit Committee settled into the view from Q2 2010 onwards that subject to (a) taking steps to restrict the allocation of costs to Sales and Marketing to amounts which were comparatively immaterial (b) requiring some paperwork to lend some general support to the notion of the hardware sales assisting the software business and (c) repeating in each quarter their advice to make disclosure, whilst accepting that ultimately this was a commercial decision for Autonomy’s board, the hardware reselling was not a matter of concern.

What would have awakened Mr Bloomer and the Audit Committee’s concern was anything which suggested that the hardware reselling was in the nature of a separate business or otherwise than intended to drive software sales, or in any sense clandestine. Mr Bloomer made clear that:

(1) He never thought that hardware reselling was a “secret”; he assumed (including from what he described to me as “meetings with at least one of the shareholders where the shareholder was aware of the hardware sales” ) and “that there were a variety of things which to me seemed that the market knew about hardware” ; but he acknowledged later in his cross-examination that he was not suggesting that the market was aware that in 2010 Autonomy resold over $100 million of third party hardware “without any Autonomy software” ;

(2) Had he appreciated that, contrary to what was stated in the Q3 2010 Note and Deloitte’s Report, hardware sales had not been “flagged” to the market, he would have wanted to know why they had stated they had;

(3) He was not aware that every quarter management would set Mr Sullivan revenue targets and that Mr Sullivan was regularly asked by management to generate further revenue from hardware sales, particularly towards the end of the quarter, nor that bonus payments were offered and paid to Mr Sullivan linked to the amount of revenue that he generated from hardware sales;

(4) He was not made aware (and nor according to Mr Welham’s witness statement was Mr Welham) that separate revenue targets were set for hardware each quarter, nor that Dr Lynch received updates on the revenue from hardware sales as each quarter progressed.^{^[215]}

What the Q3 2010 Note and Deloitte’s report did in the round was confirm the Audit Committee’s mindset that the hardware sales were part and parcel of the software business, and that the issue as to costs allocation was an immaterial detail which could be left to Deloitte. Nothing was said to unsettle the Audit Committee’s preconceptions, or to suggest the need for discussion and review of the real issues raised by the continuation of the hardware reselling strategy, including (a) what more exactly the market had been told (b) what, if the market remained unaware of the nature and extent of the programme and its contribution to overall revenue the reaction of the market would be if it were to be revealed (c) why, if immaterial, the issue of cost allocation was of such apparent and abiding importance to management, (d) why a proper Linkage Analysis had not been provided, and whether any linkage demonstrated was sufficient to justify loss-making sales of non-IDOL goods ^{^[216]} and more generally (e) whether the programme, as it had developed, could properly be said to be wholly focused on driving software sales.

Mr Hussain’s Q4 2010 Note

The Claimants’ allegations in respect of Mr Hussain’s Q4 2010 Note largely mirrored their allegations in respect of his previous Notes in the Relevant Period.

As in previous quarters, in Q4 2010 the Audit Committee were led to believe that there was a strong linkage between hardware sales and software sales. Thus:

(1) Mr Hussain’s Q4 2010 Note stated, under “ Gross margins ”, that the “ lower margin sales have generated significant new software business for us ” (emphasis added). Deloitte stated in their report to the Audit Committee for Q4 2010 that management’s analysis demonstrated a “ strong linkage between the loss making hardware sales and subsequent highly profitable software sales ” (emphasis added).

(2) Again, as in earlier quarters, the Note stated, in the “ Strategic sales ” section, that “ these types of sales were flagged in the previous quarters’ results presentation so the market is aware of them ” .

(3) Mr Hussain’s Q4 2010 Note went on to state:

“ We sold to Bank of America following on from the software sales made this quarter and several deals in q1. … Large deals in Q4 included a large sale to Bank of America (we won the Merrill Lynch legacy data migration to our hosted digital safe environment). This is a major win for Autonomy as we are now a major strategic supplier to the BofA worldwide ”.

Notably, Mr Bloomer and the Audit Committee were given to understand that the analysis showed a strong linkage between the hardware sales and subsequent software sales. He confirmed that the Audit Committee was not provided with the Q4 2010 linkage analysis prepared by Autonomy, which he understood had been reviewed by Deloitte.

That might have led to a train of enquiry which would have revealed, as was in my judgment the fact, that none of the statements summarised in paragraph 1456 above could be justified:

(1) There was no linkage between the hardware sales and software sales in any causative sense.^{^[217]}

(2) The market had no visibility in relation to the substantial pure hardware sales that Autonomy had undertaken: they were not “ flagged ” in Q1 or Q2 2010, nor does it appear from the transcript that they were disclosed during the Q3 2010 earnings call. When faced with this in cross-examination, Dr Lynch suggested that the hardware sales were “ mentioned in Q3 2009 ”:

“ I think there was language added to the Q3 earnings call about something along the line of strategic sales ”.

No such language was identified. There was no reference to hardware sales, whether “ strategic ” or otherwise, during the earnings call for Q3 2009 . There were the following cryptic sentences in the Q3 2009 Quarterly Report:

“ During the quarter we saw some of our large customers promote Autonomy to strategic supplier status. This has led them to adopt a broader set of our solutions in a number of significant deals ”.

But without prior awareness it would be difficult to extract from those sentences that Autonomy was making hardware sales.

(3) As to (3) in paragraph 1456 above, no software deal was concluded with BofA in Q4 2010. As already mentioned in paragraphs 1154 to 1155 above, Autonomy had been pushing hard to close a deal with BofA. When a BofA deal could not be completed in the quarter, Autonomy instead entered into deals with two VARs, DiscoverTech and Capax Discovery, on the last day of the quarter, 31 December 2010. Dr Lynch attempted an argument that in conversations between him and Mr Mackenzie-Smith (head of BofA in London) he reminded Mr Mackenzie-Smith of the discounted hardware sales Autonomy had arranged for it: but I have determined that even if they took place, such conversations had no material effect, and more generally that the hardware sale to BofA in Q4 2010 played no part in the deals with DiscoverTech or Capax Discovery, nor in the subsequent direct deal between Autonomy and BofA.

On the issue of cost allocation, and similarly to the previous quarter, Mr Hussain’s Q4 2010 Note stated, under “ Gross margins ”, that:

“ We have charged the majority of the cost of the strategic sales to the cost of sales line even though we had agreed with our suppliers that the 50% of the cost would be used for marketing purposes. ”

However, no such agreement had been reached with any hardware supplier; nor were the sales “ strategic sales ”; they were sales of (adopting the Claimants’ phrase) ‘pure hardware’ .

It follows that Mr Hussain’s Q4 Note was misleading, intentionally so in my judgment. As in the case of his previous Notes, Mr Hussain’s Q4 2010 Note and Deloitte’s own report were calculated to reinforce the Audit Committee’s preconceptions and the Defendants’ justification of the hardware sales: the Note does not establish the falsity of that justification, but the justification being false, it once more perpetuated it.

Before turning to Mr Hussain’s Q1 2011 Note, a point of detail which arose at this time in relation to continuing review of the issue of segmental analysis (see paragraphs 854 to 863 above)^{^[218]} illustrates the disregard for the truth that had developed within the cabal whenever it was necessary to depart from it to achieve a desired accounting treatment or approval. On 18 January 2011, Mr Chamberlain provided a paper to Deloitte which was intended to support the conclusion that Autonomy had only one operating segment. One of the points made in the paper was that “ No information is provided to MRL on hardware revenues ”. That was false. As set out in detail above, Dr Lynch discussed and agreed the hardware revenue targets in each quarter, and Mr Hussain provided Dr Lynch with regular updates on hardware revenue throughout each quarter. It is clear that Deloitte were misled in this regard: Mr Welham’s unchallenged evidence is that he was not aware that separate revenue targets were set for hardware each quarter or that Dr Lynch received updates from Mr Hussain on the revenue from hardware sales as each quarter progressed.

Mr Hussain’s Q1 2011 Note

In addition to familiar complaints about references to “strong linkage…and subsequent highly profitable software sales” , to which the answer was the same as in earlier quarters, the Claimants focused especially on words added in the final version circulated to the Audit Committee to the draft Note sent to Dr Lynch (which I have underlined below for the purpose of identification):

“ strategic package sales that included approximately $20m (9% of total sales) of lower margin business included JPMC and Bank of New York. These sales are part of strategic sales to these companies — for example JPMC is one of our largest customers of compliance software. ”

The Claimants took issue with the underlined words. They suggested that those words had been put in at Dr Lynch ’ s request, and they alleged that the suggested strategic link was spurious, not least because no software deal had been entered into by Autonomy with JPMC in Q1 2011 (it had happened much earlier, in Q2 2010).

In addition to their submission (which I accept) that careful consideration of the sequence of drafts shows that Dr Lynch was unlikely to be the person who inserted the words,^{^[219]} the Defendants submitted that in any event, the words are unobjectionable: contrary to the Claimants ’ suggestion, they did not say that there was a software sale to JPMC in Q1 2011, and any linkage which might be taken to be suggested was not inaccurate.

That submission did not, however, negate the fact that it was the overall message which was received which counts; and that message perpetuated and reinforced the Audit Committee ’ s mindset. Both message and mindset can be seen from Mr Bloomer ’ s evidence in his witness statement. Mr Bloomer explained that in Q1 2011 he understood that:

“ As they had done in their previous reports, Deloitte reviewed Autonomy ’ s strategic hardware sales and concluded that the linkage between the loss making strategic hardware sales and subsequent profit-making software sales justified the allocation of the loss to sales and marketing expense ”.

As before, the Audit Committee was not provided with the linkage analysis for that quarter. Mr Bloomer said that both he and the other members of the Audit Committee would have relied on what Deloitte and Mr Hussain had said in their respective reports regarding linkage.

Mr Hussain ’ s Q2 2011 Note

The position in relation to Q2 2011 was much the same. The pattern had been set, the message and the mindset established. In the usual way, Mr Hussain produced a Note for the Audit Committee meeting for Q2 2011, which he shared in draft in advance with Dr Lynch, seeking his comments. Both the draft (which Dr Lynch had reviewed) and the final version presented to the Audit Committee contained the following passage:

“ Strategic package sales - new strategic package sales in the quarter included approximately $16m (6% of total sales) of lower margin business (JPMC, Bank of America and Bank of New York). In addition, we had deferred lower margin business of $5m from previous quarter due to delivery. These sales are part of strategic sales to these companies - for example at JPMC we sold ediscovery software, continued archiving services and strategic package sales. And at Bank of America we have sold web content management software, archiving and is one of our largest customers of compliance software.”

The Claimants advanced two specific allegations:

(1) First, that the statement that $5 million of hardware sales had been deferred from the previous quarter “due to delivery” was false. At least some $2 million of hardware revenue was not recognised in Q1 2011 because the Defendants chose instead to recognise the DiscoverTech/Prisa VAR transaction. The Claimants contended that this had nothing to do with the delivery of hardware, and was instead driven by the desire to achieve a particular level of revenue without unduly affecting gross margin and earnings per share. The Claimants relied on Mr Bloomer’s acceptance that he had understood that $5 million of hardware sales revenue had not been recognised in Q1 2011 because the hardware had not been delivered in that quarter and contended that, therefore, the Audit Committee were misled in relation to this issue.

(2) Secondly, that the Note was misleading in stating that the sales to JP Morgan and BofA were strategic package sales and in giving the impression that the hardware sales were linked to software sales in the same quarter. The Claimants contended that in truth, Autonomy did not enter into any software licences with BofA in Q2 2011: the only transaction between Autonomy and those companies in that quarter was a restructured hosting deal where JP Morgan were offered very substantial savings, and there is no evidence whatever to suggest that that restructuring of an existing hosting deal was in any way driven or impacted by any hardware deal that may have been done with JP Morgan during Q2 2011.

I have concluded in a previous section of this judgment (see paragraphs 1074 to 1100 above) that revenue was improperly deferred in Q1 2011, at the instigation of what I have called the clique or cabal of three, including Mr Hussain, and (so I have found) to the knowledge of Dr Lynch as its éminence grise (see paragraphs 1094 and 1099 above). I accept also, therefore, that in stating that the reason for the deferral was a problem with delivery Mr Hussain ’s Note was misleading to the knowledge of both Defendants.

In cross-examination, Mr Bloomer agreed that his understanding at the time was that deferral was in consequence of late delivery, and that if in fact it was not delivered late, obviously he would have wanted to understand what the reason was for the apparently unjustified deferral, and if an impermissible one it might have given rise to concern about the approach management was taking in relation to the recognition of revenue. He pointed out, however, that (a) he would have expected the issue as to the cut-off for the quarter (which he told me was always a point of debate which had to be sorted out) to have been carefully checked and verified by Deloitte and would have relied principally on that, and (b) an element which would have had to be considered was that of materiality: “I mean, 250-odd million of sales in these quarters, 5 million is not that material one way or the other”.

As to the depiction of the sales to JPMC and BofA as strategic package sales, I would accept that little more was claimed by Mr Hussain than that the hardware sales provided the context for, and may have helped in getting, the software sales. No causal connection was expressly claimed: indeed, the nexus asserted was that which might be claimed for any advertising campaign or customer-oriented effort to encourage sales or customer loyalty. I accept that the hosting restructurings offered sufficient incentive of themselves to make unnecessary any further incentive from discounted sales of hardware; but that does not necessarily mean that the context did not also include such hardware sales, nor that they did not provide a further incentive.

But once more, as it seems to me, this parsing of the separate statements overlooks the fundamental message being conveyed in Q2 2011 as in previous quarters, to the effect that hardware resales were driving software sales, and the rest was anodyne and immaterial detail. Once more the message reinforced the Audit Committee ’s bedded-in perception of the purpose of the sales, and ensured passive acceptance, rather than quizzical discussion.

The messa ge as the Audit Committee received it is again apparent from the evidence of Mr Bloomer:

(1) Deloitte’s report to the Audit Committee for Q2 2011 stated the following:

“Management has further extended its analysis determining the strong linkage between the loss making hardware sales and subsequent highly profitable software sales. This continues to show a high degree of correlation between hardware sales and much more profitable licence sales to the same companies” (emphasis added).

(2) Mr Bloomer understood that Deloitte had reviewed the linkage analysis prepared by management for Q2 2011, and that linkage was established. Again, management’s linkage analysis was not shared with the Audit Committee.

(3) Based on what the Audit Committee was told by Deloitte in their report, Mr Bloomer understood that the analysis showed a strong linkage and high degree of correlation between the hardware sales and subsequent software sales.

Conclusions in respect of Mr Hussain ’s Quarterly Notes

In the round, in my judgment:

(1) Mr Hussain’s Quarterly Notes embellished and amplified the message that both Deloitte and the Audit Committee had already accepted from the time of Mr Knights’ “conversion” (when he became a “wordsmith”).

(2) Deloitte were largely tied into their initial acceptance of the presented purpose of the strategy (in Q3 2009), and the Audit Committee followed their lead.

(3) The Audit Committee were in a sense distracted by conceiving as the principal issue whether or not the newly introduced segmental accounting provisions in IFRS 8 were being properly applied. Mr Bloomer’s witness statement, and to a great extent his oral evidence, appear to me to take as the litmus test of the propriety of the accounting and disclosure of the hardware sales whether by reason of their extent or otherwise such sales had come to constitute a separate segment of Autonomy’s business, necessitating separate identification, accounting treatment and disclosure of its performance.

(4) The Audit Committee had no idea, for example, that Mr Sullivan considered himself to be, and in reality was, “basically running a small side business” .

(5) Mr Hussain’s Quarterly Notes do illustrate yet again that management and one or more of the Defendants were content to dissemble if that was necessary to ensure that there was no further disclosure of the hardware sales programme which would have revealed, or would be likely to lead to a train of enquiry that probably would reveal, their true extent and contribution to Autonomy’s revenues.

It seems clear and I find that the predominant reason for dissembling, and for the false presentation repeatedly reinforced by Mr Hussain’s Quarterly Notes, was that those involved were concerned that revelation of the programme would be perceived in the market to diminish the quality of Autonomy’s earning record and prospects and would materially adversely affect analysts’ and the market’s rating of Autonomy.

Linkage Analysis

I have repeatedly referred in the preceding analysis to the Linkage Analysis and the way in which it was deployed by Autonomy’s management will by now already be apparent, as will be its purpose. In summary, it served as management’s response to:

(1) The shift from placing primary emphasis on the avowed “joint development programme” (in collaboration with EMC) as the principal justification for the programme to placing emphasis in justifying the hardware sales (in conjunction with Dell) on their success as an investment in customer relationships and in enabling Autonomy to become a one-stop supplier of both software and hardware to its customers, and

(2) Deloitte’s natural requirement for some demonstration of the nexus and growing concern that the hardware reselling strategy, which had initially been presented as a one-off initiative with EMC, was becoming “business as usual” (as Mr Welham put it in his witness statement).

In this part of my judgment, I address certain issues in relation to the Linkage Analysis raised by the Claimants, and in particular the issue as to (a) what it was designed to show and (b) whether it demonstrated the “strong linkage” and “high degree of correlation” between hardware and software sales which management asserted in its deployment.

The first is as to what Deloitte expected from it and its format. Although it seems clear that this exercise was in response to a request from Deloitte for an analysis of hardware revenues paid to customers to whom Autonomy also sold software, there was not in evidence any email or other document setting out the terms of the request, and there was a dispute as to what Deloitte gave Autonomy to understand was required.

It was Mr Welham ’s evidence that Deloitte requested (it is not clear when) what he described as:

Mr Welham was not cross-examined on this statement. However, in his oral closing argument, Mr Miles suggested that Mr Welham ’s description was not of the request but of what Deloitte received in answer; and that it accordingly amounted to no more than Mr Welham ’s own perception of the content and effect of the analysis: any request would have been in line with the simpler one already referred to (but made later in January 2011).

That would tally with Mr Stephan ’ s evidence in chief ( “ direct” evidence as described in the US) in the US criminal trial as to what he conceived his task in preparing the analysis to be. That evidence (which was adduced by the Claimants under a hearsay notice) was that the spreadsheet did not contain any real analysis of whether the hardware deal actually related to a software deal: it was simply (in Mr Stephan ’ s words):

“a mathematical exercise of matching names of customers…So if we sold a hardware to Customer A, we ’d be looking for Customer A through the historic revenue to try and find software sales to them.”

Ms Gustafsson ’s evidence of her understanding of the analysis, in its original form but also as updated, was to the same effect.

In conjunction with the dispute as to what Deloitte had asked for, there was also a dispute about what Deloitte really understood the analysis to show when they received it (and successive versions). The Claimants ’ case is that Deloitte were misled by the linkage analysis (which was updated quarterly) into thinking that there was a direct causal link between individual hardware sales and individual sales of software, and that then they and Mr Hussain together (unwittingly in the one case and dishonestly in the other) misled the Audit Committee into thinking there was “strong linkage”. The Claimants described it as “yet another exercise designed to disguise the true nature of what Autonomy was doing.”

Apart from what the spreadsheets themselves showed, the Claimants ’ case was very largely based on the evidence of Mr Welham. Mr Welham ’s evidence in his witness statement was that Deloitte understood that the linkage analysis prepared by management:

“ provided directional evidence of a causal link between the hardware sales and subsequent software or hosted revenues ”.

Mr Welham was not cross-examined on this either; but in his oral closing submissions, Mr Miles emphasised the words “ directional evidence” in Mr Welham ’s statement and submitted that they made clear that Mr Welham cannot have been suggesting that the spreadsheets constituted an analysis of the causal link between particular sales of hardware and particular software sales.

In the absence of cross-examination, it is not easy to determine quite what Mr Welham did mean. But I agree with Mr Miles that the two expressions Mr Welham used ( “ directional evidence” and “ causal link” ) cannot be taken together to denote “ causal link” . I think that the nearest synthesis is probably that the analysis suggested a correlation which might indicate, but did not establish, a causal link.

In any event, if Mr Welham was seeking to suggest that Deloitte understood that the linkage analysis provided evidence of something like a causal link, that is not what it appears others in Deloitte thought. In his oral closing submissions, Mr Miles referred me to notes of an FRC interview with Mr Mercer on 4 September 2013 in which Mr Mercer made clear that he and Deloitte appreciated the limitations of the spreadsheets. He plainly regarded the sales of hardware as being in large part a marketing exercise and emphasised more than once both the truism that “marketing is an art, not a science” and the aphorism “half of marketing works” : “You just don ’t know which half. So, so this is what they ’re doing and overall, we can see it is working.” When asked by the FRC representative “…what was this analysis of the linkage that management did?” Mr Mercer replied:

“It ’s analysis of sales that they made of hardware and losses historically…So, it is, “We have made these sales of hardware and losses to these customers. We have made sales to these selfsame customers of software of (blank). It ’s the same customers we are selling…that we are making this marketing effort and incurring this marketing cost by selling software at losses that we are selling highly profitable software to…” So it ’s not...you are absolutely right. You can ’t…show a hardware sale on 30 June 2010 and have an analysis that it ’s…since they started doing this in a significant way, they ’ve tracked who they ’ve sold this stuff to and them who they ’ve made - how much of the software posted revenues they ’ve made to the same people.”

In other words, Mr Mercer and Deloitte did not treat the spreadsheets as intended to show a causative link, or even a clear correlation between a particular hardware sale and a subsequent software sale: they treated them as showing that, even if a direct nexus was not capable of being demonstrated in particular individual cases, in overall terms the loss-making hardware sales appeared to lead to more high-margin sales to the same group of customers: as in all marketing. Mr Mercer agreed “absolutely” that it was not more than “sort of a reasonable correlation” across the cohort of customers, adding:

“It ’s by no means perfect and you can ’t see one following the other but you can see that the vast majority of these companies that they are selling hardware to, they are now selling significant amounts of software to…”

That accords with Dr Lynch’s understanding ^{^[220]} :

“A. … it was like when you do advertising, you don't know exactly which sale comes from the advertising but you can do something like - you can look at the amount of advertising in a region and then see what sales you have in that region. It doesn't mean that there's a causality on a one-to-one deal basis.

…

I think they [i.e. Deloitte] considered there was a link but not on an individual basis, in the same way as there would be a link in advertising something in Northampton and then looking at the number of shoe sales in Northampton. That was my understanding. ”

More difficult to answer in those circumstances is the question why this broad correlation, as both Autonomy and Deloitte do seem to have perceived was what the linkage analysis was designed to show, came to be presented by Autonomy to Deloitte and by Deloitte to the Audit Committee as demonstrating a “strong linkage” and their resort to it as a significant factor in justifying the allocation of costs to Sales and Marketing expenses.

In that regard, it may be recalled that Mr Hussain ’s Q2 2010 Note for the Audit Committee asserted that the “ lower margin sales have generated significant new software business for us ”; and Deloitte ’s report to the Audit Committee for Q2 2010 stated:

“These are further examples of a number of strategic hardware transactions [sic] completed by Autonomy to major international banks or other large blue chip companies completed in order to open up new market opportunities and to become the preferred supplier for all such clients’ archiving requirements, including both software and hardware.

… Management ’s rationale for entering into these loss making contracts is that Autonomy is seeking to develop a long term strategic relationship with the end-users in order to secure future profitable software sales.

Management has prepared an analysis demonstrating the strong linkage between the loss making hardware sales and subsequent highly profitable software sales …”

In my judgment, it is clear from the contemporaneous documentation, taken as a whole, that both Autonomy and Deloitte understood the limitations of the linkage analysis, but nevertheless deployed it to support the depiction of the nexus between hardware and software sales urged by Autonomy ’s management. This was to the effect that there was a sufficient correlation between hardware sales and subsequent software sales to the same customers not only to justify taking a loss on the hardware sales, but also to justify an accounting treatment which would allocate a proportion of the costs to Sales and Marketing expenses.

For example, in an internal Deloitte email dated 14 October 2010, Ms Antonia Anderson said of the linkage analysis that Mr Stephan had sent her earlier that day that “ [t]he important thing on this tab is their analysis of the software and hosted sales they generate with customers that they sell this hardware to ”. Similarly, in his email to Mr Stephan on 20 January 2011, Mr Leo He of Deloitte requested a copy of the linkage analysis in order to “ justify the rationale that the purpose of Autonomy selling hardware at a loss is to generate significantly more software revenue from these customers, and therefore it is reasonable to classify the hardware loss to be S&M expenses ” . Mr Welham went further in his witness statement and said that when Deloitte accepted that the loss suffered by Autonomy on the hardware sales could be allocated to sales and marketing expenses, with the balance of the cost of the hardware being allocated to COGS, they did so:

“ on the basis of the linkage analysis provided to us by Autonomy each quarter, which, based on the representations Autonomy management had made to us about the business strategy, appeared to supply adequate support for the allocation of the loss on the hardware sales to sales and marketing expenses ” .

I consider and find that there were two main reasons why Deloitte in fact treated the linkage analysis as offering more support than it was designed to show and on its face it did:

(1) First, I agree with the Claimants that it is important to remember that Deloitte did not receive the linkage analysis in isolation. Autonomy had stressed to Deloitte that its strategy of reselling hardware at a loss was in order to become “ the single source supplier for all of the major banks in relation to [their] data management activities ” and that the net loss represented “ an investment in the customer relationship ” which had “ helped enormously in the procurement of significant software sales ”. (emphasis added). That was Deloitte’s mindset; and, read in that light, the Linkage Analysis complemented and reinforced it.

(2) Secondly, the truth is, as it appears to me, that Mr Knights was very open to persuasion: my impression is, as I have previously indicated, that by now he had changed from sceptical assessor to become an active participant in finding a way to justify the allocation of costs to sales and marketing expenses which would satisfy the Deloitte Reviewers. I have in mind particularly an email dated 18 July 2010, from Mr Knights to Mr Hussain, Mr Mercer and Mr Welham (copied to Mr Chamberlain). In that email, he explained that he, Mr Mercer and the Reviewers had had to draw the line when Autonomy’s management had tried once more to get approval for 50% to be allocated as such expenses (which he described as “not something our compliance people will get comfortable with…we only got comfortable with this in Q1 on immateriality grounds” ). Whilst that might have appeared to show (and both Mr Knights and Mr Mercer have since insisted that their refusal to accept the 50% split did show) the audit function properly working, he then immediately put forward what was in effect a “work-around” with the same objective. Under a heading in bold type which read “But there is a solution that makes sense - particularly as in the Q1 call you already highlighted the $10m of hardware in inventory which you highlighted was to be sold in Q2” he laid out the following plan:

“My solution would be:

· Record the hardware sales at nil gross margin for IFRS reporting

· Take the “ loss” as selling expense - (around $4-5m I think)

· The market already knows that you will be making Q2 hardware sales as you highlighted this at Q1 and had inventory on the b/s. So any IFRS gross margin one off drop is reasonable and can be explained as part of the strategy.

· In the adjusted gross margin [original emphasis] strip back out the hardware element to a “normalised” level and add an explanation –

By the time you wrap up the $10m hardware b/f and the $4-5m that is in selling expense surely we are almost there??

Just to be totally clear all of us fully get the strategic element to this and the opportunity to open up new markets. The evidence of follow through sales is apparent.”

(Mr Welham ’s again unchallenged evidence (which I accept) was that the last sentence was a reference to the linkage analysis that had been supplied to Deloitte by Autonomy management.)

In summary, therefore, it was not so much the linkage analysis itself as its context and presentation, together with Mr Knights’ readiness to accept Mr Hussain’s depiction of it as in some way (unexplained) showing “strong linkage between the loss making hardware sales and subsequent highly profitable software sales” (to quote from Deloitte’s Q2 2010 Quarterly Review), which persuaded Deloitte to regard the linkage as sufficient for them to approve Autonomy’s allocation of a significant proportion of the costs of the hardware sales to sales and marketing. That was vital, as I have emphasised before: it was that which ensured that the hardware reselling did not appear to have a concerning adverse effect on gross margin such as would be likely to excite further enquiry, and thus at one and the same time both ensured full realization of the true purpose of the hardware reselling and so much helped to disguise and attenuate the likelihood of inquiry about its true extent.

Other examples showing the determination to avoid disclosure to the market

Finally, in assessing the purpose of the hardware reselling strategy I turn to discuss three aspects of the Defendants ’ conduct especially relied on by the Claimants as showing that both Defendants were prepared to resort to lies about the true extent and nature of the hardware sales in order to avoid their disclosure or enquiry which would result in the market becoming aware of them and their improper purpose, and thereby Autonomy ’s dependence on hardware sales to maintain its image as an extraordinarily successful “pure software” company.

Q&A scripts

First, I address the scripts prepared by Autonomy and management in advance of every earnings call. These included draft “Qs and As” worked on after ‘brainstorming’ sessions to ‘spot’ topics of likely interest or concern to those attending the calls. Many of the questions ‘spotted’ were never asked, nor therefore were the draft answers given. Obviously, a scripted answer never in fact uttered and never made available cannot be the basis of any claim under FSMA or in misrepresentation. However, the Claimants relied on them nevertheless as evidence of what would have been said if the questions had been asked, and thus of the Defendants’ state of mind and intentions.

More particularly, the Claimants relied on individual scripted answers as evidence of (a) a willingness on the part of both Defendants to mislead the market and (b) the fact that neither of the Defendants had any honest belief that the hardware sales were known to the market, or that their purpose was legitimate, or that there was any proper basis for not disclosing them fully.

The Defendants contended, to the contrary, that the scripts could not safely be relied on as evidence of what would have been said if the questions were asked. According to the Defendants, the answers relied on by the Claimants in the scripts were drafts, which went through many iterations and which were prepared by a large number of people, including Autonomy’s brokers and investor relations. Scripted answers might or might not reflect input from the operational part of the company to which they related; might or might not have been checked by the Defendants; and might or might not be used. On this view, they were little more than placeholders.

Dr Lynch’s elaboration of the process was as follows:

(1) The process involved some 20 people (some junior or only “interns”) “brainstorming” with a view to identifying any questions, and generated a mass of them which accumulated over time, many of which might remain on the script unless and until focused on at a second stage of review: “some of the questions…hang around for years” and the scripted misleading answer may have been a hangover from previous versions of the Q&A script which had not been reviewed or had simply been insufficiently checked and verified.

(2) Some questions were “too sensitive” for the large group and so a smaller group of people, who were privy to “inside information” would undertake a “final phase” where the “actual” answers would be developed, sometimes at a stage too late to be included in the printed version so that the final answers had to be inserted in handwriting. The scripted answer may in effect have been a placeholder (my word, not his), which would have been amended and changed, had it been focused on, and would not have been used in anything like that form if the question had been asked.

(3) Even after review of the drafts there was a further important stage:

“Then the last stage, which is the sensitive one, is those documents are taken and then the questions are amended for things that we know that the rest of the group don't know that are important. And that would be any inside information or anything that we think is competitively very sensitive. So that will not have been surfaced and taken out. The answers that will have been put in in the drafts would have been ones just to move on to the next question, because obviously you can't highlight this to the brokers and the other people that are sitting there. And then we come up with an actual set of answers that we would use. And that's why the answers in those documents are not the same as the answers you hear when the questions are asked sometimes on the call because there's actually a final version that's done.”

The last amendments would not be made in the printed version, but in manuscript. Thus, the Defendants submitted, the inclusion of a proposed response on a particular draft cannot be relied on as showing anyone speaking on the call intended to use the response in question: and, they concluded, the only responses that can be relied on - and the only information of any possible relevance to the claim before the Court - are those that were in fact given on the calls. Furthermore, in the case of each of the scripted answers on which the Claimants sought to rely, Dr Lynch’s evidence was that the statement in question was not in fact, and would not have been, used.

The Claimants submitted that Dr Lynch’s repeated efforts to dismiss the scripts as containing draft answers which were later revised and which in any event would never have been (and in the event never were) used as an elaborate concoction which Dr Lynch knew to be patently untrue. They submitted further that Dr Lynch’s resort to lying about both the process of developing the answers and the use to be made of them when cross-examined not only fortified their main argument as to the insight shown into the mindset of both Defendants, but also was fundamentally revealing of Dr Lynch’s personal dishonesty and his propensity to lie to the court as well as the market.

This aspect thus acquired more importance in a FSMA claim than might be expected of representations never in fact uttered. Indeed, the Claimants spent considerably more time on the Q&A scripts, which were not mentioned in the pleadings, than on what was actually said at the Earnings Calls. They cross-examined Dr Lynch at length on the process by which they came to be finalised for use in each quarter in the Relevant Period, and on the content of the Q&A drafts (with particular emphasis on the scripts prepared for the Q3 2009 Earnings Call and the Q2 2011 Earnings Call, marking the beginning and end of the sequence in the Relevant Period).

I do not think it is necessary to go through each script quarter by quarter. Instead, I examine first, the dispute as to the way the scripts were produced and my conclusion in that context, before discussing various illustrations of the questions and answers instanced by the Claimants as obviously false and misleading from the Q&A scripts prepared for Q3 2009, for Q1 2010 and for Q2 2011, and Dr Lynch’s responses in respect of them. I then set out my conclusions as to the allegations made in respect of the Q&A scripts compositely.

The documentary evidence seems to me to reveal that:

(1) The final printed version of the Q&A document typically went through a number of drafts: for example , that for Q3 2009 appears to have gone through at least 15 drafts;

(2) Autonomy’s Investor Relations (headed by Mr Geall, working with Mr Goodman) would typically “own” the document, with suggestions and amendments channelled through them;

(3) Financial Dynamics (Autonomy’s external financial PR advisers) would also be involved;

(4) The evidence of larger meetings of 20 or so is sparse. Although in Q3 2009 at least one meeting seems to have taken place, where individuals not in core management may have been present to throw around questions and answers, there is no clear evidence of any such meetings in later quarters, and it is not easy to see when they would have fitted into the chronology;

(5) Even if a group of 20 or so people did assist in “brainstorming” sessions to identify possible questions and answers at the inception of the process, it is plain from the little evidence that there is (confined to Q3 2009) that they took place very near the beginning of the process of developing the script;

(6) The process was soon in the hands of a small group and the drafts were worked up over email, as opposed to in much larger meetings;

(7) Further, in every quarter, the Defendants seem to have been engaged in the final drafts: and in Q1 2010, for example, the source of the questions and draft answers highlighted by the Claimants appears to have been them;

(8) Certainly in Q1 2010, the source of the objectionable/ noxious questions and answers seems to be Dr Lynch and Mr Hussain. As late as 8.03pm the day before the Earnings Call, Dr Lynch was editing the document containing those answers and not correcting them. It seems implausible that any meeting was going to take place after that time when any analysts or interns would see that document;

(9) There are no extant copies in evidence of Q&A scripts with manuscript amendments; and typically, the final printed version would have little room for the insertion of manuscript additions;

(10) There does not appear to be any evidence that Deloitte saw the scripts or the Q&A drafts (or final version) prior to the quarterly earnings calls.

The following scripted questions and answers prepared for the Q3 2009 earnings call (numbered 27, 37 and 107 in the final printed draft) are examples particularly relied on by the Claimants as demonstrating that the Defendants intended to give seriously misleading answers to key questions on hardware sales which had been identified as likely to be asked:

27. How much h/w did you sell in the quarter?	We do sell a bit of H/W every quarter eg telephony cards as part of our call centre solution
37. You used to give licence organic growth in the past. What is it and why don’t you give it now?	All we sell is IDOL so that is the organic growth number. The reason for the old categories was when we had legacy things like ultraseek. That is no longer relevant.
107. Do you include hardware revenues in Hosted?	There are some instances where we might sell an appliance solution. But this is rare and the quantum is low as a percentage of group revenues.

The contemporaneous documents show that both Defendants were closely involved in the preparation of the final script. Dr Lynch had to accept that had the answers scripted been given, they would have been seriously misleading.

In addition to his explanation as to the process (see above) and his ultimate defence that the scripted answers would never have been used, Dr Lynch suggested that the Q&A scripts also tended to accumulate historic answers which were (as he described them) “hangovers” from previous quarters which were no longer to be used. But he had to retract the suggestion when a further document was put to him later in his cross-examination.

In respect of the Q&As prepared for the Q1 2010 earnings call, the Claimants relied particularly on what they submitted were a series of misleading or diversionary answers to be given to hardware-related questions that had been picked out as likely to be asked. These included:

(1) Q&A 15 was “How much was the hardware related sales incurred in Q1?” to which the scripted answer was “[$5]m went into the solutions.”

(2) Q&A 16 was “Hardware: Why?” to which the scripted answer was “Arcpliance. Up to customers strategic eca”.^{^[221]}

(3) Q&A 19 was “HW: What was it in the past?” to which the scripted answer was “No idea”.

(4) Q&A 71 was “ You used to give licence organic growth in the past. What is it and why don’t you give it now? ”, to which the scripted response was “ All we sell is IDOL so that is the organic growth number. ”

(5) Q&A 85 was “ Why haven’t you disclosed the hardware sales in the past? ”, to which the scripted response was “ Not a material part, normally pure software, sometime software and services or software and hardware…new move to arcplaince. ”

(6) Q&A 108 was “ Inventory - have you done many hardware deals in the past ”, to which the scripted answer was “Not usually a significant amount.”

Of these answers, all but (1) and (4) were exactly as drafted by Dr Lynch, except for the correction of typographical errors which were typical of his usually hurried emails. (1) and (4) also appear likely to have been largely drafted or suggested by him as they were inserted further to a final email on the Q&A script from him to Mr Hussain on the day before the Q1 2010 earnings call.

Turning to the Q&A script prepared for the Q2 2011 earnings call, the Claimants relied particularly on the following:

(1) Q&A 24 was “ Hardware sold in the quarter ”. The proposed response was as follows: “ Arcpliance sales were low well under their usual range of a couple of million. 1 would note that this is dissimilar of what they were a year ago ”.

(2) Q&A 25 read “ Do you sell hardware other than arcpliance ”, to which the scripted answer was: “ Yes occasionally for example we sell a tiny amount of memory cards for PCs that go into call centres ”.

(3) Q&A 66 posed the question “ Cost of sales - how much was hardware? ”, to which the scripted response was “ Arcpliance sales 2.5 to 3 [SH to confirm] ”.

All of these Q&As were included in an attachment to an email from Mr Hussain to Dr Lynch, Mr Kanter, Mr Chamberlain and Mr Derek Brown (who was also in Autonomy’s Investor Relations department) dated 26 July 2011.

Dr Lynch, with some surprising equivocation that it might depend on the context, had to accept in cross-examination that if those answers had been given, they would have been misleading. Plainly they would have been so: in Q2 2011, revenues from hardware sales amounted to about $20.8 million, compared to total revenues of some $256 million.

Dr Lynch sought to defuse this on the same basis as before: (a) the answers had not been used, and so could not have been misleading; and (b) the answers did not show what would have been said:

“You’re asking me about an answer that may be generated by a 22 year old intern…This is a document which is produced and worked on by 20 people around a table, including in some cases even interns, 22 year olds, people throw in the questions, we write answers. In the answers we do not give inside information because it is - and it includes external people to the company as well. And then this is taken and the actual answers are done and the proof of that is very simple, which is the answers in this document are not the ones you see when questions are asked on a conference call. And that’s true for the last, you know, whatever it is, 40 quarters.”

The Claimants submitted that these efforts were confounded by the documents: far from the scripted answers having been “ produced and worked on by 20 people around a table ”, they were contained in a draft which Mr Hussain described as “ my latest ” and which he sent to Dr Lynch, Mr Kanter, Mr Chamberlain and Mr Brown only. Furthermore, and as noted above, the fact that Autonomy was selling large quantities of pure hardware in order to report what would appear to be higher levels of software revenue was not “ inside information ”. Dr Lynch’s answer that the scripted answers were never used missed the point, which was that the question was never asked.

When put to him that the position that was being taken in these responses was dishonest, Dr Lynch said as follows: “ If it had been dishonest, the answer would have been used. The answer was never used ”. But, yet again, that misses the point: as it transpired, the question was never asked but the existence of the planned answer suggests, strongly to my mind, that Dr Lynch and Mr Hussain were prepared dishonestly to mislead the market.

Dr Lynch’s broader point that the transcripts prove that questions which were asked were often answered differently than appeared on the prepared script is less easily dismissed and in fact was not addressed by the Claimants. Dr Lynch did tend to depart from the script: indeed, he almost invariably did so, providing much more detailed answers than had been suggested. For example, questions in Q2 2011 about the growth of Cloud business and developments in Promote were dealt with by Dr Lynch in a far more detailed way than the scripted answers (and indeed some of the questions were left blank in the script).

I have concluded and find as follows:

(1) Even if there were meetings of 20 or so people including “interns” to “spot” questions and suggest possible answers, such meetings had very little influence on the shape and content of the finalised scripts, which had gone through a number of drafts and reflected the input of Autonomy’s Investor Relations department, Financial Dynamics and ultimately, and most significantly, Mr Hussain and Dr Lynch. I cannot accept that the scripted answers were the work of interns: they were the work of the Defendants.

(2) There is nothing to support Dr Lynch’s evidence that the scripts were or would have been amended in manuscript. The Defendants’ revisions of the relevant scripts typically continued until the eve of the earnings call in question. There was no reason for such a process: and there was no space provided, nor any time for such an exercise.

(3) I do not accept either the suggestion that placeholder answers were necessary to protect confidentiality or reflected the fact that those compiling them had no access to confidential information. There was nothing confidential in the answers; and they were in many instances drafted by Dr Lynch himself.

(4) Dr Lynch appears to have paid particular attention to the answers to be given to questions ‘spotted’ relating to hardware. His contribution to the Q1 2010 scripts is especially clear. The answers he scripted in this context would have been misleading if given.

(5) All that said, however, I doubt that the scripted answers show anything more than that Dr Lynch did not intend to disclose the truth about the hardware sales. I doubt that the answers Dr Lynch would have given would have followed the script had the question(s) been asked. Even where he had drafted the scripted response, I am doubtful that he would have used it: that was not his habit or style. As his performance in the witness box also showed, Dr Lynch is articulate and skilled at presentation. I think at most the scripted answers were aide-memoires, and sometimes I suspect little more than placeholders.

(6) For example, the scripted answer to Q&A 27 in Q3 2009 was a clumsy lie and Dr Lynch is not clumsy; neither would either Defendant, or any of those speaking on behalf of Autonomy, have fixed on that answer as likely to pass unnoticed by, or without comment from, Deloitte and Mr Welham (who listened into the earnings calls). Dr Lynch may well have hoped that the question would not be asked, did not want to commit to any focused answer in advance, and trusted himself to equivocate or distract attention if and when the question came, with the comfort of knowing that the practice was not to allow more than one follow-up question.

(7) Dr Lynch had taken the advice from Deloitte, as adopted by the Audit Committee, to be that “the information did not have to be disclosed” (as he stated when cross-examined): he would not simply have stated what the hardware sales were. Nor would Mr Hussain, who was very reluctant to say anything unscripted anyway: he would have referred the question to Dr Lynch, and said nothing more.

(8) The truth is that the Defendants were prepared to mislead to keep analysts and the market away from the nature and extent of the hardware sales.

(9) The Q&A scripts further undermined Dr Lynch’s reliability and credibility in my eyes. They support the conclusion urged by the Claimants as summarised in paragraph 1503 above.

Representations to disguise the hardware reselling strategy made in Earnings Calls

That conclusion is also supported by various representations actually made by the Defendants in the course of Earnings Calls in the Relevant Period. The Claimants’ pleaded allegations in this regard related to the earnings calls for Q3 2009, Q1 2010 and Q2 2010: they did not advance allegations in their RRAPoC or in their opening in respect of the other Earnings Calls in the Relevant Period. Mr Rabinowitz did, however, cross-examine Dr Lynch on the other earnings calls in the Relevant Period in support of the general case of dishonesty against both Defendants.

Whilst I have borne in mind the others, and have mentioned them in passing in the section on the chronology of Autonomy’s relationship with Dell and hardware reselling, it is for present purposes sufficient to focus in this context on Q3 2009, Q1 2010 and Q2 2010. Each one illustrates a different set of what the Claimants contended were misrepresentations to conceal hardware reselling and examples of serious dishonesty on behalf of the Defendants. I address at some length the Q3 2009 Earnings Call, which was principally related to the costs attributable to a new product of Autonomy’s called SPE. SPE was an issue of some importance in the Claimants’ case ^{^[222]} (though it was one which the Defendants tended to shy away from and seek to marginalise). The Claimants contended that (a) it demonstrated the lengths to which the Defendants would go to conceal hardware reselling and costs and (b) it reveals “a willingness on the Defendants’ part to tell lies to Autonomy’s auditors, the market and to this Court.”

Q3 2009 Earnings Call

In its Quarterly Report for Q3 2009, Autonomy had capitalised $11.7 million in R&D expenses. This figure was much higher than the R&D costs capitalised for the equivalent quarter the previous year which amounted to only $3 million. The Q3 2009 R&D figure was stated in that report to be “primarily due to the new IDOL SPE product reaching commercial exploitation phase.”

The principal focus of the Q3 2009 Earnings Call was Dr Lynch’s announcement of the launch of SPE. He described SPE during the call as “a complex piece of technology”, a “radical technology”, which could “clip on to pretty much any relational database” and was “probably the second most important product in the Company’s history”. He knew, he said, that it was “geek stuff at the moment” but it was soon, he hoped in 2010, “going to open up markets in a very different way for us.” He talked of a “very large addressable market” , an “$18 billion market”.

On the Q3 2009 Earnings Call, Dr Lynch told the market that the sales and marketing expenditure incurred in Q3 2009 in relation to SPE was “$18 million/$20 million, with about $4 million extra on the Quick Start program”. In explaining the significant increase in Sales and Marketing expenses, Dr Lynch talked of ^{^[223]} :

(1) the time spent on a BETA programme model necessary for such a complex piece of technology;

(2) the work done on “advertising in the trades” and working very closely with the (computer) analysts;

(3) the ‘Quick Start initiative’ for which he said there had been “over demand” ;

(4) tax breaks negotiated by Mr Hussain which had given “a little bit of leeway… to increase the Quick Start program considerably above where we expected it to be” ;

(5) “flying key customers from around the world to special selective seminars…small affairs…done for about 15 customers at a time….over two days… and that worked very well”

Mr Kanter added that the IDOL SPE launch had been “acclaimed by industry analysts”.

Mr Hussain said in seeking to clarify (a) the reasons for the reduction in gross margin in the quarter to 86% compared to 92% in Q3 2009 (as stated in the Q3 2009 Quarterly Report), and (b) the decline in Autonomy’s operating margin in Q3 2009 (which fell to 34%, compared to 42% in Q3 2008 and 47% in Q2 2009), as follows:

(1) “Moving on to gross margins. At 86% this quarter, we saw the effect of a couple of percentage points of the over-demand on the Quick Start program for the new product release” ; and

(2) “Operating margins were at 34.5%. This was affected by seasonality and the additional spend on the new product release . Now excluding the effects of the new product release, i.e. the additional marketing spend, the lower gross margin effect of the over-demand, and adding back the net effect of the capitalised R&D, then operating margins would have been 43%, around that number”.

He also clarified, in answer to an analyst’s question, that “the spend on SPE in Q3 was a one-off in Q3 only.”

The Claimants’ case can simply be stated: the R&D capitalised costs of $11.7 million did not arise in respect of the development or launch of SPE or its related product ‘Quick Start’^{^[224]}. They were a disguise to explain increases in marketing expenses and R&D costs, and decreases in gross margin, without any mention of the effect on these metrics of the rapidly developing and substantial, but loss-making, programme of hardware sales instigated further to the Loudham Hall meeting in July 2009.

According to the Claimants, who relied principally on the evidence of Dr Blanchflower, behind all the fanfare, SPE was not a new standalone product at all. It did not involve writing any new code. It was simply a relatively straightforward repositioning of IDOL’s existing meaning-based computing to structured data, and there were no additional statistical or probabilistic features in SPE which did not already exist in IDOL.

The Claimants’ case was that the real costs of SPE’s development were small: Dr Blanchflower estimated them at around $50,000. Other costs might take the aggregate to $100,000; but no more. The Claimants’ other main witness on the point, Mr Lucini, accepted a slightly higher total amount for development costs: $70,000, rather than $50,000, based on an industry charging rate.

The Claimants contended that the documentary evidence revealed no significant SPE sales and marketing expenses in Q3 2009 and that it is impossible for anything approaching $18-20m to have been spent on sales and marketing for the launch of SPE and that the true figure was unlikely to exceed $100,000.

They submitted that, with the exception of a small sum spent on advertising, there was no documentary support for there having been any expenditure on any of the specific categories mentioned by Dr Lynch in the Q3 2009 Earnings call. According to the documents, and the evidence of Dr Blanchflower and Mr Lucini:

(1) The total advertising spend was $78,342 and Dr Lynch personally limited such expenditure to no more than $80,000 in an email on 23 August 2009 to Ms Eagan.

(2) There were very few documented discussions with analysts. The Claimants accepted that there was a video conference held with Ms Sue Feldman of IDC on 14 September 2009 but submitted in their written closing that there is no evidence of any payment made in respect of this and that if there was it is likely to have been immaterial. The Claimants accepted that briefings were held with Mr Muncaster on 15 September 2009: but on the documents, that appeared to be that.

(3) There was no documentary evidence of any Beta program (being sample software testing) in existence in relation to SPE in Q3 2009, and the Claimants relied also on the evidence of Dr Blanchflower and Mr Lucini that there was none. They noted further that (a) when, following the press release on 16 September 2009, customers asked whether they could join the Beta program, they were told with Dr Lynch’s knowledge and/or by his direction that the Beta program was closed due to over demand; and that Dr Lynch personally created the pretence of a Beta program when none existed; and (b) Dr Lynch, by email on 23 October 2009, had directed a response to be given to Mr Goodman of Fidelity who had registered to download the Beta version of SPE, which was to tell him it was closed due to overdemand. Lastly in relation to this point, they further relied on a prepared answer in a Q&A script prepared on 20 April 2010 for Autonomy’s Q1 2010 earnings call stating the product had been sold so Beta was not relevant, which the Claimants submitted was a way of ducking the question because Dr Lynch knew the Beta program had not existed.

(4) Contrary to what Dr Lynch told the market, and though it was stated to account for some $4 million of the sales and marketing costs, there was no Quick Start initiative or programme in existence. The evidence of Dr Blanchflower and Mr Lucini was to the effect that they were unaware of the existence of any such program in circumstances where they would be reasonably expected to if such a program existed. There was no documentation to substantiate systems engineers (“SEs”) coordinating with customers in order to install SPE such as when to turn up, with what equipment, and identifying how SPE may be useful to such a customer, and a lack of witness evidence to the effect that SPE was installed at customer sites in Q3 2009.

Apart from a single seminar held at Cliveden House from 09 September 2009 to 10 September 2009, there were no such selective seminars in Q3 2009 which involved flying key customers from around the world. There was none of the supporting documentation which might be expected: no guest lists, invitations, flight bookings, hotel bookings, schedule of events, nor any agenda for the seminars. The Claimants suggested that if such seminars had really occurred then they would have involved Mr Lucini; and he denied any involvement. Dr Lynch had not called a single witness who was claimed to have been a customer, though this had been a pleaded issue. Further, the Claimants relied on, as being obviously a concoction, Dr Lynch’s suggested response to an email dated 14 September 2009 sent to Mr Hussain by an analyst called Mr Raimo Lenschow asking for an update on progress on the product:

“Premarketing well underway, ad starting to run, customer seminars with them being flown in from around the world already done and on going, demo appliances out, industry analyst briefed etc……..sadly we missed off financial analysts as they have no budget…!”

The Claimants added to this barrage asserting, in effect, that the Q3 2009 Earnings Call had amounted to a pack of lies, that a memorandum prepared by Ms Gustafsson (of Autonomy’s finance department) (“the R&D Memo”), which had been produced to justify to Deloitte the R&D figures in the accounts, was based on figures which had no factual foundation and were untrue. The figures had been provided to Ms Gustafsson by the cabal, and in particular, Dr Menell and Mr Chamberlain. Ms Gustafsson made clear that she had very largely relied on the information that had been thus supplied to her.

There were two primary elements in the calculation of R&D capitalisation in the R&D memo:

(1) $4.8 million was attributed to time spent by SEs and

(2) $2.5 million was attributed to time spent by the development engineers.

The Claimants detailed at considerable length their reasons for dismissing the amount of time attributed to SEs, which involved an extended account of the usual work they undertook and the calculation of the monetary value to be attributed to their time. For present purposes, the following should suffice:

(1) The SEs were uncontroversially described in the R&D Memo as follows:

“ Systems Engineers are a team of individuals traditionally associated with ensuring that software supplied by Autonomy meets the needs of the customers[’] existing systems …

Typically the work of an SE is the tailoring of Autonomy software, or providing ‘proof of concept’ to customers to demonstrate that the software meets their requirements”.

(2) The basic approach was stated in the R&D memo as follows:

“From discussions with Pete Menell, a list was provided showing SEs who spent all of their technical time in Q3 09 developing SPE through trialling the software with real databases. 75% of Q3 SE costs have therefore been capitalised in line with the proportion measured and agreed with HMRC . ”^{^[225]}

(3) Dr Menell simply provided a list, later expanded by Mr Chamberlain without apparent justification, of all Autonomy SEs, and the figure in the R&D Memo represented 100% of all their ‘technical time’ (being 75% of total time).

The Claimants submitted that it was plain and should be found that (i) none of the SEs in Mr Chamberlain’s list spent all of their technical time in the first three quarters of 2009 developing SPE by trialling the software with real customer databases; on the contrary, there is no evidence that any of these SEs spent any of their time in the first three quarters of 2009 doing that, with one minor exception of one SE, who spent some time assisting Mr Lucini with the construction of the SPE demo in Q3 2009. They submitted further that that conclusion was further justified by the complete absence of any of the evidence that might be expected if the SEs had been trialling SPE with clients, such as feedback and customer comments.

The Claimants also submitted a shorter answer to the analysis in the R&D Memo based on Dr Blanchflower’s evidence to the effect that SPE did not come into existence until midway through Q3 2009 itself. They pointed out that it was a logical impossibility for the SEs, at any earlier stage during Q3 2009, let alone in Q1 or Q2 2009, to have been trialling with customers a product that had not yet come into existence.

The amount ($2.5 million) allocated to development engineers (see paragraph 1534 above) for work described in the R&D Memo as having been a “direct result of the new product, SPE” was calculated from spreadsheets showing the time spent on R&D by each of the development teams within Autonomy. Each development engineer’s salary as recorded for national insurance was extrapolated and then multiplied by the days recorded as having been spent by that engineer on R&D work. Ms Gustafsson told me in cross-examination that the attribution of $2.5 million to SPE, out of a total of $6.94 million capitalised for R&D work for development enquiries, was based on a “sort of qualitative context around the development of SPE that you can’t do in a spreadsheet.”

On that somewhat opaque basis, the R&D Memo asserted, for example, that it was likely that “over half” of the $62,589 of R&D time incurred by the Cambridge team was attributable to SPE, and a round figure of $400,000 was taken as the value of the time spent by that team on SPE. Ms Gustafsson told me, when asked what was the source of the assertion of the proportion of R&D time spent on SPE by the Cambridge team, that she could not remember specifically but “I would speculate that it would have been Pete”. She confirmed that she had no recollection of having spoken to Dr Blanchflower, who was based in Cambridge and head of R&D, about how much time he and his team had actually spent. I accept Dr Blanchflower’s evidence and find that she had not done so. His evidence was that the value of time spent by the Cambridge team was a maximum of £24,000.

The Claimants made the further point that on Dr Lynch’s own case SPE predominantly involved reconfiguration of IDOL. The R&D Memo asserted that a number of other Autonomy development teams - Discovery Mining, iManage and Web Content Management (“WCM”) - in total incurred $2 million worth of time in the development of SPE. But IDOL was the responsibility of the Cambridge development team led by Dr Blanchflower . His evidence was that other teams did not have access to the IDOL source code, which was only available to the Cambridge development team, and they played no part in the development of IDOL. Accordingly, it is inherently unlikely that they would have played any part in the development of SPE. When it was put to Dr Blanchflower that he was not in a position to know what work was being done by the WCM team in 2009 he answered firmly but fairly:

“Was I aware of everything that the WCM team were doing? No. Were they working towards SPE? No. But I didn’t know all of the details of what they were doing, correct.”

The Defendants’ response to this barrage sought to re-address the origin, substance and worth of SPE:

(1) SPE was a real product which predated the hardware sales: it was “not some kind of a blind for the hardware costs” .

(2) Long before the Loudham Hall meeting, Dr Lynch had stated at the January 2009 earnings call that:

“Now onto the really interesting one. Our probabilistic structured technology is new technology which we’ve been working on. We think that’s going to be very powerful technology and that’s going to lead to a large market. And so that’s something that we will see revenue start to come in, in 2009.”

(3) Contrary to the impression given in some of his evidence at trial, Dr Blanchflower had regarded this as a potentially huge new product line with “the potential to open an entire new market”. That view was shared within Autonomy.

With particular reference to the issue of R&D capitalisation:

(1) Much had been made by the Claimants about the capitalisation of R&D costs in respect of SPE but this was a “peripheral area of the case”, with no relevance to any part of the Claimants’ loss: the Claimants’ criticism of lack of evidence and witnesses to speak to the matters was misplaced.

(2) The Claimants’ approach to the capitalisation of SEs’ time was flawed. In particular, the SEs were in many ways and over a long period “contributing to the development effort” with regard to structured functionality.

(3) Not only Ms Gustafsson and the finance department but also Deloitte were evidently satisfied as to the amount of SE time being capitalised, after careful testing.

(4) None of the Claimants’ witnesses could speak to what the true costs which should be capitalised were. Dr Blanchflower “did not have the expertise to comment on which costs should or should not be capitalised”. Mr Lucini could not comment on what was involved: “He had a customer-facing role and was not usually involved in the development of products” nor was he in any position “to accurately quantify the exact time spent by SE’s in relation to SPE.” Ms Gustafsson was the only person directly involved in the accounting.

The Defendants ^{^[226]} dismissed as unsupported by the evidence and unwarranted the suggestion that the statements about the derivation of marketing costs in the Q3 2009 Earnings Call were false.

Dr Lynch pointed out that Mr Welham had listened to the call; Mr Welham confirmed that if he had heard anything misleading on that call, he would have raised it with his engagement partner, and that he did not have occasion to do that. Dr Lynch also maintained that none of the witnesses on whom the Claimants had relied had been in a position to give reliable evidence on how much was spent on marketing costs. Dr Blanchflower was not part of the marketing department, and nor was Mr Lucini. Mr Lucini told Morgan Lewis in an interview in December 2012 that it would be difficult to quantify the marketing costs for SPE.^{^[227]}

Dr Lynch sought to downplay the role of Dr Blanchflower, describing him in written closing submissions as merely a “titular head of research and development…largely involved in IDOL development” to whom “not everyone in R&D reported” and who himself had “very little interaction with members of senior management other than Dr Menell.” He “did not have the expertise to comment on which costs should or should not be capitalised” ; he could only speak to work on coding and not as to work streams more generally: it was stated that “he did not know what the SEs were doing on structured data, or to what extent that could properly be included in the R&D capitalisation.” More generally, his evidence on this aspect of the matter should be discounted: he “had a propensity to make firm statements as to matters that he was not involved in and could not recall”.

Mr Lucini, similarly, was said not to have “handle[d] the area”. In cross-examination he admitted being aware of a long list of people in Autonomy involved in marketing, with at least some working on SPE,^{^[228]} but acknowledged that he did not know precisely what they were doing:

“ Q. You were not in the marketing department, so you would not have known in detail the breakdown of the marketing expenditure?

A. That is correct.

…

Q. Or what all the people in the marketing department were doing?

A. No.

Q. You didn't have day-to-day responsibility for managing the marketing of SPE?

A. I did not.”

Further, Dr Lynch contended that the documents in fact showed a number of activities being undertaken during the launch period, including (it was said) a “focused” advertising and briefing campaign for analysts and journalists and advertisements placed in trade publications. In addition, Autonomy spent considerable time and money briefing trade journalists and analysts about the product, which was said to have generated immediate interest amongst industry analysts and writers. Mr Lucini explained that there were also roadshows to what he described as “ quite a lot of people ”, specifically analysts, press and customers.

Dr Lynch had been advised there was a soft launch of the product through the beta programme. Despite being subjected to a lengthy cross-examination regarding the roll-out of this programme, Dr Lynch did not have a detailed understanding of its delivery, and made it clear that his understanding derived from documents he had seen subsequently. He told me he could only “ speculate ” about these things that went on “far below his level”.

He emphasised that in any event, at the time (i.e. October 2009), he believed SPE to be in the beta phase, and submitted that was a reasonable belief on the basis of the documents. His view was that a beta programme would typically involve two aspects: a process of gathering up interest and testing of the product. Such programmes were normally run through the Customer Liaison Support Program,^{^[229]} who were the experts in this area. He contended that the evidence showed this happening in the case of the SPE beta programme: participants registered for the beta programme via Autonomy’s website,^{^[230]} and the registrations were forwarded to telesales to take up with the customer.

Though Dr Lynch said that this likewise was a process that took place “ way below ” his level as CEO, he told me he believed that the Quick Start programme was also ongoing as of October 2009 and referred to documents such as a request to approve the purchase of servers for the “ SPE Early Starter program ” in October 2009. He relied especially on Dr Blanchflower’s acceptance that VMS participated in the Quick Start programme in early 2010. With his usual reverence for anything “secret squirrel” Dr Lynch told me that he believed that some of the first customers that received SPE through the beta/Quick Start programmes were likely to have been intelligence customers.

Dr Lynch also maintained that, contrary to the Claimants’ presentation, there was also expenditure on customer seminars, and customers had been flown to seminars, called “Customer Innovation Forums” or “CIFs”, in (so he believed) Cliveden, Napa Valley and Mexico. There was some documentary material, not mentioned by the Claimants, demonstrating both that CIFs were taking place in Cliveden (for UK/EU customers) and Napa Valley (for North American customers) in October 2009, and that Autonomy was responsible for customers’ air fares (though as I elaborate below, nothing in the document seemed to suggest that they were SPE-specific).

Dr Lynch’s written closing submissions also went into some detail about the marketing and use of SPE after its launch, instancing a number of case examples of it being presented (and more occasionally deployed by Autonomy in Interwoven products). Also instanced was an internal Autonomy email to (amongst others) Mr Lucini and Dr Blanchflower dated 3 February 2010 reporting that a demonstration of SPE to a VMS representative had gone very well: “he was pleased with the results and sees great potential in SPE…[and]…took away from this both a very positive impression of SPE as well as a great concrete example of how it could solve problems for his own needs…” .

However, Dr Lynch’s over-arching point was that he was not directly involved. He had understood that the finance department had determined properly R&D costs by reference to the proportion of SE and Development Engineers’ time costs at the rate agreed by HMRC but he was not involved in the R&D capitalisation, did not know any detail in relation to it, gave no instructions in relation to it, and was not involved in the production of the R&D Memo or speaking to Deloitte about it.

Similarly, he maintained that he depended on what he was told as to marketing costs. He said he was not involved in the beta or Quick Start programmes: he had been informed of the plan for the soft launch of the product via a document and was entitled to rely on its accuracy. He acknowledged having approved an item of spend for SPE, but that was all: as CEO, he was not involved in approving everything spent.

Likewise, he asserted that he was entitled to rely on the documents given to him as the basis for his statements on the Q3 2009 Earnings call, including the figures he gave.

Summary

After this rather lengthy digression into the detailed evidential dispute about SPE, I turn to my assessment as to whether, as the Claimants submitted in their written closing, (a) both Defendants “knew that the statements in the Q3 2009 Quarterly Report and on the Q3 2009 earnings call about Autonomy’s R&D capitalisation being primarily attributable to SPE were false” and (b) in the Q3 2009 Earnings Call, knowing already from (and as set out in) the Strategic Deals Memorandum that Autonomy was intending to treat $19.5 million of hardware costs as sales and marketing expenses,

“Dr Lynch lied about the sales and marketing and COGS expenditure on SPE in Q3 2009…to enable Autonomy to cover up the very substantial costs it had incurred in the quarter on buying pure hardware and which it was treating as a sales and marketing expense.”

It is reasonably clear, in my view, that the R&D Memo was to a considerable extent a concoction. The figures given for the time cost of SEs was greatly exaggerated (well beyond any acceptable margin). It was based on (a) a list of SEs initially provided by Dr Menell but then enlarged by Mr Chamberlain who (I accept and find) cannot have had any sufficient reason to suppose had done any work on SPE; (b) treating all those on this artificially enlarged list (as I find it to have been) as having spent 100% of their technical time on SPE, which was entirely unrealistic.

My impression that the amount attributed to SE time costs, which was the most important element of the capitalisation costs, was contrived is fortified by: (a) the lack of any contemporaneous documentation showing any work done by the SEs in relation to SPE; (b) the lack of any witness evidence from SEs (including none from Mr Zanchini, all of whose technical time in Q1, Q2 and Q3 2009 was treated as having been spent on SPE, who works for Darktrace and must have been amenable to a witness summons); (c) Mr Chamberlain’s response to an email from Deloitte sent also to Mr Hussain and dated 14 October 2009, in which he blamed Deloitte for not having raised the issue of timesheets for SEs until too late, assured them that “the 75% is not a number picked out of the air” (which deflected from the real issue as to what, if any, time that was properly to be applied to) and under the guise of dealing with the issue “in a practical and sensible way” studiously ignored the actual request which was for an opportunity “to talk to 5 systems engineers to understand their role in the development”.

Similarly, in my view the time cost attributed in the R&D Memo to development engineers was simply the say-so of Dr Menell and Mr Chamberlain with no feet in the facts. Had there been any intention to base the figures on fact, it seems to me inconceivable that Dr Blanchflower would have been entirely ignored in the process.

The attribution of a capitalised sum of $2 million to the work of other Autonomy development teams in the development of SPE lacked any evidential support; and I agree that it is inherently unlikely that any of those teams played any such role in the development of SPE, given that they had no access to the IDOL source code.

My view that the R&D memo was concocted disposes also of the notion that the Defendants (and more especially Dr Lynch) were entitled to rely on Deloitte’s approval of the R&D capitalisation figures. It was Mr Welham’s unchallenged evidence, and it seems to me to be obvious (and I find), that Deloitte would not have accepted the $7.3 million capitalisation if they had known that the R&D memo had no real footing in fact. I accept and find also that the Audit Committee, who were dependent on what Deloitte reported to them, were thereby also misled.

Focusing next on Dr Lynch’s justification of what he said at the Q3 2009 Earnings Call about marketing costs and launch spend in the sum of “I think nearer $18/20 million, with about $4 million extra on the Quick Start program” :

(1) Dr Lynch’s suggestion, not pleaded and unstated in his witness statement and skeleton argument, was that his reference needed to be read in the context of what he said at the Q2 2009 earnings call and in a website entry: it was, he said, a reference to Autonomy’s aggregate marketing and launch spend for a couple of new products and was “more than just SPE”. I do to think it is necessary for me not to say more than that (a) having reviewed what he said in Q2 2009 there was nothing in it to support this construction, and indeed it appeared to reinforce that he was talking about SPE only; (b) the website similarly did not assist him. That was further reinforced by evidence of the drafts provided by Dr Lynch to direct Mr Hussain’s responses in email exchanges with an analyst at Citibank called Mr Hoi Chuen Lam. In short, Dr Lynch’s suggestion of some different meaning to what he said at the Q3 2009 Earnings Call in this regard was a false expedient for which the only apparent explanation is that he knew he had given inflated figures: and I find he did know that accordingly.

(2) As to the five categories of SPE expenditure he referred to:

i. The Claimants always accepted that some $78,342 had been spent on trade advertising.

ii. There was no evidence of material expenditure in relation to analysts.

iii. I would accept that, contrary to the Claimants’ case, there may have been some events or meetings which might be said to have been part of a “soft launch” but very little (if any) evidence of an established beta programme.

iv. Similarly, there was little or no support for the suggestion that material sums, let alone $4 million, were referable to the Quick Start initiative with strategic customers. As the Claimants observed, even Dr Lynch himself did not positively assert in his witness statement that the Quick Start initiative had taken place, and could not name a single site at which it was said to have taken place apart from vague reference to there being one “within the UK intelligence community”. None of his witnesses (including Ms Pereira, who did address SPE) provided any support either. No supporting evidence was provided. Dr Lynch relied on assertions by his Investor Relations team and in the draft “Q&A script” for which no more foundation was provided either.

v. The evidence that Autonomy had arranged and flown customers in for trade seminars seemed to me to illustrate how little support Dr Lynch had for his statements. None of it appeared to be SPE-specific; the seminars were concerned with Autonomy generally. It may be that SPE was to be mentioned at the seminar as one of the “latest product developments” : but the only material to which I was referred seemed focused principally on Autonomy’s recent acquisition of Interwoven and “Autonomy’s iManage brand”.

(3) Whilst the figure of only $100,000 suggested by the Claimants as all that had been demonstrated to be referable to marketing SPE may be too low, I am satisfied that what Dr Lynch said at the Q3 2009 Earnings Call about marketing costs and the Quick Start initiative had no foundation.

The three questions which seem to me to remain are (a) what involvement Mr Hussain had; (b) Dr Lynch’s overarching point that he himself relied on others to verify what he was to say, and he believed what he said to have been true; and (c) subject to that, whether the falsities were deliberately aimed at disguising the hardware sales and costs.

As to (a) above, Mr Hussain was the CFO, in charge of the finance department. The question of capitalisation of R&D costs was a matter within his responsibilities. There are clear signs that he conceived the use of R&D capitalisation to disguise hardware costs. I have taken the following from the Claimants’ analysis, which I accept and adopt:

(1) On 1 October 2009, Mr Hussain emailed Mr Chamberlain and Dr Menell suggesting that the R&D for Q3 2009 should be $9.5 million, and asserting (without any possible basis) that this was all “ to do with the enormous work pete and the team have been doing on the structured probabilistic engine ”.

(2) Mr Hussain was the sole recipient of Mr Chamberlain’s email of 7 October 2009 identifying a $2 million shortfall in R&D.

(3) He was copied on the email exchanges on 8 October 2009 regarding the identification of SEs to include in the capitalisation, and was kept informed by Mr Chamberlain of the fact that Ms Gustafsson would be meeting Dr Menell to revise the R&D Memo.

(4) Mr Hussain was also copied on Mr Chamberlain’s exchanges with Deloitte on 14 October 2009.

(5) In the context of justifying the R&D capitalisation, Mr Hussain also provided to Mr Chamberlain, for relaying to Deloitte, a forecast of the revenues that SPE was expected to bring in.

(6) Deloitte’s working paper dated 15 October 2009 duly recorded these figures as Mr Hussain’s estimate of the revenues that SPE would bring in.

To my mind, this has all the hallmarks of the working of the cabal. There is no reason to suppose that Mr Chamberlain (Mr Hussain’s trusted deputy) or Dr Menell would have constructed a dishonest fraudulent analysis to support the R&D capitalisation, without Mr Hussain’s knowledge and approval. In my judgment, Mr Hussain knew that the R&D capitalisation was based on concocted figures. It seems more likely than not that he would also have known of the false ‘marketing cost’ figures.

The Claimants speculated that this perhaps may explain why, on 18 October 2009, shortly before the Q3 2009 earnings call, Mr Hussain wrote to Dr Lynch saying:

“ I am burnt out...given my “anal” nature i am spending all my time (awake and asleep) worrying about the 10 minutes on Tuesday where i have to answer the analyst questions and i’m not doing anything else, i need you to take the questions this time round unless they are really easy, i don’t want to deal with the analysts anymore. ”

As to (b) in paragraph 1563 and Dr Lynch’s overarching point, far from being “ best placed ” to know as the Claimants asserted he was, Dr Lynch’s case was that:

(1) He was not involved in the R&D capitalisation, and there is no evidence at all that he knew any detail of it or, as the Claimants suggested, gave any instructions in relation to it.

(2) He was not involved in the R&D Memo and there is no suggestion he spoke to Deloitte as part of the process.

(3) Nor was he involved in the detail of the sales and marketing work. He was not involved in either of the beta or Quick Start programmes, both of which went on far below his level as CEO. He was told there was to be a soft launch of the SPE product in September, and that customer seminars were ongoing. It was not put to him that anything in that document was inaccurate, or intended to mislead; on the contrary, it appeared to be accepted by the Claimants that this document accurately represented Autonomy’s plan for the launch of SPE. Dr Lynch was, accordingly, entitled to rely on it.

(4) The statements made by Dr Lynch on the Q3 2009 earnings call were based on information he had been given and he had no reason to doubt its veracity. As Dr Lynch explained, he was given drafts of the presentations in order to prepare for the calls. The documents, which show the Investor Relations team sending Dr Lynch the PowerPoint presentation and Q&A, stated that the beta and Quick Start programmes were underway, that the feedback had been positive, that $20m had been spent on the new product launch, and that c. $3-4m had been incurred due to overdemand on the Quick Start programme. This information came from others in the organisation. If the information was inaccurate, Dr Lynch was unaware of that.

(5) The Claimants did not take issue with Dr Lynch’s evidence that the figures conveyed to the market had been given to him and prepared by others: this was stated at §228 of his witness statement and was not challenged even when repeated in cross-examination.

(6) In fact, very few documents involving Dr Lynch were put to him at all on this area of the case. This underlined how little involvement Dr Lynch had in any of the SPE development or launch process. For example, the Claimants could only point to one instance of Dr Lynch approving a marketing expense in relation to SPE. That email dealt only with one aspect of the marketing campaign, namely advertisements in printed trade magazines. The Claimants had produced no evidence that Dr Lynch was personally involved in approving any other spend in relation to SPE (in relation to its development or marketing).

I accept that the documents do not establish his personal involvement. However, in my judgment, Dr Lynch would have known (and for the avoidance of doubt I find that he did know) that:

(1) There was no basis for the figures for marketing that he gave in the Q3 2009 Earnings call. No document has ever been identified in which anyone in Autonomy’s finance department advised Dr Lynch that anything approaching $18 or $20 million had been spent on sales and marketing in relation to SPE in Q3 2009. Ms Gustafsson confirmed that she was not involved in collating or calculating any such figures and had no specific recollection that anyone else had done so either. Moreover, it is wholly implausible that Dr Lynch, whose written approval had to be sought for advertising spend of $144,830 (which he declined to give and reduced to $80,000) would, unblinking and without detailed backing, have accepted expenditure of $24 million.

(2) The R&D capitalisation figures had been produced by the cabal and were similarly false. I cannot accept that all three would have kept from him that the figures for SE and Development Engineers time costs and the capitalisation amounts were a concoction. In my view, it is more than likely that he was kept informed, and I find that he was so.

The last issue to consider in the context of the Q3 2009 Earnings Call is whether the purpose of the falsehoods was indeed to disguise the hardware sales and costs (see (c) in paragraph 1563 above). The principal argument Dr Lynch advanced against this (in addition, obviously, to his case that he had no knowledge of any falsehood, which I have not accepted) was that “the Claimants’ arguments would have involved time travel” because (as briefly mentioned above):

“Autonomy had announced its development work on SPE by early 2009, before the strategic hardware sales had even commenced. SPE was plainly a real initiative, which predated the hardware sales by a considerable period. Further, Autonomy did not know until after the end of Q3 2009 that Deloitte would permit the allocation of any hardware costs to sales & marketing; it makes no sense to imagine that the development of SPE, and its launch in September 2009, was conceived as a ruse to hide the fact that Autonomy ’s sales and marketing expenses might include costs relating to hardware.”

Of course, the question is not whether SPE and its development was a ruse. The question is whether what was said falsely about it was in order to justify the R&D capitalisation figures and the marketing and Quick Start costs which were presented.

The answers provided by the Claimants, which I have done little to modify in my recitation of them below, seemed to me to expose the “time travel” argument as a bad point. The timing seems to me more to support the Claimants’ case than undermine it. Thus:

(1) In fact, the figure of $10-15 million for sales and marketing expenses was announced, not in Q1 2009, but on the earnings call for Q2 2009.

(2) That earnings call was held on 16 July 2009, not “ back at the beginning of 2009 ”.

(3) By 16 July 2009, the meeting at Loudham Hall (on 8-10 July 2009), at which Dr Lynch accepts that a programme of reselling hardware was decided upon, had already taken place.

(4) Accordingly, the suggestion that Dr Lynch announced a figure for the sales and marketing expenses at a time when he could not have foreseen the hardware sales is simply wrong. The hardware reselling programme had already been decided upon.

(5) It is true that Dr Lynch would not necessarily have known, as at 16 July 2009, precisely how much of the hardware costs would end up being allocated to sales and marketing expenses. But that is a point against Dr Lynch, not in his favour, because the figure of $10-15 million that Dr Lynch announced on the Q2 2009 earnings call ended up being out by a significant margin. As explained above, on the Q3 2009 earnings call, Dr Lynch identified higher sales and marketing figures of $18-20 million and an additional $4 million on the Quick Start programme.

(6) The reason for the increase was that by the time of the Q3 2009 earnings call, the actual hardware costs were now known to be $47.34 million of which $28.37 million was allocated to sales and marketing.

(7) Accordingly, the Claimants’ case does not depend on the proposition that Dr Lynch had perfect foresight of the quantum of future hardware costs, let alone that he had the power of time travel.

I have concluded that both Defendants knew that the Q3 2009 Earnings Call so far as it related to SPE contained a series of falsehoods with the objective of concealing the hardware sales and costs.

Q1 2010 Earnings Call

The Q1 2010 Earnings Call took place on 21 April 2010 during the Icelandic volcano eruption which spewed ash into the atmosphere and into flight paths and caused flights to be cancelled throughout the Northern Hemisphere. Dr Lynch ’s evidence (which I accept) was that he participated on a mobile phone from California, having been unable to return to the UK.^{^[231]}

As it is clear from the Q&A scripts ^{^[232]} prepared in advance to provide possible answers to questions that might be asked at the Q1 2010 Earnings Call, the Defendants anticipated that they would be asked about the large increase in Autonomy ’s inventory to $10 million at the end of Q1 2010; and so it transpired.

As they had planned, the Defendants answered these questions by saying that the inventory comprised $10 million of hardware purchased at the end of Q1 2010 to take “advantage of discounted offers” in anticipation of an increase in appliance sales in Q2 2010 and beyond. Mr Hussain was careful to explain that the margin for such sales of Autonomy ’s appliance product (Arcpliance) would still be “dominated by the software element”. Mr Daud Khan of JP Morgan took from this, reasonably in my view, that “the value of the appliance product that Autonomy was selling was in the software, and that there would be no particular impact on gross margin when the application was sold.” That extrapolation from what Mr Hussain said was not challenged or contradicted.

The increase in Inventory and the explanation given for it generated a number of questions from analysts on the call:

(1) Michael Briest at UBS asked the following:

“ Could you talk a bit more about the Arcpliance product because clearly the level of hardware purchase, and I guess maybe the level of sales there, would be more than we’re used to?”

(2) In response, Dr Lynch explained that Arcpliance was “a hardware box which has all the software loaded on it” . He went on to state that the “main driver for a sudden pick-up in Arcpliance purchases we think is early case assessment” and continued:

“It must be said that this is just one of the forms in which customers can consume our technology. They can of course buy ECA, add software, set up their own servers and put the software on it, or indeed it can be done as a hosted offering as well. So really our approach is just to make the technology available in whatever form that customers want. We don ’t know whether this sudden burst of Arcpliance-type sales is a one-off because of a response to things the regulator is doing at the moment or whether it is part of a trend. I suspect it ’s probably more likely to be a one-off, but we ’ll see how that unfolds.”

(3) Mr Briest asked a supplemental question: “Could you say what the Arcpliance or appliance sales and hardware sales would have been in a normal quarter? It seems a large number for Q1, going into Q2” .

(4) In response, Dr Lynch said: “Yes. It would be a fraction of that kind of number”.

Dr Lynch was cross-examined especially closely on his answers. His flat and apparently unflustered delivery, sometimes so modulated as to become monotonous, could not disguise the fact that he was at a loss to explain his own answer. In his discomfort, he relied on external circumstances: his absence in California, the middle of the night, the poor line quality, his bad hearing, the possibility (which, as he went on, he promoted to a fact) that the transcript (one of two prepared each by different transcribers) was inaccurate ( “all rather questionable” ): but none was convincing, still less an answer. He was left asserting, without conviction, that in any event he was not sure, but he did not think, his statement on the call was misleading, and that in any case, “I don’t think the question as written makes a lot of sense to me in the context of the call.”

The impression I formed that he was in this context left grasping at straws was consolidated by the further researches of the Claimants demonstrating that:

(1) The two transcripts, one by Bloomberg, the other by Thomson Reuters, were almost identical.

(2) Dr Lynch’s suggestion that in that case both were wrong was not supported by the audio version which the Claimants uploaded onto Magnum (and which I have listened to at their request): the transcripts seem to me accurately to record Mr Briest’s question and Dr Lynch’s response.

Further, although it is a subjective assessment, I do not consider that the audio version supports the suggestion that the line was very bad. There was some distortion, and some echo and hiss from time to time: but the quality was fairly good, and even making allowance for Dr Lynch’s hearing difficulties, I could not detect in the way or tone in which he responded any sign of uncertainty as to what the questions were. In particular, Mr Briest’s question was clear in the version I listened to, as was Dr Lynch’s response.

Later during the call, Mr Khan asked a follow-up question to that asked by Mr Briest:

“ Just a follow-up question to Michael’s question on the Arcpliance. If I understand it right, the $10m that has gone into inventory, I’m just wondering what the revenue would be for that amount of inventory and when we are likely to see that. Are we going to see that in Q2 or is that going to be spread across Q2 and Q3?”

Dr Lynch’s response (which appears in almost identical terms in the Bloomberg transcript which demonstrates that both the transcripts were accurately transcribed) was as follows:

“So the first question was on the Arcpliance. I’m afraid we are not going to give you an exact number, because that’s rather commercially sensitive. What I would say is that the software component of the revenue is far higher than the hardware component. So the software is still the bit that dominates in terms of the cost of an Arcpliance. It’s not the hardware; it’s the software. Rather like on the Cloud side you may remember it’s rather different from, for example, a normal hosting situation.

In terms of the sale, most of that inventory has already been sold in Q2. There might be a little bit that goes over the end, but at the moment it looks like it may well all happen in Q2.”

Dr Lynch ’s answer was predicated on the fact that this hardware was to be used in connection with Arcpliance, in which the software element would dominate, and that most of the inventory had already been sold by that point in Q2 2010. The Claimants contended that the predicate and every part of his answer was false: the inventory related to pure hardware sales, the sales had occurred and the goods had been delivered in Q1 2010, but Autonomy had decided not to recognise the revenue in that quarter.

Dr Lynch asserted, but provided no evidence to substantiate, his predicate that the inventory related to ‘ Arcpliance’ or at least to a package sale of hardware and software from which the revenue related in substantially larger part to the software component. Even accepting for that purpose Dr Lynch ’s evidence that when referring to ‘ Arcpliance ’, he meant any “regulatory appliance” and that the term “ appliance” could include not only hardware pre-packaged with software but also hardware supplied to be used with Autonomy/IDOL software to be supplied later (see his first witness statement), an analysis of the Q1 2010 transactions which had given rise to revenue which Autonomy had deferred on the ground that, though shipped, delivery of the goods had not completed (as explained by Mr Chamberlain in an email to Mr Stephan (copied to Mr Welham dated 20 April 2010)) (see paragraphs 991 above) does not appear to support the predicate.

Dr Lynch was asked by another analyst, Mr David Toms at Numis, to “just clarify the $10m or so of hardware revenue” . His answer, which was almost identically recorded in the two transcripts and for which the transcripts seemed to me consistent with the audio file, was as follows :

“David, I think you may have misunderstood what that revenue is. It’s not hardware revenue. What it is is the selling of an appliance. … We have very little interest in just selling hardware, and consequently the revenue that that goes for is not related to the hardware cost. It’s solely a component of that sale. So what we are not doing here is acting as a generic company that resells hardware, like a Morse or something like that. Obviously those people do that business and we have no interest in it.”

The Claimants rejected this answer as a lie. They submitted that Dr Lynch had no credible explanation when Mr Rabinowitz put to him in cross-examination that the hardware in question was not related to an appliance; and that acting as a hardware reseller was exactly what Autonomy was doing.

However, although Dr Lynch may have been a little uneasy with his analogy with Morse, he seemed to me to be much more comfortable and consistent, and thus more convincing, in essentially repeating to me Autonomy’s justification for the hardware reselling strategy, to the effect that the programme was not an end in itself, or a separate business; that its purpose was to protect and promote Autonomy’s core software business; and that Autonomy was not acting as a hardware reseller because it had no interest in selling hardware except to promote its software business and what it did was all ancillary to that business, as demonstrated by the fact that it was doing so at a loss.

The difficulty for Dr Lynch, as it seems to me, is not in the distinction; it is in the fact that it was not an explanation he uttered or even hinted at in the Q1 2010 Earnings Call (or in any of these Earnings Calls under review). Ultimately, what was not said at any of the Earnings Calls was the most important point of all.

Unsurprisingly, therefore, Analyst reports published following the call stated that Autonomy’s hardware inventory related to appliance sales. Thus, a KBC Peel Hunt report written by Mr Morland referred to what in truth was a balancing item for deferred revenue as the “ $10m Arcpliance stock ”.

Q2 2010 Earnings Call

The Q2 2010 Earnings Call took place on 22 July 2010. The only part of the call which the Claimants relied on in the context of their hardware claims related once more to the $10 million inventory entry that had been discussed on the previous call.

As usual, Q&A scripts had been prepared. As in the case of the Q&A scripts for Q1 2010 I address the content of the scripts earlier. In summary, the scripts show that it was anticipated that questions would focus on what was the hardware content in Arcpliance, and what would be the effect of sales of the Inventory on organic growth. Again, that was the focus of some of the questions in fact asked at the Earnings Call. An analyst, Mr Rajeev Bahl of Piper Jaffray, who prefaced his explanation saying that he “wanted to dig into the gross margin question a little” , asked whether this “ spike in appliance sales ” had affected gross margin.

In answer, Dr Lynch described the impact of appliance sales on gross margin as an “ exceptional ” situation, with “ the appliance model [accounting] for about 1% ”. He remained altogether silent about the hardware reselling strategy: he said nothing about the loss-making hardware sales or their effect on gross margin. Dr Lynch maintained that he was entitled to say nothing, since the question did not ask about hardware sales generally: it was confined to the effect of appliance sales.

The problem with that line, in my view, is that Dr Lynch never admitted to any hardware sales except “appliance” and Arcpliance sales. Indeed, although he initially confined “appliance sales” to sales of hardware “in order for Autonomy software to be loaded on it” , by now Dr Lynch increasingly deployed the term “appliance” and Arcpliance sales to cover and disguise hardware sales. In such circumstances, I do not consider Dr Lynch’s response to be complete and truthful.

Conclusion on Hardware Sales Issue 1

In my judgment, this exploration of the contemporaneous documentation plainly supports the Claimants’ case that, whatever its purpose at inception, the hardware reselling strategy soon and, increasingly, became an addictive means of meeting market expectations of revenue maintenance and growth. What the Defendants asserted to be the purpose of the programme either initially was, or as soon as the size of the potential revenue source was apparent became, a pretext to justify accounting for the hardware revenue as part of the software business revenues without separate identification. Neither the fact that the pretext appeared rational nor that Mr Sullivan and Mr Egan accepted it, alters that conclusion: the justifications and pretexts offered by the Defendants in their relentless search for revenue were invariably rational and believable but, in my judgment, they were pretexts nonetheless.

As I mentioned earlier, although the ‘purpose’ case and the ‘disclosure’ case have for the purpose of analysis, been separately examined, they are inseparable.

The real objection to the hardware sales was not that they were to raise revenue and were loss-making; it is the fashioning of the pretence, principally for the consumption of Deloitte and the Audit Committee, that their purpose was ‘marketing’ or to bring more substance to that elastic term, to promote the software business by deepening relationships with hardware suppliers and leveraging discounted hardware sales to generate more high margin software sales. Their true driving purpose was never, and could never be disclosed: because that true purpose was to generate revenue in order covertly to make good shortfalls in software business revenues by including hardware revenue without differentiation or disclosure in Autonomy’s accounts and reports as if it had been generated by Autonomy’s ‘pure software’ business.

Purpose, pretence and non-disclosure combine and support each other to expose and prove the impropriety. Had the true purpose of the hardware sales been ‘marketing’ (as that word is above elaborated) the Defendants would have wished to publicise it energetically. Had the pretence not been maintained, Deloitte and/or the Audit Committee would not have approved the accounting treatment. Both concealment of the sales, and in particular their cost in Autonomy’s accounts and other published information, were necessary because revelation of Autonomy’s use of hardware sales, and the erosion of gross margin would have nullified their true purpose and exposed that Autonomy’s software business was not generating the accelerating revenue and profits which the market looked for and which sustained its share price.

In my judgment:

(1) The hardware reselling programme was conceived, expanded and implemented in order to enable Autonomy to cover shortfalls in software revenue by selling hardware and including the revenue without differentiation in revenue shown in the accounts as generated by Autonomy’s software business.

(2) To succeed the hardware reselling had to be concealed from the market, but sufficiently revealed to Autonomy’s auditors and audit committee to secure their apparently fully informed approval of the company’s accounts.

(3) The imperative that the reselling should be concealed from the market required a number/variety of accounting devices which had to be presented in such a way as to secure the approval of Deloitte and the Audit Committee. In particular, their approval had to be secured to treat the costs of the hardware reselling programme not as COGS which would have eroded gross margin and encouraged both analyst and market inquiry and concern, but instead as ordinary sales and marketing expenses which had no such adverse effect on key investment parameters.

(4) The means by which this difficult balancing act was achieved are set out in my judgment. Suffice it to say that Deloitte and the Audit Committee were persuaded to regard the purpose of all hardware sales as being to generate revenue and new orders for the software business, and to account for hardware costs accordingly.

(5) The strategy also required that the contribution of hardware reselling revenues to overall revenues should be disguised or concealed, and that again the auditors and Audit Committee nevertheless being satisfied that such disclosure as was given was sufficient. That balancing act also was successfully achieved.

(6) The true purpose of the hardware reselling strategy/programme lacked any commercial justification and was dishonest. The true purpose had to be camouflaged in the way it was presented to Deloitte and the Audit Committee in order to obtain their approval for an accounting treatment which concealed it.

(7) The Defendants were well aware of this.

(8) Although I presently have doubts that this justifies the quantum of loss in the amount claimed in respect of it, since Autonomy was still a valuable company with an “almost magical” signature product, in terms of liability the Claimants’ hardware case has clearly been established.

Issue 2: Was Autonomy’s published information untrue or misleading by reason of the hardware sales, and did the Defendants appreciate this?

Notwithstanding the fact that the Claimants spent the larger (or certainly the longest) part of their written submissions on the ‘purpose’ case (and I have followed suit), the Claimants did state that irrespective of the Defendants’ purpose in relation to the hardware reselling strategy, Autonomy’s published information was untrue or misleading by reason of the hardware sales and since the Defendants appreciated this, that was enough to establish liability, even though if the purpose was as alleged by the Claimants the position was a fortiori. Indeed, in his oral submissions and especially in reply, Mr Rabinowitz described this as the primary way in which the Claimants put their claim.

My conclusion that the Claimants should succeed on their a fortiori case, if correct, renders Issue 2 superfluous. However, in case I am found to have been wrong, and in deference to the fact that, sometimes at least, the Claimants presented this as their primary case, I address the rather different foundations of this way of putting their case below.

Before doing so, however, I should make clear that, contrary to what I took Mr Rabinowitz to be suggesting (especially in his Reply), I do not regard this alternative way of putting the Claimants’ case (whether primary or not) to offer a simpler solution for them. As I have sought to explain, in my view, purpose and intention, pretence and concealment are all ultimately inter-related. It seems to me unrealistic in dealing with the third (concealment), which is what, in effect, the Claimants argued the statements in question amounted to or achieved, to be indifferent as to purpose. Further, it is unclear whether, in assessing whether the statements in question were misleading, the exercise to be undertaken in the ‘primary’ case is to be so on the basis that the purpose was as the Defendants asserted it to be. That is bound to affect the issues of intention and honesty. This may all also help to explain why I have taken the ‘purpose’ case first.

The Claimants’ alternative case

There are four main strands to the Claimants’ alternative case:

(1) whether Autonomy made statements in its published information that were untrue or misleading in light of the fact of the hardware sales and, if so, whether the Defendants knew those statements to be false or were reckless as to whether such statements were true or misleading;

(2) whether Autonomy’s published information omitted matters relating to the hardware reselling strategy which were “required” to be stated, and, if so, whether the Defendants knew the omission to be (quoting s. 90A (3) of FSMA) “dishonest concealment of a material fact”.

(3) whether the accounting treatment of the costs of the hardware reselling of itself caused the accounts and Autonomy’s published information to be untrue or misleading, and, if so, whether the Defendants appreciated this; and

(4) whether Autonomy was in any event required to make clear i n its published information what its accounting policy was with respect to hardware reselling and its costs, and if so, whether the Defendants appreciated this or knew the omission to be (quoting s. 90A (3) of FSMA) “dishonest concealment of a material fact”.

Statements made in Autonomy’s published information

The Claimants described this first strand, relating to positive statements made in Autonomy’s published information, as involving “the simple exercise of comparing what Autonomy chose to say to the market with the true facts…”. It did not depend on expert evidence: the experts agreed, and it might have gone without saying, that Autonomy was not free under any applicable accounting statements to make untrue or misleading statements. The arresting fact that provides the context for this strand of the Claimants’ case is that Autonomy’s published information did not, anywhere, make clear that it was selling material amounts of hardware otherwise than as part of a sale of Autonomy software (what the Claimants called “pure hardware” ). Even Autonomy’s corporate broker (UBS) was unaware of that.

The Claimants relied on the following statements by Autonomy as having been untrue or misleading:

(1) The description of itself as a “pure software company” , combined with references to “appliance” sales as comprising only a “small part” of Autonomy’s business;

(2) The description it chose to give in the narrative or ‘front end’ of each of its Quarterly Reports from Q3 2009 onwards and in its 2010 Annual Report of the categories comprising its revenue, particularly in “Supplemental Metrics” purporting to give a breakdown of its total reported revenues which made no mention of hardware sales and instead treated them as sales of “IDOL Product”;

(3) The description in its 2009 Annual Report in the “Revenue Recognition” section within the “Significant Accounting Policies” of the nature of the transactions entered into by the Autonomy group during 2009 as “the same as in 2008 in all respects” , which disguised the fact that whereas in 2008 Autonomy had not carried out any material pure hardware sales in 2009 it sold $53.6 million of it;

(4) Its inclusion within the metric “organic growth” (connoting, so the Claimants contended, growth attributable to sales of IDOL software) revenues from loss-making hardware sales, which were not sales of IDOL software at all.

The relevance of what the Defendants intended the representations to mean

It is important to have in mind at the outset the relevant legal tests to be applied in the context of determining whether (as the Claimants submitted) the positive statements they identified falsified Autonomy’s published information and give rise to liability on the part of the Defendants in respect of that published information under FSMA or in the tort of deceit.

The Claimants pleaded and tended to present this as a matter of determining what the words used would have conveyed to a “reasonable reader.” Thus, for example, their pleading as regards the statements, said by the Claimants to have been false and misleading and to have misled them, was (in relevant part) as follows:

“In describing itself as a “pure software” company, Autonomy was, in part, seeking to distinguish itself from companies which derived a significant portion of revenue from the provision of services. However, the description of Autonomy as a “pure software” company in Autonomy’s published information, the express reference in that published information to appliance sales, and the provision (as pleaded in paragraph 61.5 below) in the published information of a breakdown of revenue categories which together added up to total reported revenue for the period in question and yet made no reference to hardware, would, individually or cumulatively, have conveyed to the reasonable reader:

53A.1. that Autonomy was not engaging in any (or any material) sales of hardware apart from appliance sales or, at the very least, was not engaging in any (or any material) pure hardware sales; and/or

53A.2 that any revenue from hardware sales (apart from appliance sales) and/or from pure hardware sales was lower than the revenue from appliance sales (which were stated to be a small part of Autonomy’s business); and/or

53A.3 that any revenue from hardware sales (apart from appliance sales) and/or from pure hardware sales was lower than the reported revenue from Services.”

However, though what the “reasonable reader” would have understood is a necessary question to be answered, the answer is not sufficient where the gist of the claim is an allegation of dishonesty and deceit.

As the Defendants contended and, as a matter of law, ultimately the question is not what the objectively ascertained ‘best’ interpretation of the statements is, but what the Defendants understood and intended the statements in question to convey. As Mr Miles observed, this is not a contractual dispute in which the search is for the meaning most likely to accord with the intention of the parties as objectively derived from the words they used and the context in which they used them, set in the context of the factual matrix. This is a claim of dishonesty; and a man is not dishonest simply because he uses words which are misunderstood: he must be shown to have had an intention to mislead, or at the very least, to take advantage of a misunderstanding.

I have analysed the case law as to the proper approach in the Introduction to this judgment at paragraph 460 et seq above. Put shortly, it seems to me to be well established that:

(1) The claimant must prove what his/her own understanding of the statement in question was.

(2) Where the words used are entirely unambiguous, plainly misleading and admit no other understanding of them, and the claimant’s understanding accords with the only available meaning, it will be difficult, if not impossible, for the defendant to show that he/she intended some other meaning. The Claimants ultimately (in Mr Rabinowitz’s oral reply) submitted that to be the position in this case.

(3) Where, however, two or more interpretations are possible, the search is not for the most likely objective meaning of the statement in question, but for the intention behind its use.

(4) It is then for the defendant to make clear that he/she intended the statement to have some meaning other than that understood by the claimant; and in practice it will be incumbent on him/her to assert clearly which of two or more available meanings he/she intended, and the basis of that assertion. That is a factual matter.

(5) Theoretically, a defendant could have had in mind a meaning which the court considers not to be available on the words of the statement: and that could theoretically suffice. But in such a case, the difficulty of demonstrating such an extravagant or unlikely meaning may be insuperable as a matter of evidence, especially given that there would then be a question of whether the defendant was aware of the obvious meaning, and thus reckless as to how it would be likely to be understood, even if he/she did not intend that meaning.

The representation that Autonomy was a ‘pure software company’

As already explained in the context of the first issue, focused on the purpose of the hardware reselling and its concealment, the Claimants submitted that the statements in Autonomy’s Annual Reports that Autonomy was a “pure software company” , especially when combined with references to the fact that “appliance” sales represented but a “small part” of Autonomy’s business, were untrue or misleading in view of the nature and extent of the ‘pure’ hardware sales actually being transacted. The Claimants’ alternative case was that even if the purpose of the “pure hardware” sales was as asserted by the Defendants, the phrase was nevertheless intentionally misleading.

It may be helpful to reiterate why that was said to be so. According to the Claimants, the phrase would have conveyed to the reasonable reader that:

(1) Autonomy was not engaging in any (or any material) sales of hardware apart from appliance sales, or in the very least was not engaging in any (or any material) pure hardware sales, and/or

(2) any revenue from hardware sales and/or pure hardware sales (apart from appliances) was lower than the revenue from appliance sales, and/or

(3) any revenue from hardware sales (apart from appliance sales) was lower than revenue from Services.

The Claimants submitted that a “reasonable reader” of these statements could not have understood that, although Autonomy did not depend on any significant income from the provision of services, its business model did, by contrast, depend on material income from what they described as “other non-pure software activities”, such as conducting a significant hardware reselling business.

In that regard, the Claimants submitted that by carefully identifying very limited occasions or cases when it sold appliance hardware, Autonomy further strengthened the impression given that those were the only exceptions to its “pure software model”.

As to the comparison with the true facts, the Claimants contended, the total revenue generated by hardware sales over the Relevant Period amounted to approximately $200,000,000 and constituted approximately 11% of the total reported revenue from Q3 2009 to Q2 2011. These amounts dwarfed both appliance sales and the revenue from Services. Accordingly, the Claimants contended, each of the representations which they submitted were conveyed by the statement was false.

Against this, however, the Defendants denied that any such representations were conveyed by the statements made. They submitted that the Claimants’ case is based not on what was actually said, but on their interpretation of the phrase’s broader message, which the Defendants did not accept.

In any event, the Defendants objected that that was not the meaning the Defendants understood to be conveyed by the phrase, nor did they anticipate that it might be so construed.

The Defendants did not accept the Claimants’ contention that the statement was unambiguous. Mr Miles submitted that even if the statements in question could be read as connoting that Autonomy made no or no substantial hardware sales, the words certainly did not unambiguously convey that meaning. I agree with Mr Miles:

(1) In their pleading, the Claimants accepted that the description of Autonomy as a “pure software company” was intended to mean “in part” that it did not derive “a significant portion of revenue from the provision of services”. The possibility that the phrase could have meant only that appears to be countenanced even by the Claimants.

(2) The fact that the representations could be read and understood as simply distinguishing Autonomy from software companies which had a substantial servicing business is plain from the fact that they appear so to have been understood by Deloitte and the Audit Committee.

(3) Similarly, it seems to me an available meaning was that Autonomy carried on a one segment software business (in line with the segmental analysis) and no separate or self-standing hardware business.

(4) In other words, what the Defendants contended they meant was well within the range of possible meanings to be attributed to the phrase: it could mean that, or it could mean more.

I also agree with Mr Miles that in such circumstances, what the Claimants had to prove was both they and that, contrary to their evidence, either the Defendants understood and intended the words to denote that Autonomy made no substantial hardware sales, or they were aware that the words were ambiguous and intentionally used the ambiguity for the purposes of deception.

As to their own understanding, the Defendants maintained that what they thought the statement that Autonomy was a “pure software model” or a “pure software company” meant was that Autonomy did not have a substantial percentage of revenues deriving from professional services.

That, Dr Lynch told me, was worth stating because it was rare amongst software companies at the time because generally software sellers in the enterprise sector had to do a lot of customisation work on the product for every implementation. By contrast, he said, standard IDOL required little customisation and Autonomy’s business model was to contract out implementation work to be done by approved partners. This was an important point of difference marking out Autonomy’s business modelling: and there was good reason to bring it out in its published information.

The Defendants’ case was that they thought that this was clear from the context in which the phrase appeared, and from the Reports in which the statements appeared read as a whole.

Thus, the 2009 and 2010 Annual Reports stated that:

“ Autonomy is one of the very rare examples of a pure software model. Many software companies have a large percentage of revenues that stems from professional services, because they have to do a lot of customisation work on the product for every single implementation. In contrast, Autonomy ships a standard product that requires little tailoring, with the necessary implementation work carried out by approved partners such as IBM Global Services, Accenture and others. ”

Read in context, they submitted that the phrase says nothing about whether Autonomy sold hardware, or how much it sold.

The Defendants relied also on the explanation in its 2010 accounts, as follows:

“Many software companies have a large percentage of revenues that stem from professional services, because they have to do a lot of customisation work on the product for every single implementation. In contrast, Autonomy ships a standard product that requires little tailoring, with the necessary implementation work carried out by approved partners such as IBM Global Services, Accenture and others.”

The point was repeated at page 18 of the document, where Autonomy set out the proportion of its revenues derived from services, as it had to under IAS 18:

“Services. Services revenues relate to third party and internal implementation consultants and training. Services revenues remained flat in 2010 at approximately 5% of revenues (or $10 million to $11 million per quarter) (2009: $9 million to $11 million per quarter). Autonomy operates a rare "pure software" model under which our goal is that most implementation work is carried out by approved partners. This optimises Autonomy’s ability to address its horizontal technology to multiple vertical markets and regions in the most efficient way.”

The Defendants added to their own insistence that the statements could not be taken to mean that Autonomy did not sell hardware ^{^[233]}the contention that in point of fact that was not how the statements were understood either by Autonomy’s auditors, or by the Claimants’ own witnesses; nor indeed by HP and those acting for it, until later when the exigencies of formulating a claim made it expedient for them to contend otherwise.

As to Deloitte, who had a detailed knowledge of the volume and nature of the hardware sales:

(1) Mr Welham stated in his evidence when cross-examined in the US criminal proceedings that he understood the reference to “ pure software model ” as distinguishing Autonomy from software companies that have a large percentage of revenues that stem from professional work^{^[234]} : and that he did not understand the phrase to mean that Autonomy only sold software. He expressly confirmed that evidence when cross-examined in these proceedings. Further, he told me that he did not think that the reference to the “ pure software model ” in the 2010 Annual Report, “in the context it was placed in” , was in any way inconsistent with the fact that Autonomy had made hardware sales of around $100 million in 2010.

(2) Taylor Wessing made this point on behalf of Deloitte in their January 2015 letter to the Claimants’ Solicitors^{^[235]} :

“Contrary to your letter of claim the hardware sales were not inconsistent with Autonomy's description of itself as a 'pure software' company. It is quite clear from the Business Overview sections of Autonomy's financial statements on which you rely that the company was not representing that it only sold software. The words on which you rely appear immediately after a paragraph addressing Autonomy's sales of appliances and states: "Autonomy is one of the very rare examples of a pure software model. Many software companies have a large percentage of revenues that stems from professional services, because they have to do a lot of customisation work on the product for every single implementation. In contrast Autonomy has a standard product that requires little tailoring ... ". It is clear that Autonomy uses the term 'pure software company' to distinguish itself from other software companies which derive significant revenues from selling services alongside software.”

(3) Similarly, the Deloitte Defence in the FRC proceedings noted that:

“The words ‘pure software model’ were not inconsistent with any assertion in the financial statements. The audit team reasonably understood these words to draw a distinction between Autonomy’s business model and those of rivals who derived a substantial proportion of their revenues and profits from professional services. … There is no assertion in the financial statements that hardware was or was not sold and a statement that Autonomy followed a “pure software model” was entirely consistent with the assertions in the financial statements.”

In support of their submission that the Claimants and their witnesses themselves understood that the statements that Autonomy was a ‘pure software company’ were simply differentiating Autonomy from the norm of a company deriving substantial revenue from services the Defendants contended that:

(1) Mr Apotheker had a similar understanding of the phrase:

“Q. ... Go back to page 15 {K15/369/15}. It's making it quite clear, isn't it, what it means by this? "Financial model": "[...] very rare examples of a pure software model. Many software companies have a large percentage of revenues that stem from professional services [...] In contrast, Autonomy ships a standard product that requires little tailoring [...]"

A. Yes.

Q. That's the point they're making, that they're not providing lots of services?

A. I agree.

Q. That's all they mean by this phrase "pure software model"; that's how you understood it?

A. Yes, I understood that this company was in the business of providing excellent software with as little services as possible.

Q. Right.

A. That's how I understood it.”

(2) Similarly, Mr Khan and Mr Gersh both understood that the phrase was being used to distinguish Autonomy from a company that sells services.

(3) When cross-examined, Mr Holgate was constrained to accept that in context, the phrase was used to denote that Autonomy was not a service company. He agreed also that Autonomy was saying “we’re not like those other companies that sell lots of services”. However, he qualified this by saying that he “would not infer from that description that they sell hardware on an undisclosed basis…” though I do not think that had been suggested by any party, and certainly not by the Defendants, the issue raised being whether a positive statement was made that Autonomy did not sell hardware.

Nor, the Defendants submitted, could the Claimants get any assistance from the reference to appliance sales being a small part of Autonomy’s business: the paragraphs dealing with appliance sales in the 2009 Annual Report said nothing about whether Autonomy sold any non-appliance hardware, either expressly or by implication. Further, and contrary to the Claimants’ argument, they cannot somehow be combined with the reference in a different part of the Annual Reports to the “ pure software model ” to convey the impression that Autonomy was not engaging in sales of non-appliance hardware.

The Defendants also contended that:

“the market could not assume, nor is there evidence suggesting it did assume, that the term “pure software company” denoted that the company sold no hardware, given that, even on the Claimants’ case, it was public knowledge that Autonomy sold some hardware appliances. Although Mr Gersh, when cross-examined, sought to depict a sale of an appliance as a sale of software, and the hardware on which such software was loaded, not as hardware but as “the form of the delivery of the software” or a “software solution”, he eventually had to concede that it nonetheless constituted in part at least, a sale of hardware.”

This was not a complete and accurate presentation of Mr Gersh’s evidence, as I shall explain when addressing later the issue of HP’s pre-acquisition knowledge (see paragraphs 1799 to 1813 below). But the Defendants maintained that in any event, this was not the only evidence suggesting that HP itself did not really understand the statement ( “pure software company” ) to connote that Autonomy sold no hardware (other than limited appliance sales, using that phrase in the limited sense the Claimants insisted upon), and that in truth HP knew before the acquisition that Autonomy engaged in hardware sales, from which it is to be inferred that it would not have interpreted the statement to connote otherwise.

Accordingly, the Defendants rejected any suggestion that the reference to Autonomy being a “pure software” model or company was false or misleading in light of the hardware reselling. That was not in fact how it was understood or liable to be understood, as both Deloitte and the Audit Committee must also have been satisfied. In any event, the Defendants said that they did not intend that it be taken to mean that Autonomy made no hardware sales, nor did they calculate that it was likely (or, at least, liable) to be taken to mean that either.

In my judgment:

(1) As I have previously concluded, the purpose of the statement was to convey a special selling point, the success and self-sufficiency of its software business without the need for other revenue streams.

(2) That is how it was understood by the Claimants. The Defendants’ quotations from Mr Apotheker’s evidence were too selective. What I understood him to be telling me (and I accept the truth of this evidence) was that he was pleased to note that Autonomy provided very little servicing because that made him “believe that Autonomy provided brilliant software because it didn’t need a lot of services” , but he also thought that the statement meant that Autonomy regarded itself as different from other companies that produced software because, as he put it, Autonomy was

“a software company that only does software, so it makes total sense that they are trying to do as little as possible services. But they basically do two things in this annual report: they sell software and they provide some services.”

(3) A determination of what the Defendants themselves intended to be conveyed by the phrase is inevitably informed by my conclusion that the real purpose of the hardware reselling strategy was to generate recognised revenue from hardware sales but (a) account for them without differentiation as if they had been generated by software sales and (b) attenuate and disguise the adverse effect of almost invariable losses on the hardware sales by accounting for them as sales and marketing expenses.

(4) Having reached that conclusion, it seems to me that it would be odd to conclude that nevertheless the Defendants did not mean to further the concealment by the use of an ambiguous phrase.

(5) In other words, in light of my other conclusions, I have further concluded that the phrase was intended by the Defendants to conceal hardware sales.

(6) But that is primarily because of my conclusion that the hardware strategy and its potentially adverse effect if properly accounted for, needed to be concealed to succeed. The question now posed is what the position would be had I concluded that the hardware reselling programme was genuinely in furtherance of its software business and there were good commercial reasons for not disclosing it or its effect such as to justify it not being disclosed.

(7) It is not easy to recast my approach, but it seems to me to follow that had that been my conclusion, I would also have concluded that the Defendants might well honestly have intended the phrase to convey in 2010 as it did at the beginning of 2008 (before the hardware reselling programme began) that unlike many other software companies, services were not a material part of its business or revenues. There are certainly examples of that being spelt out. For example, at the Q3 2010 Earnings Call, Mr Kanter (speaking with and at the invitation of Dr Lynch) stated:

“…there’s a continuous debate on whether the Autonomy business model should be pure software or whether there should be a shift towards more professional services in the revenue mix. To date, the model has been maintained at the pure software end of the spectrum, i.e., little services revenue in the mix.”

That brings the question back again to purpose, pretence and concealment. In my judgment, without proof of those three, the alternative way of putting the case might have failed.

That the disclosed revenue categories comprised all sources of Autonomy’s revenue

Secondly, the Claimants contended that the revenue stated in Autonomy’s disclosed revenue categories in its Annual and Quarterly Reports was equal in aggregate amount (across the categories) to Autonomy’s total reported revenue. This, they contended, gave the same false impression, and would have conveyed to the reasonable reader, that there were no other material sources of revenue, and thus that there was no (or no material) revenue from hardware sales or indeed any other sources. In addition, the inclusion of revenues from hardware sales within the IDOL Product category, without any disclosure or explanation that this was so, also falsely inflated the revenue falling within IDOL Product. They submitted that all of this must have been apparent to the Defendants.

The Defendants did not accept this either. They countered the Claimants’ argument that the non-IFRS “supplemental metrics” included in Autonomy’s Quarterly Reports from Q3 2009 to Q2 2011, and in the 2010 Annual Report, which broke down Autonomy’s revenues into IDOL Product, IDOL OEM, Services and Deferred Revenue Release, and at a later point IDOL Cloud, were defined in a way that could not properly include hardware and further, left no headroom for hardware revenue, as follows:

(1) As the Quarterly Reports and 2010 Annual Report made clear, these were metrics “ provided for background information and may include qualitative estimates. ” The metrics were not precisely defined.

(2) Deloitte scrutinised the calculation of the non-IFRS metrics each quarter. They were well aware that Autonomy’s hardware sales were included in the category “IDOL Product”. Each quarter, they prepared a working paper, the stated aim of which was “ To agree the metrics used in the quarterly press release to the supporting schedules and to test the validity of these schedules. ” Mr Welham confirmed that, as part of the audit or review process, Deloitte would have reviewed the underlying contracts for at least some of the transactions included in those schedules, all those that were over $1 million in value, and those that were part of the sampling process. He said that Deloitte knew that certain deals included within IDOL Product were hardware deals: indeed, that is clear from the face of the working paper, and from Deloitte’s tickmarks in their working papers.

(3) Deloitte were aware of how IDOL Product was described and what its category comprised in the Annual Reports, as demonstrated by the ticked off versions, where the number is ticked next to the description in question.^{^[236]} They knew that some of the hardware sales did not include an IDOL software component.^{^[237]} They did not consider that this rendered Autonomy’s Quarterly or Annual Reports misleading, as Mr Welham (unsurprisingly) confirmed under cross-examination:

“Q. Then looking at what you did know, going back for example to IDOL Product, you knew that as part of the total amount that was being stated as IDOL Product, that included the hardware deals that we've looked at?

A. Yes.

…

Q. … So you knew those facts, you didn't think that the way that then Autonomy presented itself to the financial markets through its published information was misleading in any way, did you?

A. We did not, no.”

The Defendants also maintained that, contrary to an important premise of the Claimants’ argument, the market was also aware that IDOL Product could include hardware revenue. Thus, for example, they cited Mr Morland’s evidence in cross-examination as demonstrating this:

“ MR SHIVJI: Mr Morland, you knew at the time at the time, this is 2009 to 2011, that Autonomy included some sales of hardware in its revenue figures?

A. Yes.

Q. And you must have realised that they would be included in its IDOL product category?

A. Yes. ”

However, that citation of Mr Morland’s evidence, which would on its own be highly material, was incomplete. A little later in his cross-examination, when Mr Shivji came back to the point, Mr Morland clarified what he had meant:

“Q. But you understood that within licences there was a hardware element?

A. An immaterial element, yes.”

In any event, Dr Lynch emphasised that he had no involvement in the compilation of the “ supplemental metrics” ; and Mr Welham agreed that Deloitte dealt with the finance team and the Audit Committee in the course of the audit process and Dr Lynch really had no substantive involvement with Deloitte, or certainly none with him.

If set in the context of the conclusions I have reached as to the true purpose and concealment of the hardware reselling strategy, none of the Defendants’ arguments is attractive or indeed tenable. Set in that context, I would regard the “supplemental metrics” as having been intentionally calculated, as part of the imperative to conceal the hardware sales to present Autonomy’s business as comprised only of the elements identified, and as not involving any other material activities. I would not regard reliance on Deloitte as a viable argument in circumstances where Deloitte had been unaware of the true purpose of the strategy and its concomitant, the imperative to conceal it. And I would not accept that Dr Lynch was not involved in this furtherance of the strategy by these additional means.

However, if the building bricks of purpose and its concealment are removed, the position is much less clear. If accepted that the hardware sales were truly driven by the protection and promotion of the software sales, so as in effect to be a cost of it (as indeed Dr Lynch sought to depict the strategy) the “supplemental metrics”, with the warning to which they were subject, would not appear to me necessarily to have been calculated to mislead. Nor would I accept subjective intention to mislead would have been demonstrated. I would not accept, even then, the defence that Dr Lynch was not involved; but that would not save the plea.

Once again, in my judgment, without proof of purpose, pretence and concealment, the alternative way of putting the case might have failed.

The representation that transactions in 2009 were the same in nature as those in 2008

Thirdly, the Claimants contended that the statement in the 2009 Annual Report under “Revenue Recognition” section within “Significant Accounting Policies” that “The nature of the transactions that the group has entered into during 2009 is the same as in 2008 in all respects” was untrue or misleading because Autonomy had not carried out any ‘pure hardware’ sales in 2008 and its resales of $56 million in 2009 constituted a clear change.

The Defendants dismissed this as “not a valid complaint” . They submitted:

(1) This statement should be seen in context. It does not appear in a section describing Autonomy’s business, but instead in a note dealing with significant policies. The paragraph in which it appears reads as follows:

“The group discloses revenue within two categories, namely sale of goods and rendering of services, as required by IAS 18. During 2009 there has been no change to the group’s revenue recognition policies in any respect. The nature of the transactions that the group has entered into during 2009 is the same as in 2008 in all respects. To assist the reader in understanding the group’s business the accounting policy set forth below has been reviewed and clarified, but does not represent any change in the group’s accounting policy for the recognition or measurement of revenue.”

(2) Thus, the phrase is dealing with the nature of the transactions seen from the narrow perspective of revenue recognition. The reader would not expect to see the types of goods sold by Autonomy enumerated in that context, or any change in the types of goods identified.

(3) The Defendants did not accept that there was any relevant change. As stated previously, Autonomy had always sold some hardware.

(4) There was no suggestion in cross-examination that Dr Lynch was aware of, or responsible for, this alleged misstatement.

(5) Nor was there any evidence that HP had noticed or placed any reliance on the statement. It would be surprising if it was a point of any importance at the time of the acquisition over 18 months later.

I would not accept the contextual point in paragraph 1642(1) in any event; and in my view, there plainly was a change. The sales of hardware were of a different order in 2009 when compared to 2008. Nor, again, would I accept that Dr Lynch was not relevantly involved, though I could see an argument that he may not have focused on the presentation, and though that would be careless and a breach of duty it might not be sufficient to establish dishonesty or recklessness for the purposes of a FSMA claim. However, I need not finally determine the point: because, in my judgment, once again without proof of purpose, pretence and concealment, the alternative way of putting the case, would have failed. Deloitte and the Audit Committee were satisfied, on what they understood of the hardware sales and assumed to be their genuine purpose, that they were not material.

Falsity of inclusion of hardware sales in “Organic Growth”

Fourthly , the Claimants contended that the inclusion of revenue from ‘pure’ hardware sales within the revenue figures given for “organic IDOL growth” , without any note or warning, was obviously and materially false and misleading.

They highlighted especially the Financial Review section in the 2009 Annual Report, which stated that the increase in revenues in 2009 “ is a combination of strong organic growth and the successful integration of Interwoven ” (a company Autonomy had acquired in early 2009) .

This was, according to the Claimants, untrue or misleading, since ‘pure hardware’ sales were responsible for $53 million of Autonomy’s revenue growth in 2009. Similarly, the Financial Review section in the 2010 Annual Report ascribed Autonomy’s reported revenue growth between 2009 and 2010 entirely to the deployment by customers of Autonomy software. Autonomy reported an increase in revenues of $130.7 million in 2010: $52 million of this growth (some 40%) arose from sales of ‘pure hardware’ .

The Claimants relied especially on the evidence of one of Dr Lynch’s own witnesses, Mr Shelley (Co-Head of Corporate Broking first at UBS, one of Autonomy’s corporate brokers, from 2004-2010, and then at Goldman Sachs, who joined UBS and Citi as a corporate broker for Autonomy, in around May/June 2011). This evidence was to the effect that he understood that “organic IDOL growth” was a measure of growth in Autonomy’s core software business. Autonomy’s 2010 Annual Report described “organic IDOL growth” as the “most meaningful organic performance metric for understanding the momentum within the business.”

One of the Claimants’ witnesses, Mr Morland, concurred in that, and stated in his witness statement that as far as he was concerned,

“the key valuation metric is organic software growth because, in indicating underlying, repeatable growth, it provides the best measure of how successful and competitive a software company’s products are in the market(s) in which it operates.”

When it was put to him in cross-examination that Autonomy did not publish any figures for “organic software growth”, Mr Morland explained that when Autonomy said “organic growth” that, to him, had the same meaning, even though he accepted that it included an “immaterial” hardware element.

The Claimants prepared calculations by reference to figures given in Autonomy’s Investor Relations Bulletin (“the Bulletin”) to show that if the ‘pure hardware’ revenue had been excluded from “organic growth”, as they contended it should have been, such “growth” would have been negative:

(1) Whereas Q4 2010 growth was stated in the Bulletin as 12%, without pure hardware revenues, there would have been negative growth of 5%;

(2) Similarly, whereas FY 2010 growth was stated in the Bulletin to be 17%, without pure hardware revenues that would have reduced to 7%.

Further, when cross-examined about a draft of the Q1 2011 Quarterly Report that Derek Brown had sent him on 20 April 2011, Mr Shelley (see paragraph 1647 above) stated:

“Q. In the last couple of lines [of {K17/356.2/2}] it says that reported IDOL Product revenue was 54.4 million up from 46.5 million for quarter 1 2010 and that’s the 17% increase they’ve reported, yes?

A. Yes.

Q. Now, I’d like you to assume for a moment that the 46.5% figure for quarter 1 2010 included 6 million of revenues from pure hardware sales, yes?

A. Right.

Q. So that’s 40.5 million if you take out the hardware?

A. Yes.

Q. And assume that the 54.4 million for quarter 1 2011 included 18 million of revenues from pure hardware sales, making it 36.4 million, yes?

A. Yes.

Q. And if those assumptions are right, what it would mean is that instead of going up by 17%, the software part of IDOL Product had actually declined by more than 10%, yes?

A. Mm-hm.

Q. Can we flick back to page 1, please {K17/356.2/1}. Do you see at the bottom of the page there’s a section called “Chief Executive’s Review”?

A. Yes.

Q. And this is a statement from Dr Lynch, yes?

A. Yes.

Q. He says at the beginning: “Q1 was a strong quarter for Autonomy in which we continued to execute well with good growth in revenue, profits and other key metrics.” Do you see that?

A. Yes.

Q. In the fourth line he says: “In Q1 2011 IDOL Product, driven by licence growth, increased by 17%.” Do you see that?

A. Yes.

Q. What he’s saying there is that IDOL Product revenues are because the licence revenue was up, yes?

A. Yes.

Q. If what I just told you about the hardware revenues is correct, then the claim that growth in reported IDOL Product revenue was driven by licence growth would be completely wrong, wouldn’t it?

A. If what you’ve just told me is correct, then potentially, yes.”

The Claimants pointed out that this apparently significant evidence was not addressed by the Defendants, neither of whom made any mention of Mr Shelley in any of their submissions.

Again, the Defendants denied that Autonomy’s published information contained any false representation in this regard.

As to the Claimants’ argument that Autonomy’s statements about “ organic growth ” and “ organic IDOL growth ” were misleading in light of the hardware sales, the Defendants submitted that:

(1) Once it is accepted (as is common ground) that the hardware revenues were correctly included in total revenues, there can be no objection to the inclusion of hardware revenues in a calculation of organic growth. In ordinary parlance, organic growth means growth generated organically by a company rather than through acquisitions. This is how Autonomy used the term: as stated in the Financial Review in the 2010 Annual Report, organic growth “ excludes the contribution from acquisitions, foreign exchange impact, services revenue (not a goal of the business) and deferred revenue release (primarily maintenance income). ” There was no indication, or implication, that hardware revenue was being excluded. As Dr Lynch put it, to exclude hardware from the calculations would be to produce a different metric, “ organic growth ex hardware ”.

(2) Deloitte were ultimately content to approve the inclusion of hardware sales within organic growth. It is true that they were initially doubtful, especially about Autonomy’s inclusion of them under the heading “IDOL organic growth” : the hardware sales were obviously not sales of IDOL, and Deloitte wanted to understand and be persuaded why they were organic.

(3) Mr Knights expressed this concern directly in his email to Mr Hussain and Mr Chamberlain dated 16 October 2009 (which Mr Hussain forwarded to Dr Lynch):

“Can we have a detailed breakdown on how the figures are compiled. My biggest concern will be that the hardware sales were neither IDOL based nor organic!!”

(4) In an email to Mr Chamberlain and Mr Knights (cc Mr Welham) two days later (18 October 2009) Mr Knight of Deloitte focused on the “organic growth issue ” as almost the sole issue of remaining concern, elaborating that:

“We currently still have an issue with organic growth, all we have seen to date is a summary calculation which has “assumed interwoven sales” and no account taken for the hardware sales. The reader will probably assume this is therefore all organic software growth.”

(5) Deloitte (through Mr Knights) ultimately concluded that it was appropriate to include the hardware sales on the basis that (as Mr Welham accepted in his witness statement):

“Ultimately we agreed that growth in hardware sales was organic, in that it did not derive from the acquisition by Autonomy of a pre-existing business” .

(6) Deloitte signed off on Autonomy’s organic growth figures in the non-IFRS metrics knowing that they incorporated hardware, so long as the entry was captioned “organic growth” (as Deloitte themselves suggested) and not as “organic IDOL growth” nor as “organic software growth”.

The Defendants also contended that the criticisms made by the Claimants about the description of growth in the 2009 and 2010 Annual Reports could not withstand scrutiny.

(1) The statement in the 2009 Annual Report that the increase in revenues in 2009 “ is a combination of strong organic growth and the successful integration of Interwoven ” was, they said, correct. Although the growth in hardware revenues was one component of the increase, that was properly included in the organic growth figures. They submitted, therefore, that the Claimants were wrong to suggest that it was misleading.

(2) This statement could be of no further assistance to the Claimants than that relating to the inclusion of hardware revenue in the figures given as “ supplemental metrics” : once the hardware revenue was included within those metrics, the difference between organic and non-organic growth did come down to the difference between revenues generated by the business without taking into account acquisitions and revenues generated by the business taking into account acquisitions.

(3) The Claimants’ suggestion that in the 2010 Annual Report, “ the Financial Review section (page 15) ascribed Autonomy’s reported revenue growth between 2009 and 2010 entirely to the deployment by customers of Autonomy software ” is incorrect: there is no such statement in the Financial Review.

(4) Their criticism as put to Dr Lynch in cross-examination that he must have spotted, and ought, in all honesty, to have corrected, an inaccurate description by Mr Briest of UBS (one of Autonomy’s corporate brokers) in his analyst’s note in late 2009, referring to Autonomy’s organic growth as organic software growth was unfair: the statement was buried towards the end of a 57-page document. Dr Lynch’s evidence was that, although he had read earlier parts of the note, and corrected certain aspects of the description in it of IDOL technology which Dr Lynch considered Mr Briest “had got very wrong” , he would not have noticed the point in a different section tucked away at the end. In any case, the Defendants contended that this is miles away from being any representation by Dr Lynch, or any part of Autonomy’s published information that can be relied on for the purposes of the claims before the Court.^{^[238]}

(5) Dr Lynch added that in any event, he was quite removed from the process. Mr Miles submitted in his oral closing:

“He is not involved in the production of the accounts, he’s less involved in the review of the accounts even than other directors such as Mr Bloomer and Mr Webb, and he knows that these things are being carefully considered by Deloitte, who are considering them to ensure that nothing misleading is being said.”

(6) Finally on this point, the Defendants countered the Claimants’ contention that “if the pure hardware revenue had been excluded from organic growth, as it should have been, organic ‘growth’ would have been negative” [their underlining] with comparative tables of their own demonstrating the reverse. According to the Defendants’ calculations (which Mr Miles put to Dr Lynch in re-examination), if the hardware transactions were stripped out then both for Q2 2011 and H1 2011 organic revenue growth would have been significantly higher (in each case, compared against the equivalent period the previous year). Further, again according to the Defendants’ calculations, the Core Business Organic Growth Rate Calculation in Autonomy’s interim results up to 30 June 2011 gave growth rates of 15% (Q2 2011) and 17% (H1 2011). Adjusting these figures to remove the “pure hardware” sales listed by the Claimants in Schedule 1 of the RRAPoC, those figures would have increased to 28% (Q2 2011) and 21% (H1 2011). The Defendants submitted that these were the key periods for comparison since the key comparator at the time of the Autonomy acquisition would have been Q2 2011 v Q2 2010, as Mr Apotheker had confirmed when cross-examined: judged at that time, growth rates as at Q4 2010 were out of date, so the Claimants’ figures based on Q4 2010 results “addressed the wrong period”.

Once again, it would seem to me that the claim would be well founded in the context of my primary conclusions, as being a further manifestation of the disguised purpose and the imperative to conceal it. However, I must seek to assess the claim on the hypothesis previously explained.

I have found the issue more finely balanced in relation to this representation, principally because Autonomy was expressly warned by Deloitte, in an email from Mr Knight to Mr Chamberlain dated 18 October 2009, which was plainly relayed to both Defendants (and which I have quoted more fully in paragraph 1654(4) above), that “The reader will probably assume this is therefore all organic growth”.

Although Deloitte had eventually accepted Mr Chamberlain’s counter-argument that all that “organic growth” connoted was that the growth was not from acquisitions, that does not diminish the propensity of the phrase to mislead, nor negate the fact that the Defendants were warned and aware of that propensity and its attendant risk that “organic growth” of a “pure software company” would be read as “organic software growth”, as indeed it was (including by Autonomy’s own broker, Mr Shelley, as described in paragraphs 1647 and 1651 above). It is a short step from knowledge that a statement may be liable to mislead to a conclusion of recklessness or dishonest intention to mislead if it is used.

Even so, I have concluded that, on the premise which I have taken to be applicable, the mindset of those concerned at Autonomy, including the Defendants, would have been that sales as loss leaders for the software business generated revenues to be treated as all part of the single software enterprise in a single segment entity, and the inclusion of revenue from hardware sales within organic growth was a corollary.

Overall conclusion on the claims in respect of statements made in the accounts

There can be no real doubt, and I do not understand it to have been disputed, that the Defendants wished to avoid disclosure of the nature and extent of the hardware sales. The Claimants’ case was that it was part and parcel of the true objectives of the hardware reselling strategy and the imperative to conceal it. The Defendants rejected that but prayed in aid (in summary) commercial sensitivities. Although that is the backdrop, and is to be borne in mind in assessing intention, for the present, my focus is on what Autonomy chose to say, rather than on what they chose not to say. The question now is whether the Defendants not only sought to avoid disclosure, but also included in Autonomy’s published information statements designed, or which the Defendants knew would or might be taken, to signify that no material part of its business comprised the sale or reselling of hardware to third-party customers.

For the reasons I have given in respect of each of the positive statements relied on by the Claimants, on the hypothesis that (contrary to my primary conclusions) the purpose of the hardware reselling and the reasons for not separately disclosing it are taken to be as the Defendants asserted they were, I would have not been persuaded that the Claimants had established their claims that the statements made were untrue or misleading and that the Defendants appreciated this so as to render them liable under FSMA.

I confirm that I have also considered the statements in the round as well as individually. Although in the context of my actual conclusions, the picture is then darker still, on the hypothesis I have been implicitly directed to take, my conclusion would be no different than in relation to the representations singly. I have concluded that even together the statements cannot be aggregated to produce a false and intentionally misleading positive representation that Autonomy was not reselling hardware.

However, the overall presentation certainly did not disclose hardware reselling: and I turn to the second way in which the Claimants put their case on the hypothesis that the purpose of the hardware reselling was as asserted by the Defendants.

Did the Annual Reports/ Autonomy’s published information omit information about hardware sales which was required to be disclosed?

With the reservation that if their case on what Autonomy had positively stated in its published information succeeded, their case that there were also omissions might be unnecessary, the Claimants also contended that Autonomy’s published information “was defective because it did not disclose the fact, nature and extent of its hardware sales in its Annual Reports for 2009 and 2010.”^{^[239]}

They submitted that:

“ the relevant accounting standards and other rules required fair disclosure and explanation of the nature and extent of Autonomy’s hardware sales, including sales of pure hardware and other hardware in the Annual Reports .”

Their pleaded case in this regard as set out in the RRAPoC is summarised there as follows:

“In short, the absence of any disclosure in the Annual Reports…of the existence or extent of hardware sales (other than the unquantified reference to appliance sales being a small part of Autonomy’s business) meant that the statements [identified above] were untrue and/or misleading , those reports gave a misleading impression of the revenue and revenue growth of Autonomy’s software business and/or the Annual Reports omitted a material fact (namely that Autonomy was engaged in the business of selling significant amounts of pure hardware at a substantial loss) that was required to have been included in Autonomy’s published information.”

Leaving aside positive misstatements, only the omission of “any matter required to be included” can be relied on as the basis for liability under FSMA Schedule 10A. Further, an issuer is liable only if the PDMR (here, each Defendant) “knew the omission to be a dishonest concealment of a material fact.”

Thus, to establish liability for an omission, the onus is on the Claimants to demonstrate on the balance of probabilities that:

(1) Disclosure of the omitted matter was “required to be included” (that is, mandated by applicable legislation or accounting standards)^{^[240]} in the relevant published information;

(2) The relevant Defendant, being a PDMR, had actual knowledge, at the time of the omission of the published information in question, of (a) a material fact which (b) he also knew was required (in the sense explained above) to be disclosed but which instead (c) was being “concealed” (that is, deliberately being left out).

The issues as to the materiality of a fact which had not been disclosed and as to whether disclosure was “required” turn on Accounting Standards and Practice, on which there was detailed opposing expert evidence. The issues of what might be termed “guilty knowledge” and dishonest concealment are issues of fact. I address first the accountancy issues and thereafter the factual issues.

Did Autonomy ’s published information omit a material fact required to be disclosed?

In arguing that such disclosure was “required” the Claimants relied in particular on IAS 18, §35; IAS 1 §29; and IFRS 8, §32. Alternatively, even if disclosure was not required under specific IFRSs, they contended that additional disclosures were necessary under IAS 1, §17(c) in order to achieve a fair presentation in the Annual Reports.^{^[241]} I shall address those provisions in turn by reference to the expert evidence as to the application of these provisions and in the order that Mr Holgate dealt with them in his expert report.

Before doing so, however, the confines of the issue seem to me best observed by recording that the following matters were and remain common ground:

(1) As previously noted, the revenues from hardware sales were properly included within the overall revenues of Autonomy in compliance with IFRS;

(2) There was no requirement on Autonomy to disclose hardware revenues in its quarterly or interim reports;

(3) None of the Companies Act 2006, the Combined Code or the Disclosure and Transparency Rules required disclosure;

(4) Autonomy properly reported that it had a single operating segment for the purposes of IFRS 8 and this was not gainsaid by the Claimants;

(5) The accounts were prepared or scrutinised by the finance department (made up of a number of experienced accountants), Deloitte and the Audit Committee, and all three groups considered and decided on the appropriate accounting treatment.

IAS 18.35

Mr Holgate cited the requirement in IAS 18.35 to disclose:

“the amount of each significant category of revenue recognised during the period, including revenue arising from: (i) the sale of goods; (ii) the rendering of services …”

There is no dispute between the experts that IAS 18.35 was mandatory ( “An entity shall disclose…” ). The issue is whether the requirement to disclose “each significant category of revenue” extends to providing a further breakdown of the goods sold or services rendered within those categories.

Mr Holgate’s opinion, as given in his evidence, was that hardware sales (a) were very different in their nature and economic effect from the (disclosed) software sales, with radically different profit margins, so as to comprise a “category of revenue” and (b) the quantum of the hardware sales was such as to render them “significant”.

When cross-examined, Mr Holgate elaborated on his view as follows:

“… in many circumstances which are straightforward, then this disclosure requirement [in IAS 18 §35] would be met by saying sale of goods X, rendering of services Y, interest separately and so on. But the circumstances here, as I’ve outlined, are very unusual in terms of the growth of hardware, the impact on growth percentages and the vastly different gross profit percentages involved. So we’re not in a normal situation, I don’t believe. So it is important to read the requirements which is the amount of each significant category of revenue recognised in the period and then including this and that. But, to me, the hardware revenue is a significant category of revenue because, if you don’t know about it as a separate component, then you are easily misled about the growth, where the growth has come from, where the profitability has come from, the effect on gross margins and the overall picture of what is described as a pure software company, if there’s hardware within there with very different characteristics, economic characteristics from the main software business. To not disclose that, I think has various problems including true and fair view and fair presentation and, more specifically here, it is in my view a significant category of revenue for the reasons I’ve given.”

However, Mr Holgate accepted that there was no definition of what counted as a “category”. There is nothing in the wording of the rule that requires a further breakdown into sub-categories. Nor was Mr Holgate able to point to any literature which suggested that a breakdown within the listed categories was required.

Mr Holgate appeared at one point to argue that the interpretation of the rule should be conditioned by such matters as growth rates, margins, and the scale of the relevant sales; in other words that these things should affect the interpretation of the rule in some way. However, he did not explain how these factors could affect the interpretation of the mandatory rule.

Mr Holgate also raised the spectre of an argument, echoing that of the Claimants, that the reference to Autonomy as a “ pure software company” was somehow relevant to this rule. But he then accepted that, in context, that was used to point out the differences between Autonomy’s business and that of a service provider. Again, he did not offer any real explanation as to how that could affect the interpretation of IAS 18.35 or cause it to be read as requiring a further breakdown below the level of “categories” of revenue.

Mr MacGregor accepted that the hardware sales were indeed “significant” and that they were different in nature from sales of software. However, he took a different view about the application of IAS 18.35, focused on the words “category of revenue”.

Mr MacGregor’s view was that the categories of revenue described in this standard are general categories of sale of goods, rendering of services, royalties, et cetera . Mr MacGregor relied upon IAS 18.1, which sets out the scope of the standard and explains that it covers revenue in five categories. These are then replicated in IAS 18.35. He explained that the purpose of IAS 18.35 is to require the separate statement and disclosure of revenue in these categories, if it is significant. In his view, no further sub-categorisation or breakdown of the revenues within these categories is required.

Accordingly, Mr MacGregor’s evidence supported the approach in fact taken by Autonomy, which was audited by Deloitte. The accounts divide the revenues into the categories of (i) sales of goods and (ii) rendering of services but do not provide a further breakdown of those categories.

Furthermore, and although sometimes rather didactic in his approach, Mr Holgate accepted that in relation to this specific issue the view taken by Deloitte, that no sub-categorisation was required by the rule, was not outside the range of a reasonable accountant’s views.

Mr Rabinowitz pressed Mr MacGregor in cross-examination on what meaning he thus gave to the words “including revenue arising…” [my emphasis] which might appear to suggest that the list which followed was not exhaustive; but Mr MacGregor at least three times repeated his view that in its context the list was indeed intended to be exhaustive.

Further, Mr MacGregor rejected Mr Holgate’s suggestion that the rule may depend on variable metrics of performance of the kind he ventilated. He depicted that approach as importing a sort of segmental analysis: and he was adamant that “This is not an attempt by the standard setters to come up with a segmental analysis”.

Mr MacGregor considered that Mr Holgate’s approach would inject excessive uncertainty and subjectivity into what is stated as a simple and straightforward rule. Further, the Defendants argued that IAS 18.35 is only one of many applicable provisions: it is not the only means of bringing home the objective of IFRS. The suggestion that the Claimants’ interpretation was necessary to safeguard the objectives of IFRS ignored this.

To test his view, I suggested to Mr MacGregor in the course of his cross-examination an extreme example, of a company (I gave Siemens purely as an illustrative example) which made and sold both (say) mobile telephones and fridges, products with vastly different profit margins: I asked him whether he would nevertheless contend that they should be lumped together in one category as both being sales of goods. He said he would. Mr Rabinowitz took this up with him as follows:

“Q. And just to be clear, Mr MacGregor, …you take the view that you don’t have to disclose those separate categories, however different they may be, however important they may be to understanding how the company is working, you don’t have to disclose them separately, however material they may be for the ability of the investor to understand the financial position of the company; is that your view?

A. This is what the Standard requires...this is not the role of the Standard, to sort out segmental disclosures. If in Siemens, the mobile phone department is organised as a separate segment from the…white goods department, and…they’re managed separately and they’re reported on separately, then in the segmental information, which is required under IFRS 8, if that’s the way Siemens organises itself, then that information would be disclosed there.

On the other hand, if there was just one department dealing with all electrical goods, which includes mobile telephones and fridges, it wouldn’t. It comes down to the way management has organized itself, certainly under IFRS 8.

Q. So doesn’t that in your view, however material it would be for investors to know about it, they wouldn’t have to be disclosed under this Standard, the fact that the company was doing that?

A. This is the disclosure requirement…

…what was in the conceptual Standard doesn’t mean you disclose whatever somebody out there might think is interesting, material or significant. The disclosure requirements are set out in the IFRSs…

…

You’re trying to read into [IAS 18] something which seems to me to be along the lines of, because there is information someone out there might find reasonably useful, you need to disclose it. That is not the requirement.”

In his oral reply, Mr Rabinowitz disparaged this interpretation of “including” and the application of IAS 18.35 as “confused” and “plainly wrong” . In the Claimants’ written closing, it was submitted that the difficulty and flaw in Mr MacGregor’s approach was that it:

“produces an outcome that is entirely inconsistent with the general objective of IFRS, namely to ensure a fair presentation of a company’s financial position and its financial performance in a way that enables readers of its financial information to make economic decisions.”

The Claimants also cited in support two letters written to Autonomy by the Financial Reporting Review Panel (“FRRP”) on 9 September 2009 and 30 November 2009, in the latter of which the FRRP stated:

“The Panel further notes that it regards the categories listed in paragraph 35(b) of IAS 18 as a minimum disclosure only and would generally expect more disclosure from all but companies with relatively simple operations”.^{^[242]}

As to the Defendants’ interpretation, they submitted that whether the word “including” before a list is intended to be exemplary or exhaustive is a matter of interpretation having regard to the context. Further, the Defendants submitted, while there is room for debate about whether there could, in principle, be further categories of revenue, the question here is not whether the hardware sales should have been disclosed as another category. The Defendants’ position was that the hardware revenues represent revenues from the sale of goods, which is one of the express categories contained in the rule. The issue is whether the rule required further breakdown within that category; not whether there might be some other category of revenue beyond those listed. Hence, they submitted, the contention whether the rule is non-exhaustive is nothing to the point.

In such circumstances, it is necessary to consider the other provisions on which Mr Holgate relied: for it was the Claimants’ case that it is not just IAS 18.35 that would have compelled disclosure. The starting point is IAS 1.

IAS 1

IAS 1 is a general standard about the presentation of financial statements. It is not concerned specifically with revenue reporting, there being (as already apparent) specific standards prescribed in IAS 18 (revenue) and IFRS 8 (segmental reporting).

IAS 1.1

IAS 1.1 explains that the standard prescribes the basis for presentation of general purpose financial statements to ensure comparability both with the entity's financial statements of previous periods and with the financial statements of other entities. It sets out overall requirements for the presentation of financial statements, guidelines for their structure, and minimum requirements for their content. For present purposes, the provisions specifically relied on (as well as IAS 1.1) are 1.15, 1.17, 1.29, and 1.30. Mr MacGregor also relied on IFRS 8.

IAS 1.15

IAS 1.15 is concerned with the general requirement that financial statements “ shall present fairly the financial position, financial performance and cash flows of an entity.” It makes clear that:

“Fair presentation requires the faithful representation of the effects of transactions, other events and conditions in accordance with the definitions and recognition criteria for assets, liabilities, income and expenses set out in the Framework.”

The final sentence of IAS 1.15 states that “the application of IFRSs, with additional disclosure when necessary, is presumed to result in financial statements that achieve a fair presentation.” It continues that “ in virtually all circumstances, an entity achieves a fair presentation by compliance with applicable IFRSs ” [my underlining] although the paragraph then states that there may be some circumstances where fair presentation requires further disclosures (see IAS 1.17(C)).

IAS 1.17

IAS 1.17 similarly states that “In virtually all circumstances, an entity achieves a fair presentation by compliance with applicable IFRSs”. But it too, in sub-paragraph (c) repeated that a “fair presentation” also requires an entity:

“to provide additional disclosures when compliance with the specific requirements in IFRSs is insufficient to enable users to understand the impact of particular transactions, other events and conditions on the entity’s financial position and financial performance.”

The real dispute between the experts was, in the end, as to when “additional disclosures” would become necessary, and whether they did so in the circumstances of the hardware sales in this case.

Mr Holgate considered the “fair presentation” requirement in IAS 1 to be analogous to the “true and fair view” standard in the Companies Act. I would interpolate that this appears to me to be consistent with the approach of the FRC, and with the view expressed in an Opinion of Mr Martin Moore QC that the two are “synonymous” and “simply different articulations of the same concept.” ^{^[243]}

In that regard, Mr MacGregor’s opinion was that where all applicable IFRSs have been complied with, the threshold for further disclosure is a very high one and further disclosure will only be required rarely.

Mr Holgate accepted that the question whether further disclosure is needed under that rule notwithstanding compliance with the detailed specific provisions of IFRS is a matter of accounting judgement.

IAS 1.29

Mr Holgate placed special reliance on IAS 1.29, which provides as follows:

“An entity shall present separately each material class of similar items. An entity shall present separately items of a dissimilar nature or function unless they are immaterial.”

In Mr Holgate’s view, IAS 1.29 was a further standard that applied so as to require Autonomy separately to disclose its hardware sales because, both in relation to 2009 and 2010, those hardware sales were material and, further, those hardware sales were of a dissimilar nature to the other sales Autonomy was making (i.e. software sales). The test for materiality, in this context, is whether the information could influence the economic decisions that users make on the basis of the financial statements.

The Claimants contended that there are any number of factors that would highlight the dissimilar nature of hardware and software sales by Autonomy, including the very different profit margins that each type of sale would attract; software involves the sale of a licence, whereas hardware is a tangible product; software was created by Autonomy, but hardware was bought and resold with no added value. They also suggested that it is clear that Autonomy itself regarded selling hardware as wholly dissimilar to selling software, and indeed that Mr MacGregor did not in fact dispute that the products and their sale were dissimilar.

In terms of materiality, Mr Holgate in his report developed in some detail his reasons for concluding that hardware sales could not be considered immaterial. Thus:

(1) Mr Holgate identified, in this context, the volume of sales of hardware compared with total reported sales. His Table A was put forward as illustrating the fact that, for FY 2009, total hardware revenue constituted 7.3% of total reported revenue, and that, for FY 2010 total hardware revenue constituted 12.1% of total reported revenue. Mr MacGregor did not suggest that the revenue from hardware, for example in FY 2010 when it was over 12.1%, was immaterial in the context of the overall revenue reported by Autonomy for those periods.

(2) Mr Holgate further referred, in this context, to the effect of “ pure hardware sales ” on Autonomy’s revenue growth from 2008 to 2009 and also from 2009 to 2010. His Table B was put forward to show that, if the contribution made to revenue by hardware in each of those years was stripped out, so as to identify what growth there had been in Autonomy’s revenue from its software sales (across the board, including hosting):

i. rather than growth of 47% from 2008 to 2009, the figure would be 36.5% - a reduction in growth of some 22%;

ii. rather than growth of 17.7% from 2009 to 2010, the figure would be 11.5% - a reduction in growth of 35%.

Mr Holgate was not challenged on this analysis in cross-examination. Nor was he challenged on his evidence that “ growth in revenue is a major performance indicator for listed companies and one that Autonomy frequently cited” and that such a difference in these growth figures of software was “highly significant” . Mr MacGregor accepted that it was material, though he caveated that “It’s a smaller part than played by other parts of revenue” .

Mr Holgate further set out figures intended to illustrate both the very different gross margins that applied to sales of hardware and sales of software and the extent to which the gross margin applicable to sales of hardware affected the overall gross profit margin. Mr Holgate, in this context, also noted that it would have been important, in “ seeking to interpret an overall gross profit of 79%, … to know that the majority of sales are at a 90% margin and that a minority of sales are at a gross loss ”. Mr Holgate was not challenged on this in cross-examination.

Again, Mr MacGregor did not express disagreement about this, nor did he deny that being able to identify what gross margin applies to what products that an entity is selling was information that could be important for investors. Rather, and as foreshadowed in paragraph 1698 above, Mr MacGregor’s position was that on the basis of his understanding of Autonomy’s business, IAS 1.29 (indeed the whole of IAS 1) could not add anything to any disclosure obligation on the part of Autonomy in relation to hardware.

Mr MacGregor made the following principal points:

(1) First, in his view, disclosure of revenue was expressly covered by IAS 18.35 and no further disclosure obligation could, or perhaps should, be added by IAS 1 or IAS 1.29, which were more general in application, rather than being applicable specifically to revenue. Except in very unusual circumstances, compliance with the specific provision qualified as compliance with the general guidance as regards the specific accounting item.

(2) Secondly, Mr Holgate appeared to assume, but did not explain the basis for the assumption, that (in relation to the requirement of IAS 1.29 that “An entity shall present separately items of a dissimilar nature or function unless they are immaterial” ) hardware sales were to be treated as a separate and material “class” of “items” from software sales. The Defendants added to this that Mr Holgate had accepted in cross-examination that there was no definition of what is meant by “similar” or “dissimilar” , nor of “nature” or “function”; nor is there anything in the wording of the provision (at least nothing specific) that required a breakdown of revenue into sub-categories. Further, he had not suggested or provided any support in the literature for the view that the rule requires a breakdown of revenues into sub-categories and (the Defendants submitted) the wording of the rule does not support it.

(3) Thirdly, and as a further matter relevant to the interpretation of IAS 1.29 and (the Defendants submitted) supportive of Mr MacGregor’s view, IAS 1.30 appears to show that the reference to “ items ” in a set of financial statements is a reference to a “ line item ”. Hence revenue for sales of goods could be such a line item, revenue from services another line item, etc. Similarly, costs of goods sold are a different line item from administrative expenses. The rule says that each class of similar line items must be presented separately. It also says that if a line item (say sales and marketing expenses) is not material it need not be separately disclosed. These rules say nothing about the need to present a breakdown within a line item.

(4) Fourthly, and in any event, IAS 1, as a general presentation standard, had to be read together with IFRS 8.

Mr MacGregor made clear, however, at least when cross-examined, that his approach was informed by his understanding (on the basis of his instructions) that (as he put it):

“Autonomy considered it was essentially selling one overall product, and everything connected with the sale of that product was considered to be [the same] revenue.”

Mr MacGregor appeared to accept in cross-examination that if the sale of hardware had no connection with or relationship to the sale of software, “nothing to do with the sale of software” , then IAS 1.29 could apply and require disclosure:

“Q. So you accept that IAS 1.29 could apply? You’re just saying you don’t think it does if Autonomy - if Dr Lynch’s version of the facts is right and hardware was simply sold as a way of selling software?

A. That’s correct. I mean, if the reason for selling the hardware was not to sell the software but was for another reason, for example to pump up revenues, then clearly that is something which is material to an understanding of the business. In those circumstances. But in the circumstances where it’s incidental, connected with the sale of the software, then no. I appreciate there is a dispute over that.

Q. So just to be clear, I think you’re accepting that if the primary reason was actually to drive revenue rather than to drive software sales, there would have had to have been disclosure, including under IAS 1.29 -

A. That’s correct. Just to be clear, if what’s being suggested is that they are implying that what was in fact the sale of IDOL was in fact the sale not of that at all but it was basically being used to pump up revenues, then I would agree, that is something which the accounts should have disclosed. Those are the two - that seems to me the boundary or the bounds of the dispute and the effect on the presentation.”

That leads on to a further issue as to the application and effect of IFRS 8, and in particular IFRS 8.32.

IFRS 8

IFRS 8 is generally concerned with segmental reporting. The standard applied to Autonomy for the first time in respect of the full year accounts for 2009. It was then a new standard, adopted as part of an attempt to converge IFRS and US GAAP, and (although Mr Holgate disagreed) Mr MacGregor considered that there were uncertainties about its application.

Mr MacGregor gave as an example a debate as to the extent to which it replicated or overlapped with the requirements of IAS 18.35; and he also referred to a paper ( “Post-Implementation Review of IFRS 8 Operating Segments” dated July 2013) which the International Accounting Standards Board (the “IASB”) had issued to consider and report on various uncertainties which had caused debate and difficulty. These difficulties and uncertainties included the issue as to the meaning of “similar economic characteristics” and when operating segments fell to be aggregated ^{^[244]} .

As explained in Dr Lynch’s closing submissions, and put broadly, whether an entity should be accounted for as having one or more “Operating Segments” so as to distinguish the performance of each depends on the way the entity is managed, on the information flows within the company, and on identifying the “chief operating decision maker”. Speaking generally, where there is more than one line of business, it will be expected that there is more than one operating segment.

In the case of Autonomy, the finance department and Deloitte concluded in both 2009 and 2010 that there was a single operating segment and (as already explained) that conclusion has not been questioned in these proceedings.

IFRS 8.1

IFRS 8.1 identifies the “ core ” principle of IFRS 8 as follows:

“An entity shall disclose information to enable users of its financial statements to evaluate the nature and financial effects of the business activities in which it engages and the economic environments in which it operates.”

One of the curiosities of IFRS 8 is that although at first blush it might be taken that it applies only to determine whether an entity should adopt segmental accounting, and if it does, to provide for its more detailed application, it is common ground that at least one of its provisions can also apply to an entity with a single Operating Segment, such as Autonomy. The particular provision in issue is IFRS 8.32.

IFRS 8.32

IFRS 8.32 provides as follows:

“ An entity shall report the revenues from external customers for each product and service, or each group of similar products and services, unless the necessary information is not available and the cost to develop it would be excessive, in which case that fact shall be disclosed. The amounts of revenues reported shall be based on the financial information used to produce the entity’s financial statements.”

It is not in dispute that IFRS 8.32 in effect obliges an entity to report the revenues for “each product or service” that it provides to customers - or each “ group of similar products or services ” - unless the information is not available and cannot be made available without excessive expense. It is not suggested that this information was not available within Autonomy.

As already noted above, Mr Holgate saw no difficulty or uncertainty in the application of the standard in the case of Autonomy. He was of the view that the sale of hardware was indeed dissimilar to Autonomy’s software sales: hardware and software had to be regarded as different products or different groups of products; and that, since the volume of hardware sales was material, IFRS 8.32 clearly obliged Autonomy to make separate disclosure of its hardware sales.

Against that, and as in relation to the other potentially applicable standards, Mr MacGregor noted that the standard did not define “ similar products ”, which was therefore something in relation to which judgement was required. He also noted that materiality was, again, a matter in relation to which judgement was required. In each case the judgement would be fact-sensitive and nuanced . In his opinion, the question turned on the nature of the business being carried on by Autonomy; and I took Mr MacGregor to justify this approach in part by reference to the need to give some different content to IAS 8.32, ascribing to IAS 8 overall the objective of requiring proper disclosure of substantially separate and independently run businesses. Mr MacGregor stressed also that in his opinion, “judgment was (and is) the foundation when considering the application of IFRS 8.32” ; and he emphasised that when stating his overall view in respect of IAS 8.32:

“…judgement is required when determining whether an entity’s sales of products and services require separate disclosure. On the basis that Autonomy’s hardware sales were incidental to the sales of the core software product, I agree with the conclusions reached by Deloitte that separate disclosure was not required.”

Mr MacGregor told me that this was consistent with the way that other accountants and other companies (including those which, like Autonomy, had a single operating segment for the purposes of IAS 8) had looked at the matter, as had Deloitte in this case after careful consideration of IAS 8.32 and review also by its National Accounting and Auditing Team (“NAA”).

Applying that approach, and as in relation to IAS 1.29, Mr MacGregor explained his view that Autonomy was not required to make separate disclosure under IFRS 8.32 of its hardware revenues, on the basis that Autonomy undertook no separate or segregated business of hardware selling: the hardware sales were ‘incidental’ to Autonomy’s IDOL software sales.

Mr MacGregor’s view was that on the basis that Autonomy had only one overall product (IDOL) and the sales of such product were supported by the hardware sales no disclosure was necessary: on that basis, Autonomy was a seller of software and the other sales were, like everything else the company did, made in order to sell IDOL. As Mr MacGregor put it in cross-examination:

“…there is one overall product and that product is supported by essentially the hardware sales. That is the important thing. That is their business model.

…

My understanding is that Autonomy isn’t interested in hardware, it doesn’t consider itself to be a seller of hardware per se. What it considers itself to be is a seller of certain types of software and everything else it does is designed to support those sales.”^{^[245]}

Mr MacGregor’s general position was encapsulated as follows:

“Q. That [i.e growth figures] is plainly information of the sort that could affect the decision of users of the financial accounts, correct?

A. Well, it may do but if you’re not required to disclose it, then you’re not required to disclose it. I mean, Autonomy could have taken the view they would disclose it elsewhere but you’re not required to disclose it in a set of financial statements…

…

Q. …you do not dispute, do you, Mr MacGregor, that being able to identify what gross margin applied to what products an entity was selling is indeed something that could influence the economic decisions of users of accounts?

A. But it’s not - the breakdown of the gross margin is not something that is required by accounting standards. If you’re going to do it, then it’s done on a…voluntary basis.

Q. What’s the answer to my question? Do you accept that it is something that could affect the economic decisions of users of the accounts?

A. It might do. It might do…it might do but there’s all sorts of things that companies do that could affect the economic decisions of people out there who might invest or who might not. They’re not required to disclose them in the annual financial statements if it’s not required by the accounting standards.

Q. Just in terms of whether hardware and software sales were dissimilar in nature or indeed function to other items that Autonomy was selling, just standing back, there’s a fairly obvious distinction between the nature of hardware and the nature of software?

A. Yes, there is. If you step back and look at those two things, but my understanding is within the context of the business that Autonomy considered it was essentially selling one overall product, and that everything connected with the sale of that product was considered to be the sale of revenue.

The way I think about this when I think about one-segment companies of which there are some, you’re allowed to have one-segment companies, IFRS 8 allows you to do that, is this: the principal thing you’re selling, the other things which are sold you wouldn’t sell but for the principal thing you are selling. That is the distinction which I make in terms of trying to understand why you have a one-segment company.”

This was consistent with the approach in Deloitte’s report in relation to IFRS 8.32 in 2009 (which was approved and signed off by the NAA, as was their report in 2010) which read as follows:

“All of the software solutions provided by Autonomy to its clients are underpinned by the single core IDOL technology. On that basis, management has provided the following analysis of the revenue balance for the group’s single operating segment (Note 4):

(1) Sale of goods

(2) Rendering of services; and

(3) Interest Receivable

We note that sale of goods included all items of software and strategic hardware sold during the year. Rendering of services is the release of the support and maintenance revenue and the provision of professional services to clients.

As outlined above, management tracks all licence and strategic hardware sales as a single body of sales, being the sale of goods. This is consistent with the financial information presented to Mike Lynch and it is the basis on which he makes his resource allocation decisions. Likewise, the deferred revenue release and the professional services rendered are also reported to Mike as a single line item.

On that basis, we deem that management has appropriately disclosed a breakdown of revenue that is consistent with the information presented to the Chief Operating Decision Maker and is that used to produce the group’s financial statements.

It is worth noting that in their Q4 2010 press release, management did provide some representative revenue figures for the following virtual product categories:

-IDOL Product;

-IDOL Cloud;

-IDOL OEM;

-Deferred revenue release; and

-Services.

We note that whilst this information was able to be produced following some detailed analysis performed by management, these are not amounts extracted from the financial information that underpins the preparation of the financial statements. It was derived from a separate analysis purely [performed] for providing some information to analysts on the performance of each virtual product category. It does not represent the way that revenues are analysed out on a regular basis for presentation to Mike Lynch.

We also note that this is just one of several virtual buckets that management use to badge their different product offering to analysts, another being the Protect, Promote and Power families. Again, no separate financial information is maintained on a regular basis to evidence the results for any of those virtual brands.

On that basis we note that the disclosure provided by management is in line with the requirements of IFRS 8.”

Mr Holgate considered that IAS 8 clearly required separate disclosure of revenues from hardware sales (assuming the information was available and material). He accepted that it was a new standard at the time and there had been “some uncertainty as regards the disclosure of operating segments because it was based on a new idea of what was reported internally to the chief operating decision-maker” but he disagreed that IAS 8.32 was a source of difficulty, and described it as “quite straightforward.” He countered the suggestion put to him in cross-examination that the provision lacked a definition of “the idea of products or groups of similar products and services” by the response that “they just have their natural meaning”.

He disagreed with Mr MacGregor’s approach (and thus also with that of Deloitte and the NAA), and considered it (a) to rely on a gloss which was not justified by the words of IAS 8.32, (b) to be extreme and implausible in treating hardware and software as similar products, and (c) to be a view and approach which would not have been taken either by him and his firm or “the vast majority of auditors and the ones I’ve come across and worked with” .

However, Mr Holgate was nevertheless reluctant, when cross-examined directly whether nevertheless a range of views might encompass the decision made by Autonomy and Deloitte, to say that he considered it to be “completely outside the range of what a reasonable accountant could think.” As in the context of IAS 1, his response was that clearly that is not the way he would have accounted for hardware, and he thought that not only his own firm, but many other firms also, would have considered that disclosure was required; but he could not say that it was “outside the range of what a reasonable accountant could think.” That was the same answer as he had given in relation to IAS 1.

Materiality

The two experts had different views as to the materiality test, which seem to me also to reflect their different approaches to the provisions discussed above. These differences can be summarised as being:

(1) In Mr MacGregor’s view, “if you’re not required to disclose it, then you’re not required to disclose it” . Materiality is only a consideration in determining whether what would otherwise be a requirement of disclosure is material or not sufficiently material to warrant it. To sever materiality from the anchor of the specific disclosure requirements would lead to what Mr MacGregor called a “ free for all ”: what is in fact required is to comply with the standards.

(2) In Mr Holgate’s view, this was a surprising and extreme position, which could not be consistent with IAS 1, which is directed towards a fair presentation of the company’s financial position, and could not be right. Material information as to the company’s performance and financial position should be disclosed, if not under specific IFRS provisions then under IAS 1, to achieve fair presentation.

(3) The essential question is the relationship between the general requirement of IAS 1 (which expressly “sets out overall requirements for the presentation of financial statements, guidelines for their structure and minimum requirements for their content ”) and the specific provisions presumed in “virtually all circumstances” (quoting IAS 1.17) to implement those general requirements (albeit with the override that in what Mr MacGregor called “the very, very rarest cases” “ additional disclosure” might remain necessary (despite the presumption) to “result in financial statements that achieve a fair presentation. ”)

(4) Mr Holgate considered that this is a “rarest case”. Mr MacGregor did not, at least if (as he put it in cross-examination) “all the sales…are being designed to do one thing which is to sell IDOL software.”

My assessment

I agree that the essential dispute requires review of the relationship between the general requirement of IAS 1 and the specific provisions presumed in “virtually all circumstances” (quoting IAS 1.17) to implement those general requirements, and the application of the “override”.

I agree with Mr MacGregor that the starting point, and in ordinary circumstances, also the end-point, is what is specifically mandated by the Accounting Standards. Full compliance, mitigated by exception for immateriality, should ordinarily ensure fairness. I agree also with the Defendants that it does not follow from the fact that something might be considered “reasonably useful” that it must be disclosed. There are many aspects of a company’s financial position and financial performance which some investors might wish to know about but of which disclosure is not mandated. The Accounting Standards are intended to determine what is to be disclosed and how disclosure is to be effected: and the ordinary presumption will be that if the information is not required to be disclosed under specific Standards, then even if “reasonably useful” it need not be disclosed.

On the other hand, I agree that there are rare circumstances where the materiality of the matter in question is such that disclosure is required notwithstanding that there are no specific Standards unequivocally mandating it. If the Standards do not appear to require disclosure of a matter which, measured by its likely effect on their assessment of the company if readers of the published information were to be made aware of it, is of plain and obvious materiality, the “override” may apply. In my judgment, information may just as much be “required” to be disclosed for the purposes of s. 90A(3)(b) and (when it came into force in October 2010) Schedule 10A paragraph 3(1) of FSMA to give, overall, a true and fair view of a company’s position and performance as when specific accounting principles or statements of practice expressly require it.

It is, in other words, ultimately a question of degree, to be answered taking into account (a) any judgement as to why specific requirements, which are themselves a useful litmus test of materiality, do not apply, (b) any reasons why disclosure is thought to be unnecessary or commercially unwise, and (c) whether having regard to (a) and (b), disclosure is nonetheless required in order to provide, in the round, a true and fair view.

Sometimes, I would surmise often, there will be a nice balance; in those cases it would to my mind be surprising to conclude that nevertheless disclosure was “required”. I agree with the Defendants’ submission that where there is a range of permissible views on whether a matter is “required to be included”, an issuer will not be held liable for the omission of that matter: and in any event, if there was doubt that would be fatal to a fraud case. In my view, that is especially so where the Standards do not appear specifically to be applicable and the question of any exceptional “override” is being considered.

In the limited category of cases, where the real reason for non-disclosure is a dishonest intention to conceal, the two limbs of this analysis come together. Consciousness of a high degree of materiality in combination with a conscious determination to conceal the relevant facts and matters from published information precisely because of their materiality from the point of view of the reader will likely result in there being (a) a “requirement” by virtue of the “override” and (b) dishonest concealment such as to establish liability.

As to the first hurdle for the Claimants of establishing a disclosure “requirement” under the Standards, Mr Holgate tended initially to give the impression of a rather black and white approach. In the end, however, notwithstanding his own unequivocal views as to the correct interpretation of the Standards, he accepted that neither in the context of IAS 8 nor in the context of IAS 1 was he willing to say that the judgement made was “outside the range of what a reasonable accountant could think” (see paragraph 1728 above). I have considered whether this simply reflects polite unwillingness to express views entailing the conclusion that a respected firm had unanswerably been negligent. I have concluded that although that may have been part of it, his reluctance was primarily based on a realistic appreciation that the points made were arguable.

The lesson suggested by the expert evidence is that there was a range of permissible views on the application of the specific provisions in the Accounting Standards; that would suffice for the Defendants since a statement is not to be regarded as false or misleading where it can be justified by reference to that range of views.

In those circumstances, and in light of Mr Holgate’s concession that the case that no disclosure was required by the Standards was arguable, I agree with the Defendants that this is not the occasion to determine which of the competing views on the application of the various Standards is best as if it were a binary matter and then conclusive; for it must always be borne in mind that this is a case based on the alleged dishonesty of the Defendants. No claim is made in these proceedings against Deloitte, and there is no suggestion that Deloitte were aware of, still less colluded in, any dishonesty. In such a context, even a conclusion that in each case the applicable Accounting Standards were not properly applied, and that correctly construed the separate disclosure of hardware sales was mandated, would not be enough for the Claimants. They still have to prove that (a) the Defendants knew that disclosure of the hardware sales was required and (b) the non-disclosure was the dishonest concealment of a material fact, notwithstanding that in every case Deloitte did approve the accounting treatment and did not advise that disclosure was required.

In other words, the ultimate question in determining whether there has been a failure to disclose a matter required to be disclosed by the Standards is not the meaning of the specific provisions but whether the Defendants knew there was a requirement to include the relevant fact and determined to and did conceal it. If they honestly believed that the views taken and judgements made in relation to each relevant bit of published information was one within the range of acceptable views, having regard to the facts as known at the time, and that therefore disclosure was not required, that establishes a good defence so far as the “requirements” of the Standards are concerned. That is so even if the relevant but undisclosed fact was highly material, though the greater the materiality, the more likely it is that its non-disclosure was deliberate. However, that is not the end of the matter: the question remains whether the non-disclosure (or the form of any disclosure made) amounts to dishonest concealment because the directors knew that disclosure (or a different form of disclosure) was required in order, in the round, to present a true and fair view.

Did the Defendants determine dishonestly to conceal matters they knew to be material?

I turn to address the issue whether, if the Claimants were able to prove one or more omissions of a “matter required to be included” in the published information, the Defendants “knew the omission to be dishonest concealment of a material fact”. I do so on the premise (contrary to my finding) that the purpose given for the hardware reselling was as it was stated to have been by the Defendants. I am aware (and wary) of piling hypothesis on hypothesis.

The Defendants relied in this context on Deloitte’s advice to the effect that there was no “requirement” mandating disclosure, so that disclosure was a commercial judgement for the board of Autonomy, and contended that in such circumstances they honestly believed that there was no relevant concealment. It was stressed on their behalf that:

(1) As noted previously, only proof of actual ‘guilty’ knowledge at the time of a requirement to include the relevant fact would suffice: absent actual knowledge of such a requirement there can be no liability for dishonest concealment.

(2) Dishonesty is the gist of the claim against the Defendants and must be proved, and although in civil proceedings the standard of proof always is that of a balance of probabilities, the seriousness of the allegation means that in the ordinary course, cogent evidence is required in order to overcome the inherent unlikelihood of what is alleged and thus to prove it. Experience shows that in the ordinary course, a benign though curious explanation is often more reliable than a malign though initially plausible one; and furthermore, the seriousness of the consequences of a finding of dishonesty militates against the likelihood that the person implicated took the risk of it.^{^[246]}

(3) It is relevant to consider the advice given to the company and its directors: and where, on the basis of advice, a director is given to understand that it is not requisite to disclose a particular fact in the company’s published information, the omission of that fact is unlikely to amount to a dishonest concealment of it, unless the giver of the advice was materially misled by or to the knowledge of the recipient of it or that recipient knows full well in some way that the advice is wrong.

(4) It is clear from the language of the Schedule that a PDMR must have the relevant guilty knowledge in respect of each false statement or omission alleged to give rise to liability. Paragraph 3(2) states that “ The issuer is liable in respect of an untrue or misleading statement only if a person discharging managerial responsibilities within the issuer knew the statement to be untrue or misleading or was reckless as to whether it was untrue or misleading. ” Liability is only engaged in respect of statements known to be untrue.^{^[247]} If a company’s annual report contains ten misstatements, each of them relied on by a person acquiring the company, but it can only be shown that a PDMR knew about one of those misstatements, the company will only be liable in respect of that one, not the other nine.

The Claimants submitted that nevertheless:

(1) The Defendants’ assertion of an honest belief based on Deloitte’s advice as to the permissibility of not disclosing the hardware sales must also be tested against the background that the Defendants were prepared to tell deliberate lies to the market to ensure that the hardware sales were not detected.

(2) In that regard, the Defendants’ Q&A scripts for the earnings calls contained numerous deceptive answers to questions about Autonomy’s hardware revenues, many of which Dr Lynch now accepts were misleading (or were, at least, capable of being so).

(3) Further, what the Defendants actually said on the Q1 2010 and Q2 2010 earnings calls was flatly untrue: Dr Lynch’s attempt to deny, and put the market off the scent of, Autonomy’s pure hardware sales sits very uneasily with the contention that he genuinely believed on the basis of Deloitte’s advice that it was permissible for Autonomy to remain silent about its hardware sales.

(4) In addition, the Defendants well knew that Deloitte by no means readily blessed the non-disclosure of the hardware sales, even on the false basis as to the purpose of those transactions presented to Deloitte. On the contrary, Deloitte repeatedly advised Autonomy to make such disclosure. That advice was ignored. Gathering this together:

i. On 14 October 2009, following receipt of the Strategic Deals Memorandum, Mr Knights wrote to the Defendants noting that there would be an issue at year-end as to whether the hardware sales needed to be disclosed and that, “ if disclosure did become necessary and in the absence of any previous indication through the year, it would be the first time that this information would be made available to your investor and analyst community. This might be worthy of some consideration at Q3? ”

ii. Deloitte’s report to the Audit Committee for Q3 2009 dated 16 October 2009 noted that hardware sales represented 19% of the total revenues for the quarter and that the Autonomy board “ should consider how best to communicate this new opportunity to the shareholders as these revenues are not driven from the organic IDOL technology of the Group ”. Deloitte added that, if hardware sales in the year were significant, there might be a requirement to disclose them in the year-end financial statements.

iii. Deloitte’s report to the Audit Committee at Q2 2010 advised that “ Given the increasing significance of hardware sales to the Group’s revenues, and the resultant impact on the gross and operating margin in the quarter and half year results we would expect appropriate explanation to be given in the Q2 2010 press release ”. None was given.

iv. Deloitte’s report to the Audit Committee for Q3 2010 again advised that given the “ increasing significance of the hardware sales to the Group’s revenues, and the resultant impact on the gross and operating margin in the quarter and half year results we would expect appropriate explanation to be given in the Q3 2010 press release ”. Again, none was given.

v. Deloitte’s report to the Audit Committee on the 2010 audit stated that “ Given that Autonomy has purchased and on-sold $110 million of hardware during 2010, management now considers that the level of sales being made is equivalent to that of a hardware reseller ”. Deloitte expressed the view that, “ Given the increasing significance of hardware sales to the Group’s revenues, and the resultant impact on the gross and operating margin in the quarter and full year results, we expect appropriate explanation to be given in the 2010 Annual Report ”. None was given.

vi. The Claimants posed the rhetorical question: why were the Defendants so concerned that the market should be kept ignorant of these hardware sales, even if this meant ignoring the views of their own auditors, if there was nothing to hide?

For these reasons, the Claimants contended that the Court should find that the Defendants lacked any honest belief that it was proper for Autonomy to withhold disclosure of the fact, nature and extent of its hardware sales.

Conclusion

My starting and ending point is a strong caveat that my assessment of these opposing contentions is artificial, since I am proceeding on a premise which I have earlier rejected, and building hypothesis on hypothesis leads to a suspect structure.

However, on the hypothetical premises, I would not have been persuaded that the Defendants knew that Deloitte were wrong in their interpretation of the Accounting Standards and their application. The Defendants would have been entitled to rely on Deloitte’s interpretation of the particular provisions of those Standards, except in one regard: the exception in my judgment, relates to what I have called “the override”.

As to the exception, in my view, the “override”, which reflects or echoes the long-established company law requirement that a company’s accounts must present a “true and fair view” , can be taken to be engraved on the heart of every company director (and especially the CEO and CFO of a public listed company). Though it may be informed by the views of accountants, the assessment whether a company’s accounts, in the round, achieve this objective is a judgment for its board of directors by reference to overall fairness, which cannot exclusively be defined or confined by statements of accounting principle, and the assessment of which cannot be delegated. A question as to whether disclosure should be made, and in what form, is a paradigm example of one which cannot exclusively be determined by statements of accounting principle, or by a company’s accountants and auditors: it is ultimately a matter for the judgment of the directors, acting in good faith with a view to providing the information necessary to provide a true and fair view of the company’s position and performance.

In my judgment, the nature and extent of the hardware reselling strategy was information of sufficient materiality that its non-disclosure resulted in Autonomy’s published information in the round being false and misleading: and the Defendants knew that, and were dishonest therefore in concealing it.

The question posed by one of the Reviewers (Mr Robertson) in Q3 2009 (see paragraph 1361(8)(ii) above) rings in the ears:

“What’s the sensitivity about being more transparent on this score? If it’s a strong strategic move for them, why wouldn’t they want to explain this?”

No satisfactory answer was ever provided. As indicated in paragraphs 878 to 888 above, Dr Lynch’s justification for it was stated in his first witness statement as follows:

“HP complains that Autonomy should have provided further disclosure about the nature and extent of its “pure hardware sales”. Autonomy was under no obligation so far as I was aware at the time, to break down its revenue into hardware and software, let alone into different types of hardware. Given the defensive nature of the hardware strategy, we did not particularly want to volunteer this information to the market. We did not want competitors to know that we felt vulnerable by the move to appliances or that we perceived a risk of being removed from strategic supplier lists (and copy our strategy of selling increased volumes to these companies). In addition, we did not want our relationships with one hardware supplier to preclude us from working with another or news of favourable deal terms getting out and being demanded from everyone. The more details we disclosed about our hardware sales, the more we exposed these commercial sensitivities.”

As with most, if not all, of Dr Lynch’s explanations, this carried an initial semblance of reason and reliability: and no doubt non-disclosure did have the advantages identified, especially in Q3 2009. However:

(1) Dr Lynch’s own evidence was that in Q3 2009, the market perceived appliances to be a likely preferred means in the immediate future of delivering software. That was no secret, and nor was it a secret, or otherwise than obvious to anyone interested, that Autonomy had no appliance offering. Why keeping quiet Autonomy’s avowed efforts to develop an appliance jointly with the largest hardware suppliers in the world to remedy this apparent gap in its offering in those circumstances was necessary or appropriate is extremely difficult to explain.

(2) The Defendants could never quite make up their mind whether their case was that the market did know or that it did not know of the hardware sales. Mr Bloomer thought it did, and said that there was no attempt to keep from the market the fact that Autonomy sold hardware at a loss, and he himself discussed this with an institutional investor (Hermes): in which case the commercial sensitivity argument seems further eroded.

(3) The “shift” in the rationale for the hardware reselling strategy in Q1 2010 or Q2 2010, away from appliances or establishing Autonomy as a preferred supplier, and towards a “customer-facing” or “one-stop shop” justification, left much of Dr Lynch’s justification outmoded.

(4) Furthermore, the relevant customers were, according to the Defendants’ own case, existing volume customers of Autonomy in the financial field, further calling into question why there should have been sensitivity in revealing details lest customers use it against Autonomy. There is no evidence or sign to suggest that Autonomy would not have wished to offer the same terms to all valued software customers to incentivise software sales.

Overall, and even (a) on the premise that the purpose of the hardware sales was to drive software sales and (b) accepting that the Court will be very hesitant in second-guessing genuine business judgments made in good faith by those entrusted to make them, the proportion of the revenues derived from hardware seems to me so to have cried out for disclosure as to outweigh any contrary commercial sensitivity in determining the issue of disclosure. It may be remembered that this is broadly what Mr Ariko thought: and he was only deflected by the series of falsities in Dr Lynch’s response to his request for a review of disclosure (see paragraphs 879 to 881 and paragraphs 1395 to 1403 above), including that the market was already aware of hardware sales from what they had been told on earnings calls.

I have taken into account, in deciding that the decision not to disclose was not within the generous ambit of discretion afforded by the court in matters of commercial judgement, the counter-indication that neither the Audit Committee nor Mr Webb considered that disclosure was required either, and there is no suggestion of lack of impropriety on their part in this regard. I have explained Mr Bloomer’s outlook in paragraphs 859 to 862 above, and Mr Webb’s in paragraph 863 above.

In my view, the Audit Committee’s curious indifference or at least passivity in this regard under the chairmanship of Mr Bloomer, and in particular his and their approach in considering the measure of materiality for these purposes to be in the region of 25% of total revenues, appears to me to reflect a mistaken approach. They treated materiality according to whether the programme was of such a size or nature as to upset the segmental analysis (see paragraph 1372 above). But though related that was a different issue. In four of the ten quarters in the Relevant Period more than 10% of revenue comprised loss-making hardware sales. In Q3 2009, 22.8% of revenue was so comprised, and though this dropped in Q4 2009 and Q1 2010 (to 4.1% and 6.1% respectively) after the cessation of Autonomy’s relationship with EMC, it rose again to 14.1% in Q2 2010, 12.7% in Q3 2010 and 12.0% in Q4 2010, before dropping back again to 9.1% in Q1 2011 and 8.1% in Q2 2011. This was, in my judgment, clearly material, and whilst the Audit Committee and Mr Webb appear to have been disoriented by the segmental issue, there is nothing in the evidence to suggest that the Defendants were in the same way.

In any event, even if Mr Bloomer and his colleagues, and it also seems Mr Webb, considered that the hardware reselling strategy was well known to the market, the Defendants knew that it was not. The Defendants’ perspective was different.

At this point the accumulation of hypotheses becomes almost impossible to accommodate. But on the footing that whilst I am to assume a proper purpose for the hardware sales, and no imperative to conceal the purpose, I am not required to accept the commercial reasons given by Dr Lynch, which seem to me to be contrived. No substantial or satisfactory answer to the pertinent question posed by Mr Robertson in Q3 2009 as to what could possibly be the sensitivity about transparency of a strong strategic move having been given, and in light of the Defendants’ consistent and egregious efforts to find reasons or Delphic language to avoid it, I would have been inclined to the conclusion that the Defendants simply had no good reason not to disclose. That would leave a judgment based on a covert and dishonest objective as the only available explanation.

Fortunately, however, I need not and do not base my decision on that ground. On my actual conclusions, the ultimate relevance of the exegesis and assessment is in its confirmation of my judgment that the Defendants lacked any honest belief that it was proper for Autonomy to withhold disclosure of the fact, nature and extent of its hardware sales.

Issue (3): Did Autonomy ’s treatment of the costs of the hardware render Autonomy ’s published information untrue or misleading and did the Defendants appreciate this?

Obviously, if the true purpose of the hardware reselling strategy was to pump up revenue and the purpose put forward to Deloitte of driving software sales was a pretext, the allocation of part of the costs or losses of the programme to sales and marketing expenses, rather than as COGS, was both a further indication and a further instance of impropriety. The experts were agreed on this also. But what if the purpose was as put forward by the Defendants?

The Claimants contended that even then it is plain that the Defendants

“did not genuinely believe that it was proper to treat part of the costs of the hardware that was sold as a sales and marketing expense…that treatment was driven, not by any genuine attempt to apply the accounting standards correctly, but by reverse-engineering to get to a desired gross margin figure.”

The Claimants’ expert, Mr Holgate, accepted that there was no IFRS guidance on what a company should include in COGS or Sales and marketing but that was because it was so obvious that it does not require explanation. He considered and opined that it was obvious that it is:

“clearly contrary to IFRS and wrong in a wider sense to account for part of COGS as sales and marketing expense…

…

Making sales at low margins, or even at losses, has no impact on the simple principle that the costs of goods sold is shown as COGS and the costs of sales and marketing activities are shown as sales and marketing expenses. The effect of a reduction in sales price is exactly that: revenues are reduced.”

Thus, according to the Claimants, it was plain and obvious that Autonomy was wrongly and misleadingly treating all the revenue from ‘pure’ hardware sales as revenue, but less than all of the costs of the product that it sold as a cost of that revenue. Autonomy’s published information was untrue or misleading for the further reason that the allocation to sales and marketing of any part of the costs of the hardware reselling strategy resulted in the figures for COGS, gross profit, gross margin, and sales and marketing expenses in the Annual Reports and the Quarterly Reports from Q3 2009 to Q2 2011 all being false.

The Defendants did not accept this. They contended that within the price paid for the hardware was an element of marketing expense; and there was nothing in the IFRS or otherwise to prevent that element being allocated in accordance with what they presented to be its true nature. Further, all of the various allocations of costs to sales and marketing were closely monitored by Deloitte (including review by the Reviewers), who signed off on them in every quarter in the Relevant Period as representing a reasonable allocation having regard to the marketing purpose.

The Audit Committee likewise were content with the final figures agreed. Mr Bloomer explained in his witness statement:

“My view was that it was not unreasonable for losses on hardware sales to be categorised to sales and marketing when the benefit generated from having incurred these losses would arrive over time through further sales.”

Mr MacGregor supported this line, elaborating when cross-examined that he could see nothing wrong with treating as a sales and marketing expense “the loss which has been taken because it’s an incentive to sell the overall software product…”

Given my previous conclusions, it is sufficient to state my views on this area of the dispute in more summary form:

(1) Mr Holgate’s approach tended to overlook the Defendants’ basic argument that part of the amounts paid by Autonomy to the hardware supplier concerned were referable to, and properly treated as, a premium to fund marketing.

(2) To the extent that the payment of such a premium could be shown on the facts, I am not persuaded (on the hypothesis that the hardware reselling was for the purposes asserted by the Defendants) that it would be wrong to account for it, or some part of it, as a marketing expense.

(3) Whether that was sufficiently demonstrated by Autonomy in every or any case is a question of fact.

(4) In point of fact, I have concluded previously (see paragraphs 1335 to 1340 above) that the “residual method” was adopted in respect of the amounts paid to EMC in Q3 2009 as an expedient proxy because EMC was not prepared to acknowledge that any part of the purchase price paid by Autonomy for the hardware it resold was referable to development or marketing costs. This was not an appropriate basis on which to allocate to Sales and marketing a sizeable proportion of the costs.

(5) Similarly, and notwithstanding the provisions unilaterally inserted into the purchase orders by Autonomy to bolster an argument that part of the purchase price of hardware supplied by Dell was referable to marketing or development (see paragraphs 940 to 941 above), neither Dell nor, for that matter, Hitachi, ever agreed to part of the purchase price being committed in that way, and there was no sufficient basis in fact for allocating any substantial part of the costs to Sales and marketing. Indeed, that was the decision of Deloitte, subject to a concession in respect of losses (10%) and to Mr Knights’ “solution” (see paragraph 1495(2) above).

(6) I accept the Claimants’ contention that the amount of costs which Autonomy proposed should be allocated to Sales and marketing was dictated by (a) management’s need to reduce the impact of the loss making hardware reselling strategy on gross margins (and thereby, in effect, manipulate the figures) and (b) what Deloitte could be persuaded to accept. Internal emails demonstrated this: the examples given by the Claimants being (a) an email from Mr Hussain to Mr Chamberlain dated 11 October 2009, asking Mr Chamberlain to “see if you can allocate $4m from cogs to opex” because “need GM [i.e. gross margin] to be at the minimum 86%” and (b) an email dated 18 April 2010 from Mr Hussain to Dr Lynch telling him “need to relocate some costs from cogs to opex”.

(7) After Deloitte had confined the sales and marketing expenses allocation to losses (and thus some 10% of costs) the amounts became relatively immaterial: certainly, that was the view taken by Mr Bloomer and his colleagues on the Audit Committee.

With reference to (7) above, Mr Miles submitted in his oral closing submissions that even if in Q3 2009 the amount allocated was (a) not properly substantiated (b) disproportionate in any event and (c) calculated by the process of reverse engineering suggested by the Claimants, the Claimants did not make their decision by reference to Q3 2009, and by 2010 the impact on the accounts had been much reduced because Deloitte had insisted that only losses could be allocated, with the effect that 90% of the costs were allocated to COGS. The impact on gross margin of about 1% was immaterial. According to the argument, therefore, Autonomy’s accounts in 2010 and 2011 were not rendered materially untrue or misleading even if the losses from hardware sales were wrongly allocated to sales and marketing: there was no material deficiency in disclosure.

This argument, which was not in Dr Lynch’s written closing submissions, was not specifically addressed by Mr Rabinowitz in his oral reply. Had it been decisive, I would have required additional assistance, since I have not found it an easy point. The difficulty, as I see it, revolves around the separate but associated issues of inducement and causation, and the difficulty of determining what the effect would have been on subsequent accounts if in the Q3 2009 and the 2009 Year-end accounts substantially all the costs of the hardware reselling strategy had been allocated to COGS as I have concluded they should have been. I am disposed to think that the presentation would have altered and that Mr Miles’ argument would not be conclusive against the Claimants’ case. That seems to me to be supported by the evidence of Mr Gersh in cross-examination to the effect that had HP appreciated prior to the acquisition that part of the costs of hardware or appliance sales had been allocated to sales and marketing expenses, it would have been in a better position to assess the volume of such sales. But I prefer not finally to determine the point.

Lastly under this heading, and in case the matter were to become relevant in the future, I confirm my view that contrary to the Defendants’ case, Dr Lynch was involved from the outset (Q3 2009) in the issue of cost allocation.

Issue (4): Should Autonomy, at the least, have made clear in its published information what its accounting policy was with respect to hardware costs?

I can also be relatively brief in summarising my view in respect of the further issue about the extent to which (if any) it was necessary for Autonomy to make clear in its published information what its accounting policies were with respect to its hardware cost accounting.

Mr Holgate pointed out in his first report that the only disclosure given by Autonomy in its 2009 and 2010 consolidated financial statements (note 2(h)) was:

“sales and marketing costs comprise the costs of the sales force, commissions and costs of promoting new products and entering into new markets.”

He noted that this policy for sales and marketing costs did not refer to hardware or hardware reselling, though in 2010 (only) the policy for cost of revenue did refer to hardware in passing:

“Cost of revenues: Cost of licence revenues include the cost of royalties due to third party licenses, costs of product media, product duplication, hardware and manuals.”

Mr Holgate suggested that this clearly gave the incorrect impression that all hardware costs were included in costs of revenue and none was included in sales and marketing expenses.

He cited various IFRS requirements in relation to disclosure of accounting policies as follows:

(1) IAS 1’s requirement that:

“An entity shall disclose in the summary of significant accounting policies: (a) the measurement basis (or bases) used in preparing the financial statements; and (b) the other accounting policies used that are relevant to an understanding of the financial statements” [Mr Holgate’s underlining]

(2) A further requirement in IAS 1 with specific reference to judgements made that:

“An entity shall disclose, in the summary of significant accounting policies or other notes, the judgements, apart from those involving estimations…that management has made in the process of applying the entity’s accounting policies and that have the most significant effect on the amount recognised in the financial statements.” [Again, Mr Holgate’s underlining]

In Mr Holgate’s opinion:

“These requirements clearly point to a need for Autonomy to have disclosed that part of COGS (i.e. the costs of the hardware) had been accounted for as sales and marketing expenses. It would have been ‘relevant to an understanding of the financial statements’ for a user to have known that gross profit had been increased by moving part of COGS to a heading (sales and marketing expenses) that was presented below gross profit. Likewise, ‘disclosure would have assisted users in understanding’ the accounting treatment and the underlying business undertaken and margins achieved. In addition, disclosure of the accounting policy would assist users in understanding the ‘judgement’ that ‘management [had] made…Indeed, adding wording dealing with the inclusion of part of COGS in sales and marketing expenses would have been much more important than the disclosure that Autonomy actually gave. This is because the sales and marketing policy wording in the notes to the 2009 and 2010 annual financial statements…was both compliant and reasonably obvious (and therefore arguably unnecessary); whereas the inclusion of part of the hardware costs in sales and marketing expenses was both non-compliant and unexpected, and therefore disclosure was necessary.”

Mr MacGregor did not agree with this, and explained that the requirement was to disclose the principal policies and that it was not necessary to present policies in the granular way implicit in Mr Holgate’s approach. Just as the general presentation of information in financial statements requires the application of judgement and selection, so too does the statement of relevant accounting policies.

I would agree with Mr MacGregor’s general approach that it is not necessary to set out at great length every element of the revenues and costs of the company and its assets and liabilities that have been accounted for: some judgement is required. However, I would agree with Mr Holgate that in this case the allocation at least in 2009 was sufficiently significant in its effect that I would have expected it to have been explained.

The question whether it was dishonest not to have done so is more difficult. Deloitte were content and so were the Audit Committee. That was necessary but not sufficient; but it seems to me that there would be real difficulty in establishing dishonesty if there was no improper purpose driving the decision.

Issue (5): Did Autonomy wrongly recognise revenue in Q2 2009 on a specific ($6 million) hardware transaction with Morgan Stanley?

A discrete ‘hardware’ claim also advanced by the Claimants was that Autonomy wrongly recognised revenues of $6 million on a sale of hardware by Autonomy Inc to Morgan Stanley under an agreement entered into on 30 June 2009.

The nub of the claim is that the hardware was recognised on the premise that it was sold to Morgan Stanley and dispatched by EMC in Q2 2009 whereas, according to the Claimants, it was in fact dispatched by Hitachi pursuant to Au tonomy’s one-time reseller agreement with HDS entered into with Autonomy on 27 June 2009 with delivery not occurring until inside Q3 2009.

The Claimants’ case is that this revenue was wrongly recognised in Q2 2009, “to the knowledge of at least Mr Hussain” , because to satisfy the requirements of IAS 18.14 for the recognition of revenue the hardware had to have been delivered by 30 June 2009 and was not, though Mr Hussain pretended to Deloitte that it had been, and Autonomy’s management made a like (false) representation in a management representation letter dated 15 July 2009 signed by Mr Hussain.

The formula “to the knowledge of at least Mr Hussain” was an opaque way of saying, as was the fact, that Dr Lynch was not alleged to have played any part and was not cross-examined on the matter. Mr Rabinowitz expressly confirmed in his oral closing submissions that no claim could be made against Dr Lynch accordingly.

The Claimants acknowledged that “this episode might be said to be a comparatively minor one in pure revenue terms” , but submitted that “it speaks volumes about the willingness of Autonomy management, and in particular Mr Hussain, to lie to Deloitte to achieve their ends.”

They relied on the issue also in support of a contention that Mr MacGregor was less than objective in his support for the Defendants’ analysis, and produced an “exotic thesis…dependent on a number of hypothetical facts” none of which the Defendants even attempted to prove at trial.

The Defendants sought to dismiss this claim as a ‘cul-de-sac’, on the basis that it is not possible to see how the revenue recognition occurring a quarter earlier has any bearing on HP’s acquisition of Autonomy.

Further, Dr Lynch submitted (and Mr Hussain adopted this) that the recognition was correct on the basis that it was possible for title to pass without delivery under the contract for the reasons given by Mr MacGregor in his supplemental report.

The full factual detail is complex and convoluted; but in summary, the underlying facts can be distilled as follows:

(1) On 30 June 2009 Morgan Stanley agreed that it would purchase Hitachi hardware from Autonomy unless the parties agreed that Autonomy could supply hardware from a different manufacturer.

(2) Pursuant to that agreement, Autonomy entered into a contract with Hitachi for the purchase of hardware for supply to Morgan Stanley.

(3) Autonomy did indeed acquire $6 million of hardware from Hitachi which was, in fact, supplied to Morgan Stanley in or after August 2009, i.e., in Q3 2009. Title to the hardware that Hitachi sold to Autonomy for onward sale to Morgan Stanley did not pass even to Autonomy (still less from Autonomy to Morgan Stanley) any earlier than August 2009, after the end of Q2 2009.

(4) Autonomy management told Deloitte that once it was clear that Hitachi could not deliver the hardware on or before 30 June 2009, Autonomy had entered into an agreement with EMC dated 29 June 2009 under which EMC was to provide and deliver $5 million worth of hardware for delivery to Morgan Stanley by 30 June 2009 and Autonomy itself agreed to supply $1 million worth of hardware from its own reserves of new hardware stock.

(5) Deloitte asked Autonomy management for evidence that the EMC hardware had indeed been delivered by 30 June 2009.

(6) On 13 July 2009, Mr Chamberlain, copying Mr Hussain, told Deloitte that Autonomy was trying to get the relevant invoices from EMC, which should show the delivery date. Later, Mr Hussain and Mr Chamberlain told Deloitte that EMC had been unable to supply any documentation. Deloitte then agreed to rely on a written representation from Autonomy management to the effect that the hardware had been physically despatched by 30 June 2009.

(7) That representation was contained in a management representation letter dated 15 July 2009 signed by Mr Hussain, for and on behalf of Autonomy, which stated as follows:

“ We are satisfied that the acquisition of $9.0 million of hardware from EMC Corporation was an arms length commercial transaction. Additionally, we confirm that the despatch of hardware with a purchase price of $5.0 million was made from EMC Corporation on the 30 June 2009.”

(8) There were EMC invoices dated 30 June 2009 for the hardware that EMC had sold Autonomy. However, they were headed “ PROFORMA INVOICE ” and stated “ Invoices will be revised once order is shipped ”. Thus, those invoices made clear on their face that the hardware had not yet been shipped as at 30 June 2009, the last possible day for delivery within Q2 2009.

(9) On 13 July 2009 , Mr Hussain pointed out to Mr Chamberlain and Mr Sullivan that the invoices were “ clearly proforma ” and asked them to “ work out what you need at a minimum for the auditors tonight ” .

(10) Mr Chamberlain responded the same day saying that he needed an email from EMC confirming that the hardware had been shipped to Autonomy on 30 June 2009. He asked whether this was “ something we can get ”.

Mr Sullivan responded, “ No. Give me a call now please … ” . Mr Sullivan explained in his testimony at Mr Hussain’s criminal trial:

“… the reason I said “no” is because I knew that the - or I thought that the product had not shipped only because we signed the product - we signed the contract on, you know, the last hours of the quarter. So I don’t believe EMC would have even had time to ship it.”

(11) On 13 July 2009, Mr Sullivan asked EMC to remove the language “ Invoices will be revised once order has shipped ” from the EMC invoices. EMC did so and, on 14 July 2009, Mr Sullivan provided Mr Hussain and Mr Chamberlain with the revised form of invoices. The revised invoices obscured the fact that the EMC hardware was not delivered by 30 June 2009, but they did not provide positive evidence that delivery had taken place by that date. It is unclear whether the revised invoices were provided to Deloitte. At all events, the absence of positive evidence of delivery by 30 June 2009 resulted in Deloitte requiring the management representation mentioned above.

(12) The truth is that, as Mr Hussain well knew, the EMC hardware was only despatched by EMC to Autonomy in July 2009, and so after the end of the quarter. The as-despatched invoices show, with one minor exception (an invoice for $414,578 plus tax), this to be the case. They also show that the EMC hardware was delivered to Zantaz locations, not Morgan Stanley.

(13) The fact that the EMC hardware was to be utilised by Zantaz, not Morgan Stanley, can also be seen from other contemporaneous documents.

(14) In his evidence to the US court, Mr Sullivan confirmed that he never intended the EMC hardware to be sold to Morgan Stanley and nor would it have made sense to do so; nor did he ever discuss with EMC the prospect of selling its hardware to Morgan Stanley, which could not have been done without EMC’s permission; nor had he discussed with Mr Hussain the prospect of selling EMC hardware to Morgan Stanley.

(15) On 30 June 2009, Mr Mooney informed Dr Lynch and Mr Hussain that the “ final docs ” had been received from EMC for signature. Later that day, Dr Lynch approved the purchase order, Mr Hussain having already done so. The “ Purchase Order Request ” recorded that it related to “ EMC Storage Solution order for WW Hosted Introspect & Digital Safe platforms ” at a cost of $9 million.

(16) The hardware that was actually delivered to Morgan Stanley was not EMC hardware at all, but Hitachi hardware. This accorded wit h the Morgan Stanley purchase order dated 1 July 2009 which expressly specified delivery of Hitachi hardware.

(17) The Hitachi hardware was not delivered to Morgan Stanley until August/September 2009.

(18) On 24 September 2009, Hitachi confirmed to Autonomy that the Morgan Stanley hardware orders had been shipped and delivered.

Mr Welham ’ s unchallenged evidence was that if the hardware, which was Hitachi rather than EMC hardware, was despatched and delivered to Morgan Stanley after 30 June 2009, that is, in Q3 2009 rather than Q2 2009, then it is highly unlikely that Deloitte would have agreed with the recognition by Autonomy of $6 million of revenue in Q2 2009. He further explained that these facts, if discovered by Deloitte, would also have given rise to broader concerns regarding management ’ s honesty and integrity.

Mr Holgate ’ s opinion was that, if the Hitachi hardware was despatched to Morgan Stanley in August/September 2009, then it is “ clear ” that revenue on the transaction should not have been recognised in Q2 2009.

Mr MacGregor accepted that ordinarily the date of delivery of goods is, absent particular factors that point against it, a strong indication of the date on which revenue should be recognised. He made two caveats: (a) first, he emphasised that under “ International GAAP it ’ s perfectly possible for the risks and rewards to pass without delivery taking place…and the key issue is whether title has passed” ; and (b) secondly, that this was only a general rule and was subject to unusual complications or features, and “ this did have some unusual features”.

In th is case, Mr MacGregor suggested, the first unusual feature was that, in his view, the Morgan Stanley sale may have qualified as a ‘ bill and ho ld’ sale under which it is presumed that delivery is delayed at the buyer ’ s request but the buyer nevertheless accepts or takes title and accepts the billing, and risk and title passes even though the goods are not yet shipped. He drew attention (in his supplemental report) to IAS 18 Appendix, which gives various illustrative examples of risk and rewards of ownership passing at different times, depending on the circumstances; and he particularly highlighted the first example, being of a ‘ bill and hold ’ sale. IAS 18 Appendix 1E § 1 notes as regards such a sale that revenue is to be recognised when the buyer takes title, provided:

(1) it is probable that delivery will be made;

(2) the item is on hand, identified and ready for delivery to the buyer at the time that the sale is recognised;

(3) the buyer specifically acknowledges the deferred delivery instructions; and

(4) the usual payment terms apply.

Mr MacGregor then put forward a complex set of assumptions leading to the possibility that Morgan Stanley - by way of a possible ‘bill and hold’ arrangement or something analogous - effectively took title to the hardware that Autonomy purchased from EMC for its (Autonomy’s) own use, even though (as he conceded) the EMC hardware was not physically delivered to Autonomy until Q3 2009 (and was never delivered to Morgan Stanley). These assumptions were principally that (i) notwithstanding that Morgan Stanley contracted with Autonomy to purchase Hitachi hardware, an agreement was reached that Autonomy would instead supply them with EMC hardware, and (ii) Morgan Stanley had agreed to take title to the EMC hardware in Q2 2009 so as to give rise to the ‘bill and hold’ arrangement which Mr MacGregor posited.

As the Claimants pointed out, however, the Defendants did not even attempt to prove the factual assumptions made by Mr MacGregor; and I accept the Claimants’ submission that the contemporaneous documents, and Mr Sullivan’s testimony to the US court, did not support them.

Without descending further into the convoluted detail, I accept that the documents show that Morgan Stanley agreed to make payment to Autonomy on 2 July 2009 in circumstances where they had been told by Autonomy’s Cynthia Watkins (i) that Autonomy had “ received the PO ”, which was Morgan Stanley’s purchase order which expressly specified delivery of Hitachi hardware (see paragraph 1785 above), and (ii) that Morgan Stanley’s “ order is assembled in our Pleasanton facility ”. It was on these two bases that Ms Watkins said that “ payment can be issued as we have fulfilled the order ” and Morgan Stanley’s Kathy Macedonio confirmed that “ The order has been received in so payment can be released ”.

It is thus plain that Morgan Stanley only agreed on 2 July 2009 (i.e. in Q3 2009) to make payment because they understood (wrongly, but through no fault of their own) that the hardware that they had ordered - i.e. Hitachi hardware - had been delivered to Autonomy’s Pleasanton facility; not because they had agreed to take title to the EMC hardware by way of some ‘bill and hold’ arrangement.

I also accept the Claimants’ submission that IAS 18 Appendix IE §1 only relates to ‘bill and hold’ sales “in which delivery is delayed at the buyer’s request but the buyer takes title and accepts billing”. It is plain that Morgan Stanley did not request that delivery be delayed. Mr MacGregor seems to suggest that the same principles should apply “by analogy” where delivery is not delayed at the request of the buyer, but is instead prompted by the seller. That is not what happened in this case. But even were the principles in IAS 18 Appendix IE §1 to stretch that far - and I agree with the Claimants that it is illogical to imagine that a seller could still recognise revenue even though it was not in a position to effect delivery to the buyer - it would still need to be shown that Morgan Stanley (a) agreed to take and (b) did take title to the EMC hardware. Again, there is no factual basis to support such a conclusion.

In any event, and as the Claimants also emphasised , Mr MacGregor’s analysis is wholly removed from what Deloitte understood to be the position. It is plain that the view Deloitte took at the time was that, in order for revenue to be recognised in Q2 2009, they (Deloitte) needed to be satisfied that the hardware had actually been physically delivered to Morgan Stanley by 30 June 2009 - see (i) Mr Welham’s unchallenged evidence (see paragraph 1786 above); (ii) Mr Welham’s email dated 15 July 2009: “ I have talked to Steve [Chamberlain] re the EMC delivery point ”; and (iii) Deloitte’s file note: “ We have also examined correspondence between Autonomy and Morgan Stanley over the completion of the order and Morgan Stanley’s acceptance and beginning of the payment process as evidence of delivery of the hardware (and hence that shipment from EMC has taken place) ”. As can be seen from Mr Chamberlain’s email to Deloitte of 13 July 2009, Autonomy understood that Deloitte’s focus was on delivery: “ we are trying to get invoices from EMC which should come in overnight. These should show delivery date ”.

In short: Mr MacGregor’s analysis was ingenious but unsupported by the facts, and inconsistent with Deloitte’s approach at the time. Regrettably it seems to me that his efforts went beyond what was appropriate and became in this context, unrealistic and partisan. I agree with the Claimants’ conclusion that the simpler truth is that the revenue was wrongly recognised by Autonomy in Q2 2009.

There was no dispute as to Mr Hussain’s involvement: I find that he knew that there was no proper basis for recognising revenue from the transaction in Q2 2009.

As to the position of Dr Lynch, having fairly recognised and accepted that the Claimants could not, in those circumstances, allege ‘guilty knowledge’ on his part, Mr Rabinowitz submitted only that the episode was a further example of Autonomy not being the “bright, cleaner than clean company that did everything by the book” that Dr Lynch had depicted. That is so.

Issue (6): Did HP know about the hardware sales pre-acquisition and continue them after it?

The Claimants’ case is that prior to the acquisition, the only hardware sales that HP knew about were appliance sales, and it did not know about the loss making “pure hardware” sales. HP’s 20 November 2012 announcement of the write-down stated that the hardware sales were a recent discovery, allegedly identified after a whistle-blower came forward after the departure of Dr Lynch. That was also the position adopted by Ms Lesjak in her witness statement: she said that it was not until after Dr Lynch had left and Mr Scott had come forward with his allegations that she came to learn that Autonomy had been reselling third party hardware at a loss.

The Defendants’ case, which is of course primarily relevant to the issues of inducement (and, in the misrepresentation claim, reliance) and loss is that in truth, HP knew before the acquisition that Autonomy engaged in substantial hardware sales. Further, HP was certainly aware of the nature and amount of the hardware sales from shortly after the acquisition, but expressed no surprise, and did nothing about it, suggesting (according to the Defendants) that at the time Autonomy’s hardware sales were not a real concern for HP, contrary to the picture that is now being painted, and that it would have made no material difference to the acquisition or its terms if HP had had detailed knowledge of the hardware sales in advance.

The Claimants rejected this. They contended that there was nothing in the Defendants’ assertion that HP knew about Autonomy’s hardware sales (apart from appliance sales) pre-acquisition, and nor did HP become aware of the sales prior to May 2012.

Pre-acquisition knowledge and due diligence

The Defendants submitted that even before the acquisition was announced, HP must have known that Autonomy sold hardware; and that certainly “ it knew enough before the acquisition to make further inquiries and chose not to”, suggesting lack of concern. The Defendants’ punchline, for this as for their case that HP were certainly aware of Autonomy’s hardware reselling after the acquisition (see below), was that “Nobody thought they were buying a different company.” The issue is of most relevance to reliance (and quantum).

The Defendants supported these submissions principally on the basis of: (a) what they asserted was an industry understanding that most enterprise software companies do sell hardware ; (b) the fact that HP itself sold substantial quantities of hardware to Autonomy; (c) express references to “hardware sales, including appliances” in the course of due diligence and (d) emails sent by Mr Gersh of KPMG (acting for HP in the due diligence exercise) from which it was said by the Defendants to be clear that “he was aware that Autonomy was selling self-standing hardware” and yet “KPMG never asked Autonomy how much hardware it was selling”.

There was support for the suggestion that hardware sales by software companies were commonplace in the evidence of Mr Apotheker and also in the evidence of Mr Sullivan:

(1) Mr Apotheker, who had long experience in the industry, appreciated that hardware sales of 5 to 7% of a software company’s business “ could be quite normal for somebody who doesn’t do appliances .” (For one that produced appliances, the figure would be higher.) Thus, he recognised that software companies would usually sell some non -appliance hardware, that is, in the Claimants’ terms, “pure hardware”. Moreover, the implication is that for a software company which also sold appliances, hardware sales would be greater than 5 to 7%. He also recognised that hardware sales would often be loss making, and that this was the reason why software companies sought to avoid selling ‘pure’ hardware:

“If only because, for obvious reasons, hardware, in particular when you resell it, you will probably make a loss. At least you won’t make a profit. Or the best situation, even if you make a profit, you will significantly dilute your margins”.

(2) Mr Sullivan gave evidence at the US trial against Mr Hussain that hardware sales were common in the software industry, and that he had sold hardware to customers both at Zantaz, and also at Steelpoint, the software company at which he had previously worked. He also stated that Autonomy had made substantial sales of hardware at a profit, to Citigroup and Paribas.

There was other evidence to like effect. Mr Philip Pearson (“Mr Pearson”) “ assumed that as with any enterprise software business, [Autonomy’s] business included additional revenue streams which typically consisted of support, services and hardware ” (emphasis added); and he believed that this was common knowledge. He explained that most enterprise software firms would take on some hardware, services and support work. He was not challenged on this evidence. Furthermore, HP sold not only to Autonomy but to a wide range of enterprise software companies.

Obviously, I would accept that many enterprise software companies sell hardware; and that this would be generally expected within the limits broadly indicated by Mr Apotheker. But neither side explored, and certainly it was not clear to me, how other enterprise software companies accounted for these sales, nor what disclosure they made in their accounts; or whether they were single or multi-segmental businesses. If other hardware selling enterprise software companies generally accounted separately for such sales, or made disclosure of them, that might have rebounded on the Defendants. I do not make assumptions either way; but I do not extract much assistance from the point.

The fact that HP sold substantial quantities of hardware to Autonomy and yet did not focus on it is, at first blush, surprising. But it seems to me to be more a forensic point than one of real substance. HP was an enormous organisation, which on the Defendants’ own case was very disorganised, and divided into ‘silos’ with very poor communication between them. It should not be assumed that one department had the knowledge of another, or even ready access to it. I agree that in an £11 billion acquisition one might have expected the deal team to have made enquiries; and that this is particularly so where, as here, the buyer and the target were operating in closely related sectors, and the buyer was hoping to achieve synergies from operating as a combined business. But the question is whether those making the decision to proceed with the acquisition actually knew; and there is no evidence that they did so. The Defendants’ written closing submissions noted that it was “unclear what checks HP actually carried out in its initial records of sales to Autonomy.”

In my judgment, the Defendants overplayed the facts said to support the submission that there was discussion of Autonomy’s hardware sales in the course of due diligence and that the fact that neither HP nor KPMG displayed much interest or asked any questions in relation to them suggests that the sales were not of significance for HP in the deal. As to that:

(1) Mr Sarin recalled an exchange with Mr Hussain in relation to appliance sales; and the Defendants relied on the fact that he did not seek to find out how large those sales were, or even to apply his mind to that question: he said that he did not think about what the number for such sales might be. However, that does not reflect that Mr Sarin’s evidence was that what he had been told, he “would think” by Mr Hussain, was that (a) “sometimes customer wanted the software on an appliance or a box” and that would be shipped by Autonomy to the customer and (b) although in that way “appliances were sometimes sold to customers….it was a small part of their business”. Further, when asked “what level of appliance sales” he thought Autonomy was making, his reply was “a very small number” . He was asked to quantify that in numerical terms; but reiterated that he was “comfortable that it was a very small number” and that the demand was from “a small handful of customers” adding when pressed again that he thought it “very minimal”. Whilst more generally I formed reservations about Mr Sarin’s evidence, I found his answers credible on this point.

(2) The Defendants also placed reliance on a list of questions prepared by KPMG which Mr Sarin received on 8 August 2011. The Defendants submitted that the contracts referred to in the list “included contracts in part for the sale of hardware, as the list of questions made clear.” However, all that was included was (a) a contract for “hardware and software development” (as regards which the question was how revenue was recognised, whether on completion or as and when delivered); (b) a contract which was stated to include “implementation of networking, hardware and infrastructure procurements and deploying hardware set up and configuration” (as regards which the question was how services, including professional services, were recorded). In my view, only (if at all) with hindsight does that appear to suggest separate sales of hardware.

(3) The Defendants suggested that Mr Gersh was very reluctant in cross-examination to accept that he had been aware that Autonomy sold any hardware, even in the context of appliance sales: they described as “close to nit-picking” his position that appliance sales did not amount to sales of hardware, but rather to sales of software, the method of delivery for which was hardware. They went on to suggest that he ultimately accepted that he had known that Autonomy was selling some hardware, though they accepted that he continued to insist this was only as a component of, or at least in conjunction with, a software sale. But my understanding is that Mr Gersh was explaining that as matter of accounting:

“…the appliance is a software sale. It ’ s just the appliance refers to the method of delivery of the software and because it ’ s a sale, the cost of the sale is the hardware, which we understood to be in the cost of sales line.”

(4) His later answers, when pressed again, seem to me to show how very different was KPMG’s understanding than it is now suggested by the Defendants they had (and, indeed, to illustrate how effective was the disguise that Autonomy devised):

“ The client is buying...the client is buying software…because it ’ s buying the product which Autonomy makes, which is IDOL. It ’ s not buying - or at the time when we were doing due diligence, Autonomy is presented as a software company and it sells software. And so what are customers buying from Autonomy? We understood they ’ re buying software and in this case Autonomy is - in some cases it is selling its software already pre-loaded to a server, which they refer to as an “ appliance ”.”

(5) Mr Gersh added later that:

“the commentary around the appliance sales was that it was insignificant and didn’t change the margin profile, which to us meant that there was an immaterial amount of hardware being sold as part of the appliance product. So it was more material items in the revenue that we discussed at length”.

(6) The Defendants then contended that in fact, Mr Gersh and his team must have known from their review of contracts that Autonomy sold hardware, and that this was not restricted to appliances as HP uses the term in these proceedings. Mr Gersh confirmed that he read all of the contracts in the data room. These included a contract providing that the customer would be “ entitled at any time during the term of this Schedule to order Hardware set out in the Hardware Annex ”; and another contract to which was annexed a purchase order for 48 storage cells together with maintenance and support. In each case, Mr Gersh claimed that he had assumed these were references to “ a solution involving Autonomy’s product ” or to an appliance. He gave the following evidence:

“A . I saw contracts which had -- three contracts which described hardware which we assumed were appliances.

Q. So you assumed that they were appliances, but they didn't say on the contracts that they were appliances? It was your assumption, correct?

A. That's correct.”

(7) The Defendants claimed that there was no basis for such an assumption in the documents, and KPMG, as a firm carrying out due diligence, should not have contented itself with assumptions: its job was to probe and ask questions, and to consider critically the documents it reviewed. They contended further that it is likely that he did appreciate from these documents that Autonomy sold some non-appliance hardware, and likely also that he did not think it would be a matter that would have caused concerns for HP, since if he had regarded it as an important point, he would have asked further questions. However, that seems to me to overlook Mr Gersh’s evidence in his witness statement (which was not challenged, and which I accept) about the subject-matter and nature of the (redacted) contracts which Mr Gersh saw in the Data Room:

“ I recall that of the three redacted contracts referencing hardware that we saw in the Data Room, two seemed to me to meet Autonomy ’ s description of an appliance. The third, which I surmised was a contract with UBS, contemplated a large package solution^{^[248]} providing software, services and hardware for storage, with the core element being Autonomy software. The commercial components of this contract were difficult to break down. Although it appeared that hardware might be delivered separately and at different times from the software, we presumed it had been purchased from a hardware distributor (or, less likely, a manufacturer) and “ passed through” to the customer as part of an overall solution being provided to UBS.”

(8) Nevertheless, the Defendants then submitted that it is clear from an email sent by Mr Gersh on 8 August 2011 that he was aware that Autonomy was selling self-standing hardware. Commenting on analysis performed by one of his colleagues at KPMG on the contracts in the data room he said the work needed to be reviewed, amongst other things because “ as far as I can tell they have not captured free software or hardware pass-through. ” As apparent from the quotation in the preceding sub-paragraph, Mr Gersh told me that (on the contrary) “ hardware pass-through ” was a reference to appliance sales. The Defendants submitted that this cannot be right because of the way Mr Gersh used the phrase in another context. In his witness statement, he discussed the discovery in October 2011 of a $41m payable to Dell, which he considered too large to relate to just appliance sales. He said that “ I remember that what we collectively had heard from Mr. Chamberlain was that the payables to Dell related to pass-through hardware sales Autonomy had made and that the sales had been recorded on a “gross” basis ”. (They referred me also to how one of Mr Gersh’s colleagues used the term “ pass-through hardware ”, submitting that it was plainly not a reference to appliances as the Claimants have used the term in these proceedings.) In my view, this exercise in semantic comparison was misplaced. It is plain that the phrase was used in different senses in different contexts; for example, the Claimants referred to its use by Mr Briest (an analyst) in a research note dated 23 July 2010 in a context which plainly was intended to signify appliance sales, and was understood in that sense by, for example, Mr Shelley. The question is whether Mr Gersh’s evidence of what he meant by it in the particular context is to be rejected: and I do not consider it should be. It is entirely consistent with his explanations as described above.

In summary, I have concluded, and find, that neither Mr Sarin nor Mr Gersh knew, or on the evidence before me, had reason to know or further enquire about the pure hardware sales.

Dr Lynch’s evidence was that after the acquisition had been announced, but before it had been completed, he discussed Autonomy’s hardware sales with Mr Robison, and that they specifically considered together whether Autonomy should continue to sell Dell and EMC hardware in the run-up to the acquisition. Dr Lynch also said that it was that conversation that enabled him to give an affirmative answer when Mr Sullivan had asked him whether or not he should proceed with the Dell hardware sales after the acquisition. All this was, however, hotly disputed by the Claimants, who noted also that the alleged conversation went beyond Dr Lynch’s pleaded case of a discussion about “the synergies that might be achieved in respect of hardware” . In his witness statement on behalf of the Claimants, Mr Robison denied any such discussion about hardware and emphasised that he was “not aware of their possible existence until I heard that HP was making that contention in connection with its write-down, long after I retired from HP.” He added:

“Moreover, if Dr Lynch had informed me that Autonomy was reselling a substantial amount of third-party hardware and reported it as revenue, I absolutely would remember because that information would have immediately raised serious questions about Autonomy’s reported margins…”

Unfortunately, Mr Robison had a medical condition resulting in it being agreed that he was unfit to attend for cross-examination. Dr Lynch was cross-examined and it was put to him that he had made up the conversation. In the course of that cross-examination, Dr Lynch was shown emails that suggested that Mr Sullivan had been given approval to continue the Dell hardware sales in August 2011 and not in late September 2011. Dr Lynch reacted to this by suggesting that perhaps he had had more than one discussion with Mr Robison, including also one on the very day of the acquisition (18 August 2011):

“I think, although I’m not sure, that actually what happened, Mr Sullivan’s email^{^[249]} comes in while I’m standing next to Mr Robison. That’s why I’m able to give him an answer. And then the subject of these things come up on more than one occasion between that point and the 21^st.

What we’re doing is, as a courtesy, every time there’s something that “We need to know what will happen if the deal closes”, we’re asking Shane [Robison] what he wants to do.

And you can see in there there’s communications and Shane sends some of them on to Leo [Apotheker] as well.”

The Claimants emphasised that no conversation of that kind on the day of the announcement had ever been suggested and it was not corroborated by the documents or supported by any other evidence; they dismissed this as “simply made up…in the witness box in the hope that he could salvage his position.”

In the round, in my judgment, there is no sufficient evidence to demonstrate pre-acquisition knowledge on the part of HP of any material hardware sales (other than a small number of appliance sales); nor is there sufficient evidence that the due diligence team were put on inquiry in that regard.

Post-acquisition knowledge and discussion of hardware

The Defendants contended that whatever HP had known before the acquisition, it is clear that HP’s accountants and management had a detailed understanding of Autonomy’s hardware sales from very shortly afterwards.

They relied principally on HP’s reaction to what it would depict as a “discovery”: no complaint was made or surprise expressed, and the sales were permitted to continue, because, so the Defendants suggested, none of this came as a surprise because HP already knew.

Although they became co-mingled in the Defendants’ closing submissions, there were two strands to the Defendants’ resulting case: the first strand related to the pre-acquisition period, and in summary was that the lack of surprise shown by persons such as Mr Sarin confirmed that the deal team was aware even before the deal was finalised that Autonomy was selling substantial amounts of hardware. That point goes to both reliance and loss. The second strand related to HP’s conduct in announcing in November 2012 that it had found out about the hardware sales only after Dr Lynch’s departure, which was false and showed that HP had misled the markets, casting doubt on their corporate integrity and confirming the trumped-up nature of all the claims.

The Claimants disputed both the factual basis of this aspect of the Defence and its relevance. Ms Whitman stated in her witness statement that she did not know of any ‘pure’ hardware sales until Dr Lynch told her about them in 2012 and she instructed him to discontinue the practice (and was not cross-examined about that). Ms Lesjak and Mr David Duckworth (“Mr Duckworth”, who had the title of HP’s Mergers, Acquisitions, Divestitures and Outsourcing, Finance Integration Project Manager) all lined up to say (in effect) that the sale of hardware (or at least non-appliance hardware) did not become a point “in focus” for them until the “whistle-blower” (Mr Scott) “came forward” in May 2012. In any event, the Claimants added, Bidco’s cause of action was complete at the time of the acquisition, and “it is difficult to see what turns on when precisely after the acquisition HP learned of the earlier hardware sales.”

Turning to the chronology, it is clear that KPMG (who were in the course of preparing a closing balance sheet report) discovered the volume of hardware sales after a question had been raised about a large amount payable to Dell, and that then KPMG and E&Y (who, as HP’s auditors, were reviewing the report) each informed HP about the hardware sales in late October or early November 2011:

(1) On 24 October 2011, KPMG wrote a draft report addressed to Meeta Sunderwala in HP’s finance department, dealing with Autonomy’s closing balance sheet. It stated that:

“ There is approximately $41 million owed to Dell as of Close ($22 million in payables and $19 million in accruals). We believe these payable [sic] are related to pass-through hardware sales to customers which Autonomy records on a gross basis ” [Emphasis supplied]

(2) Later in the report they added that:

“ The remaining accounts payable balance mainly relate to data center server costs, or hardware Autonomy sells on a pass-through basis ”.

(3) By early November 2011, E&Y had a detailed understanding of Autonomy’s hardware sales. On 4 November 2011, Ms Rebecca Norris of E&Y sent an email to others on the E&Y team, stating that “We have finished our review of Deloitte’s workpapers for the 2010 audit of Autonomy” ; and the email noted that:

“The company had about $100 million in hardware revenue. This is not mentioned in the financial statements or KPMG reports. Long story short, they have VAR arrangements with some of the large hardware providers and th [sic] $100 million of hardware revenue is primarily just pass through revenue for laptops and servers. This hardware is normally sold at a loss to 4 or 5 large customers in advance of software sales. Some of these costs are allocated to sales and marketing. Their software is normally not on the hardware. We can further discuss commercial reasons and accounting on our call.”

(4) On 7 November 2011, Mr Kirk Parish of E&Y wrote to Rachel Scott, head of HP’s revenue recognition team, stating that it:

“ Looks like there may be as much as $100 million in hardware sales on an annual basis, but it was not mentioned in the due diligence report. ”

(5) On 9 November 2011, Brian Outland of E&Y raised the same point in an email to Ms Betsy Branch of HP, noting that some of the hardware was sold at a loss and the loss allocated to sales and marketing; and he said that he was going to discuss the point with Ms Sunderwala.^{^[250]}

It is also clear that on 11 November 2011, Mr Welham and Mr Murray of Deloitte discussed the hardware sales and accounting at a meeting with E&Y. E&Y’s note of the meeting shows that they were told the amount of the sales, the amount of the loss incurred, the commercial rationale and the accounting treatment:^{^[251]}

“ Hardware sales

During FY1O we observed that Autonomy purchased hardware from third parties (e.g. Dell) of $115m and sold this to customers for $105m as part of its goal to become the end to end supplier of archiving services for its customers. The associated cost of this hardware was split between cost of sales ($94m) and sales & marketing ($21m) which results in an 11% margin recognised on the hardware sales. The recognition of a margin was considered appropriate by management. However, Deloitte recorded an audit adjustment in relation to this margin. In addition, we observed that Autonomy concluded that it was appropriate to account for these pass through sales on a gross basis. HP will be required to assess the appropriate accounting and policies for hardware sales under US GAAP and HP policy.”

When shown this document in cross-examination, Mr Yelland accepted that E&Y were fully aware of the hardware sales being carried out by Autonomy.

The first strand of the Defendants’ case was focused on the reaction of KPMG and Mr Gersh, and then of Mr Sarin. I consider first the reaction of KPMG.

Mr Gersh and his team were (in Mr Gersh’s words in his witness statement) “shocked and concerned to see such large payables to a hardware manufacturer.” Mr Gersh added (in his witness statement):

“The existence of significant payables to a hardware manufacturer was inconsistent with what we understood to be Autonomy’s business from the information we received during pre-acquisition due diligence. Margins (and profitability) on hardware sales are typically much lower than margins on software sales, which could mean that HP had overvalued Autonomy. Mr Boggs^{^[252]} and I immediately began asking Autonomy management about Dell payables.”

Mr Gersh described in some detail in his witness statement the extent and results of those investigations of Autonomy’s management, and in particular conversations and exchanges with Mr Chamberlain. It is not necessary to repeat this detail: the flavour of the exchanges and their result can I think be captured in the following examples of Mr Chamberlain’s responses:

(1) After further enquiries over November and December had prompted no response from Mr Chamberlain, and his sudden cancellation of a meeting to discuss the matter, on 24 January 2012, Mr Chamberlain wrote in an email in purported answer to the details requested of the “Dell payable” as follows:

“…for certain strategic accounts we also procure hardware as well as software. This will all be sourced via HP in future but the process of setting up procurement via HP is painfully slow.”

(2) In response to an email dated 6 March 2012 from Ms Sunderwala^{^[253]} (after HP’s EFR team had, through her, been told of the continuing inquiries) asking him “What percentage of [Autonomy’s] total annual revenues is related to hardware revenue (this can be historical revenues related to Dell purchases vs a go forward view)”, Mr Chamberlain sent an email two weeks later, on 21 March 2012, informing her that this was his last day at Autonomy and in relation to the hardware question:

“ confirmed, less than 10%”

KPMG referred to the hardware sales in a draft report on Autonomy’s closing balance sheet, dated 26 January 2012. The report stated that:

“The remaining accounts payable balance mainly relate to data center server costs, or hardware Autonomy sells on a pass-through basis. Management stated these hardware sales are recorded on a gross basis.” [Emphasis supplied]

Mr Gersh explained in his witness statement that by March 2012 HP’s October 2011 financial reporting deadline that KPMG’s closing balance sheet engagement had been intended to support had long since passed, and EFR in effect took over the outstanding inquiries. He completed his witness statement as follows:

“At the time KPMG’s involvement ended, I was not aware of any satisfactory explanation for the Dell payables having been provided by Autonomy, and my team and I still did not understand what hardware Autonomy had been selling, in what arrangements, and with what associated revenue recognition.”

I consider that Mr Gersh’s reaction of surprise was genuine, confirming my previous assessment that he had not become consciously aware of the pure hardware sales in the course of due diligence.

Turning next to Mr Sarin:

(1) On 15 November 2011, Ms Kathryn Harvey (“Ms Harvey”) of HP’s EFR M&A Policies and Reporting team wrote to Mr Sarin about Autonomy’s hardware sales:

“ There was one other item that we just learned which I wanted to ask about/bring to your attention. During discussions with Autonomy folks in conjunction with our valuations, we learned that they have had approx $100M/ year in revenue coming from the sale of Dell HW products. Was just curious if you were aware of this when you put the model together? It doesn’t have any impact on our valuations, but it likely won’t be part of Autonomy’s future revenue stream and didn’t know whether it was included in their revenue forecasts/targets.”

(2) Mr Sarin responded by email on the same day as follows:

“The Dell-based revenue you reference is from which Autonomy products? They are predominantly a software company with little or no services or hardware. I am guessing they are trying to grow their “appliance” business i.e. Autonomy software bundled on industry-standard Dell hardware. I suspect this is sell-through revenue where they are getting a margin as they sell Dell appliances. Again, I am just conjecturing and don’t know for sure.

I don’t believe Autonomy breaks out this Dell-based revenue in their financials and we haven’t either…”

(3) But after that response, Mr Sarin did nothing to discuss the matter further or follow it up in any way.

In cross-examination, Mr Sarin struck me as uncomfortable:

(1) He could not explain why he had never spoken to Ms Harvey about what, on the Claimants’ case that nothing was known about hardware sales except for de minimis appliance sales, was a startling revelation.

(2) He told me that he “suspected at this time or my guess at this time was that what she was telling me was probably not right” , but he offered no explanation why he did not seek to find out how it could have occurred and correct it: he accepted that he had done nothing to follow up on the matter in any way.

(3) He resorted to suggesting that he thought it must be appliance sales: but that was hardly consistent with the notion of such sales being minimal in amount.

(4) He then resorted to seeking to pass off the matter as simply an issue for accounting; he told me:

“Again, she is just doing whatever she’s doing in the accounting department. I don’t have to go and correct everybody.”

(5) Finally, he resorted to saying that he “wasn’t tasked with that assignment and… left HP a few months after this acquisition.”

Mr Sarin’s evidence was unimpressively given and offered no answers. The question why he had responded to Ms Harvey so impassively and then done nothing to follow up remained unanswered. I do not accept that he thought that Ms Harvey was simply mistaken; nor that the explanation for $100 million was that the sales were appliance sales. In retrospect, his conduct looks strange: and I suspect his discomfort in the witness box was because he could see that.

But ultimately the question is whether Mr Sarin turned a blind eye because he knew all along what would be revealed; or whether there is some other explanation: perhaps that he was simply, by that stage unconcerned, especially since Ms Harvey had expressly stated that it did not impact the valuation; or perhaps because he put his head in the sand and hoped that something would turn up which did not demonstrate inadequacies in the due diligence exercise for which he might be blamed in circumstances where much blame was being thrown around. In my judgment, one of the latter two is the more likely, because I remain persuaded (and have so found) that he did not know of significant hardware sales. Furthermore, had he known, the obvious course, given he was told that it would not impact the valuations, was to explain it away, not leave it for others to answer.

In my judgment, the evidence does not establish that HP knew of the substantial pure hardware sales prior to the acquisition: and I find that they did not.

The second strand focused especially on the position of Ms Lesjak and when it was that she came to know of the hardware sales. I did not understand it to be suggested she had any such knowledge before the acquisition. It is most unlikely that she did: she was an opponent of the deal and it seems to me very likely that she would have deployed against it anything she knew about as showing there to be a problem.

As to the investigations which revealed the sales, she repeatedly asserted in her oral evidence that none of these matters had been brought to her attention at the time. She asserted in her oral evidence that she did not know of the work being done “several layers down in my organization”. Further, she told me that she did not recall either Ms Sunderwala or Mr Murrin speaking to her about these matters at the time. She said she was:

“not informed about it until after the whistle-blower came forward, except for some documents that I’m sure we’ll get to, that were presented to the audit committee and me later.”

The Defendants made the point that if that were right it would suggest that they were not considered to be important news. But they submitted that in fact, it appears that the hardware sales must have been drawn to Ms Lesjak’s attention at about that time.

On 11 November 2011, Ms Lesjak had one of what she told me were her regular CFO update meetings with E&Y. A short slide deck was produced by E&Y as an agenda for the meeting (2 pages plus a cover sheet). One of the agenda points in relation to the Autonomy acquisition was a statement that “Revenue includes $115M of hardware.” The same day, Ms Lesjak approved an E&Y slide deck for HP’s Audit Committee meeting the following week, saying “ Looks good. ” The slide dealing with Autonomy (one of the Q4 areas of audit focus) identified that 11% of Autonomy’s revenue was from hardware.

The Defendants submitted (and had put to Ms Lesjak) that it is highly likely the point was discussed with her, given that it was on the agenda, and given that E&Y had been flagging it in emails over the previous days sent to people in Ms Lesjak’s department. She said that she did not recall a discussion on the point, but reluctantly accepted that it was possible there was. The Defendants submitted that the truth was that she did not remember either way, but was trying to avoid accepting that she had in fact known about the hardware sales long before HP says, and said to the world in November 2012, it had discovered them. If she did not raise it with anyone else at the time, the Defendants contended that is because she did not think it remarkable that Autonomy had $115m of hardware revenues.

This was obviously a point Ms Lesjak had anticipated being raised, and her evidence had plainly been mulled over in advance. Her position in cross-examination was, in summary, that she simply did not focus on what was set out. Unsurprisingly, since she plainly valued her reputation as a safe pair of hands and may well have felt she would be supposed to have dropped the ball, her evidence came across as rather defensive. She refused to accept that it was obvious that the point referencing hardware revenue would have been discussed; she resorted to the position that it was possible that it had been, but as she could not remember, she could not accept that it was probable. I accept the Defendants’ submission that it is more likely than not that she would have read and understood the reference to revenue from hardware sales: and I find that she did so. It is less clear whether there was any discussion of the point: her evidence was that she “ did not recall ever seeing or focusing on or having a discussion about hardware.” On balance, I accept that; but t he question remains why it did not register with her as a concern. In one of her many denials that she had understood the true position, and the fact that she did not react with surprise demonstrated that she had known about it long before then, she said this:

“…I don’t actually know what happened. I’ve no recollection of a discussion around the hardware. In fact when, during an internal investigation after the whistle-blower, they showed me the decks, I was quite surprised because I did not recall ever seeing or focusing on or having a discussion about hardware.”

She suggested that this may have been because she would have assumed that any such sales were (a) properly accounted for and (b) appliances: but she accepted this was hypothesis, rather than recollection. However, she went on to make a further point to explain why the point would not have concerned her, which was considerably relied on by the Claimants: she and her team were only focusing on the accounting. As a matter of pure accounting, there was nothing wrong with “sell-through sales”, “as long as it was accounted for correctly and disclosed…” She told me in cross-examination:

“…I also think what’s really important to understand is, even if I focused on it, the most that I could hypothesise is that I thought it was appliance, and that the deal team that put the business case together also understood that there was hardware and that was captured in the business case that they put together.

…

My team, from top to bottom, did not know what was in the business case. So from their perspective, as long as this was accounted for correctly, and ultimately disclosed appropriately and the deal team knew about it and it was captured in the business case, there isn’t an issue. The issue arises if the due diligence team never understood there was hardware, never understood in what product category that hardware is being reported and disclosed to them, and it’s in that failure that the issue arises.”

Mr Yelland was also aware of Autonomy’s loss-making hardware sales, long before the investigation preceding the write-down. On 22 March 2012, he received an email from Mr Ganesh Vaidyanathan, Autonomy’s Director of Accounting, asking him to approve a payment to Dell of $942,072.33. The email explained that “ This is for the hardware orders that we source from them and sell through to our customers at a loss of approximately 10%. ” Mr Yelland accepted that he read the email, although he had not by then taken up his post as CFO. Once in post, on 17 April 2012, he received and approved a similar request, this time for $478,734.43; again, this explained that it was “ sell through to our customers ” at a 10% loss. He said that he did not know the business rationale for the sales, but he raised no objection to the payment; and he accepted that he had access to all the financial information to tell him precisely the extent of the hardware sales. Indeed, he continued to approve payments to Dell as late as Q3 2012.

Ms Harris’s evidence was also that HP had detailed knowledge of the hardware sales from an early stage. She explained that Mr Duckworth was already aware of the hardware sales when she first met him in Pleasanton (in or around November 2011), and that he knew the details of all the accounting codings. Indeed, she said that it was “ entirely impossible that [HP] weren’t aware of ” the hardware sales and hardware cost of sale s given the conversations she had had with Mr Duckworth. It was clear from her cross-examination that she was outraged when she first learned, via Mr Hussain, that HP was contending that it was unaware of the sales until a later date. If the matter had been a shock or surprise to HP, queries would have been raised, but Ms Harris never received any information requests from HP about hardware or hardware customers.

Mr Duckworth himself expressly denied in cross-examination when it was put to him that he was aware of Autonomy’s hardware revenues and hardware costs when he conducted what was referred to as “the mapping exercise”. He explained:

“ I was not. That was not my focus. There was lots of different accounts, lots of different cash accounts, lots of different, you know, accounts across the balance sheet. My structure - my focus area was not on the type of activity that Autonomy was doing. It was just to make sure that whatever they had in their general ledger had an appropriate mapping to the HP account. I actually did not remember this being in the chart of accounts mapping until after Lisa Harris had referenced me in the suit and said that we had this conversation, I actually had to go back into the chart of accounts to find on this account that there was indeed a mapping exercise, because I had no recollection of this. I also searched 3,000 emails that I had on Autonomy, and I saw no reference to any hardware revenue or hardware reference whatsoever.”

Ms Harris also gave evidence that in early April 2012, Mr Yelland had asked her “ why aren’t you selling HP hardware? ”; and she also said that she believed that he would have known the commercial rationale for the sales. She was not challenged on either of these points.

Lastly, the Defendants relied also on the open (that is, not in any way covert or disguised) continuation of Autonomy’s hardware transactions post-acquisition. After the acquisition, Autonomy sought hardware from HP to supply to its customers. In April 2012, in the wake of the supply chain problems caused by flooding in Thailand, Autonomy was having difficulty obtaining hard drives from HP, to be supplied to Citibank as part of an appliance deal (as Dr Lynch accepted in cross-examination). Dr Lynch raised the issue directly with Ms Whitman, who would have been aware then that Autonomy was selling hardware, but expressed no surprise. She was not cross-examined about this. Further, in his evidence in the US trial, Mr Sullivan stated that hardware sales continued post acquisition, and no one suggested hardware should be hidden.

Indeed, it was only in June 2012 that a decision was taken to stop hardware sales. An email dated 13 June 2012 from Mr Yelland to Mr Veghte records a number of decisions from a call the previous day. These included various “ Decision [s] regarding the business we do ”, the first of which was as follows: “ We are no longer reselling hardware unless an appliance or key enable [sic] of software sale. ”

The Defendants concluded on this aspect of the matter that:

“This history belies HP’s claim that knowledge of the full details of the hardware sales would have been significant to it. It knew enough before the acquisition to make further inquiries and chose not to. It knew everything after the acquisition and was relaxed. Nobody thought they had bought a different company. The history also shows that HP misled the markets in November 2012 when it claimed that it had found out about the hardware sales only after Dr Lynch’s departure.”

My assessment

There is no doubt (and I find) that by November 2011, relevant persons within HP and its investigating accountants (E&Y and KPMG) knew that Autonomy had been undertaking hardware sales to the value of in excess of $100 million per year through the Relevant Period.

However, HP was a huge company with 325,000 employees, and organizationally both bureaucratic and fragmented. It is important to be careful to distinguish between (a) what was known prior to the acquisition by the ‘deal team’, (b) what was known prior to the acquisition by Ms Lesjak and the finance department, (c) who came to know of those sales after the acquisition, and (d) what the reaction of each group to the information provided by E&Y and KPMG was and what (if anything) it reveals about their then state of knowledge or understanding. It is convenient to approach the matter in reverse chronological sequence, starting in consequence with post-acquisition knowledge.

There is no doubt (and I find) that Ms Sunderwala, and thus the accounting side of HP’s EFR team, knew of the volume hardware sales as from late October 2011, both from KPMG and from E&Y. So too did Ms Betsy Branch who was also in the EFR team. However, I do not consider that they or the tone or content of the investigations made by KPMG and HP’s EFR department through Ms Sunderwala suggest that they were already aware of the nature and extent of the hardware reselling strategy. On the contrary, they seem to me to suggest genuine surprise and concern.

The position in relation to Ms Lesjak is curious. She had built her career on caution and carefulness, the “safe pair of hands in a crisis”. Yet her response to the suggestion that in a two-page briefing note and slide stack she had been told expressly that “Revenue includes $115 million of hardware”, now said to constitute serious dishonesty, is that she simply did not focus on any of it before approving the slide deck for submission to HP’s Audit Committee saying “Looks good”.

The Defendants invited me to find that (a) the inference from Ms Lesjak’s approval of the slide deck was that she had read it, (b) she had known long before HP now says, and said to the world in November 2012, it had discovered them but (c) the reason she did not take any action upon finding out was that it was neither a surprise nor a concern that Autonomy had this source of revenue from hardware sales.

In my judgment, it is more likely than not, and accordingly I find, that Ms Lesjak did read the slide deck and covering note, and also that she attended a discussion about the hardware revenue (though I accept she did not recall it). The question then is what conclusions should follow from that.

Contrary to the Defendants contentions as summarised in paragraph 1849 above, Mr Rabinowitz submitted (in his oral reply) that Ms Lesjak’s lack of focus and concern was of no material significance: the reference to hardware sales would only have been of concern to Ms Lesjak had she thought it was something missed by the ‘deal team’ in constructing and presenting “their business case” (for the transaction); and she did not think that. Mr Rabinowitz referred me to her evidence in cross-examination to the effect that it did not register with her as of significance until after the whistle-blower had come forward and detailed the nature of the hardware sales, their use and how they were accounted for. In other words, although she did know from then on of the hardware sales, she did not focus on what they might signify was not something she grasped properly until later.

It seems to me that the point that ultimately I have to determine is whether Ms Lesjak’s defensiveness and/or her lack of surprise and concern, supports the Defendants’ case that HP was well aware all along, and certainly before the Acquisition went unconditional, that Autonomy was generating revenues from third party hardware sales. In the end, I have concluded that her discomfort was because, having been told of the hardware sales, she felt that it might be thought that she had dropped the ball in neither appreciating their significance nor digging deeper at the time, and not because she was dissembling in this regard. I had no sense from her evidence, her defensiveness or her demeanour, and I do not consider, that she had any appreciation then, or until the whistle-blower caused her to investigate further, of the true use, purpose and vice of third-party hardware sales to disguise and provide a stop-gap for shortfalls in software revenues.

Conclusion as to when HP became aware of “ pure hardware sales”

As the Claimants noted in identifying the issues to be determined in relation to their Hardware case, the Defendants raised but did not fully explain the relevance of post-acquisition knowledge of the hardware sales; but pre-acquisition knowledge would be relevant to the question of reliance.

I have concluded that neither Mr Sarin nor Mr Gersh, nor anyone else whose knowledge could be imputed to HP/Bidco, knew that Autonomy was routinely selling material amounts of third party hardware separately from any software sales, still less did they know that (as I have determined) their purpose was to cover shortfalls in software revenues and perpetuate the appearance of meeting revenue forecasts, and that to that end their true nature was being concealed and their effect on Autonomy’s trading performance was being disguised.

THE CLAIMANTS’ ‘VAR CASE ’ : THE 37 IMPUGNED VAR TRANSACTIONS

Overview of the parties’ respective cases on Autonomy ’s VAR sales

The Claimants ’ case

The Claimants ’ claims in respect of Autonomy ’s resort to the use of intermediaries known as Value Added Resellers ( “VARs” or “ resellers”) vied with their claims in relation to Autonomy ’s sales of ‘ pure ’ hardware in the importance the Claimants attached to them.

Although they accepted that Autonomy also entered into a large (but unspecified) number of unobjectionable VAR transactions in the Relevant Period ^{^[254]} , the Claimants identified and impugned 37 high-value VAR licence sales ( “the impugned VAR sales”) as being contrivances to which Autonomy resorted when an intended sale by Autonomy to an end-user of the same subject-matter as the impugned VAR sale was unexpectedly delayed (or in some cases aborted), leaving a revenue shortfall. Almost all of them were with one of a small cohort of ‘friendly ’ VARs.

According to the Claimants ’ case, the context of every one of the impugned VAR sales was Mr Hussain ’s urgent need, at the end of a given quarter, to find revenue which could be presented as satisfying accounting rules for the recognition of revenue, in order to enable Autonomy to meet revenue forecasts.

The purpose of every impugned VAR sale, according to the Claimants, was simply to establish a contract with a VAR, in place of the proposed end-user, from which recognised revenue in the amount of the VAR sale would appear to be generated even though in reality, revenue would only ever be received when the contemplated (but delayed) sale to the proposed end-user actually took place. In some instances, no sale to an end-user ever eventuated. In such cases, the Claimants ’ case was that other contrived transactions (referred to as ‘Reciprocal VAR Transactions ’ ) were entered into in order to enable Autonomy, in effect, to fund the VAR to enable it to appear to discharge its obligations. In a smaller group of cases, the VAR ’s obligations were, if no end-user sale eventuated, written off altogether; but that was not a solution favoured; it ran the risk of exposing the lack of substance in the VAR sales.

This use of VAR sales took advantage of the fact that under IFRS, unlike under US GAAP, there is no requirement for there to be “sell-through” to an end-user before revenue is recognised. Under US GAAP, the use of VAR sales to generate recognised revenue where no end-user sale had yet been put in place would not have been feasible.

Three special characteristics of software also enabled this use of VARs in such a way at the last moment at the end of a quarter:

(1) Software can be delivered electronically and more or less instantaneously at the last minute.

(2) Acting as a VAR carries transactional risk in respect of its outlay; but that will eventuate only if no onward sale can be arranged; and whilst the commissions are generous, a VAR need carry no inventory and its overheads may be relatively small.

(3) Typically, what the VAR acquires is a licence to software: the seller retains the rights, and the ability to sell the licence (an intangible asset that can be infinitely and cheaply reproduced) to an infinite number of other users. A producer/seller (such as Autonomy) is in a position to fulfil with ease a licence sale to the end-user, notwithstanding an earlier licence sale to the VAR, if that is what it chose to do.

The acceleration of the recognition of revenue apparently generated by the impugned VAR sales (and in some cases the recognition of revenue which should never have been recognised at all) enabled Autonomy to cover a sudden or unexpected shortfall and meet quarterly and periodic revenue forecasts. The Claimants’ case was that the objective was thus very similar to, and could be deployed to complement, the objective of Autonomy’s ‘pure’ hardware sales, and both expedients were improper. Unlike the ‘pure’ hardware sales, however, the VAR sales were fully disclosed and their ostensible purpose was never in dispute.

The Claimants contended that Autonomy ’s use of VAR transactions in this way started in earnest in Q4 2009, immediately after what they described as the Defendants’ “ lever of choice” in Q3 2009 and its favoured strategy of reselling pure hardware to plug any shortfalls in revenue came briefly unstuck in Q4 2009, when Autonomy ’s main hardware supplier, EMC, withdrew from any involvement in “Autonomy ’s hardware reselling strategy.” On this analysis, the use of VARs was another lever or device to which Autonomy resorted, with the Defendants ’ knowledge and authorisation,

“to bridge the gap between [Autonomy ’s actual revenue for the quarter] - which was invariably far short of [analys ’ ts and market] consensus - and Autonomy ’s reported figure, which was generally in line with or ahead of consensus.”

The Claimants depicted the VARs which Autonomy engaged for its VAR sales as friendly ‘placeholders ’ which were willing to go along with a nominal role accepting nominal risk which involved no real effort or exposure in return for commissions in the form of “Marketing Assistance Fees” or “ MAFs”.

The Claimants sought to demonstrate a tell-tale pattern of a VAR sale at the very last moment of a quarter. The VAR would nominally accept a payment obligation and thus risk, but would know that Autonomy would be entirely responsible for negotiating an end-user sale, and would never call on the VAR to pay the amounts due under the sale transaction out of its own resources. The shared understanding, of which they were in every case assured by Mr Egan, was that Autonomy would wait until either payment eventuated from an onward sale to an end-user, or the VAR could be put in funds by some ancillary transaction orchestrated and funded by Autonomy.

The Claimants accepted that the 37 impugned VAR sales had contractual effect, and nominally transferred to the VAR risk and reward, and managerial control, of the software licence sold. However, the Claimants contended that the impugned VAR sales, all in amounts very considerably greater than any of Autonomy ’s other transactions with VARs, were not in economic substance sales at all, but arrangements for Autonomy to recognise revenue in advance of the true sale to an end-user.

The Claimants acknowledged that Deloitte had in every case approved the accounting treatment of the impugned VAR sales. Their position was that this, though much relied on by the Defendants, did not provide either a justification or a defence because Deloitte did not and could not know the full picture. In particular, Deloitte were invariably told, and saw in the contract, that there was an express prohibition against any side agreement which might undermine the main VAR contract; but they were never told of the shared understanding between the VARs and Autonomy that, one way or another, Autonomy would see to it that the VAR was insulated from the nominal legal risk.

Normally, as indicated above that would be achieved either by Autonomy crediting the VAR (directly or indirectly) with the proceeds from an end-user sale or by Autonomy funding the VAR through a reciprocal transaction. But the Claimants relied also on what they presented as a ‘clean up ’ operation engaged in by Autonomy personnel in what the Claimants referred to as ‘the Dark Period ’. This was the period after the acquisition by HP of Autonomy but before its integration, and whilst Autonomy ’s pre-acquisition management was still in control of its affairs. The Claimants alleged that during the ‘Dark Period’, past VAR and ‘ reciprocal ’ transactions which had not been paid for or completed were in various ways reversed or expunged. They relied on this as further demonstrating both the impropriety of the entire programme of impugned VAR sales and the knowledge and recognition of their impropriety.

Overview of the Defendants ’ case

The Defendants depicted the Claimants ’ case as based on assumptions which presumed impropriety where none existed, and on evidence in chief which was not reliable, and which, after cross-examination, was revealed not to be supportive of the Claimants ’ case.

Although VAR is an acronym for “Value Added Reseller”, and the paradigm would be a VAR providing additional value to a software company ’s products by providing services to the end-user, resellers commonly undertake different and very varied roles. The Defendants emphasised that there is nothing per se objectionable in the use of VARs as a channel through which software sales are made, sometimes simply as a ‘ fulfilment partner ’ with no active role: indeed, it is common in the software business.

According to the Defendants, the Claimants had sought to portray as suggestive of lack of substance and impropriety features which were commonplace in the market, and entirely “ neutral” . They suggested that the impugned VAR sales were, except for their unusual value, not materially different from Autonomy ’s many other sales to VARs: as being part of the ordinary course of business of a software seller and entirely unobjectionable.

Thus, they insisted that it was not unusual or improper to use resellers (as the Defendants preferred to refer to VARs) merely as sales channels, or simply as ‘ fulfilment partners’ assisting the supplier with the paperwork required to fulfil a sale to the end-user through the VAR sales channel; or, as in most of the VAR sales in this case, simply to enable the supplier to secure a sale (to the VAR) from which the supplier (Autonomy in this case) could book and recognise revenue in a targeted quarter if a contemplated sale to an end-user was delayed.

Accordingly, they presented the 37 impugned VAR sales as higher value examples of a commonplace sales strategy involving what in Dr Lynch ’s estimation were “thousands” of other unobjectionable VAR transactions with “hundreds of resellers” in the same period. They maintained that there was nothing improper, illusory or wrong about Autonomy ’s VAR sales, just as there was nothing wrong with selling hardware to boost software sales, in effect as a promotional gambit.

Mr Welham confirmed in cross-examination that Deloitte, for example, understood and accepted as unobjectionable the use of VARs simply to “de-risk” a delayed transaction and book revenue from it (subject to compliance with specified criteria in IAS 18.14 as discussed below). The Defendants submitted that, in the sector, it was not unusual for the reseller to act (in effect) as a guarantor. The effect of smoothing revenue to minimise peaks and troughs was not objectionable; and nor was a VAR transaction with that in mind if it satisfied the legal criteria of IAS 18.14.

The Defendants relied especially on the detailed contracts which in each case established the impugned VAR transactions. The terms of these contracts were clear. They provided for an unconditional sale, an unqualified payment obligation (regardless of whether or not an end-user sale eventuated), allowed the VAR to determine the price of any onward sale and included in every case an unqualified “entire agreement” clause. The Claimants, who (the Defendants repeatedly emphasised) expressly disavowed any suggestion of legal sham, accepted that they were valid and enforceable in accordance with those terms. In such circumstances, the Defendants contended that there was no basis for characterising the substance of the transactions as anything other than in law what the parties had defined them to be: the substance was the contract.

Mr Egan for Autonomy and both Mr Baiocco and Mr Szukalski on behalf of ‘friendly’ VAR s, confirmed that such a VAR sale was not unusual, and was a useful way of ensuring that a supplier such as Autonomy would not be pressurised to sell to a procrastinating or opportunistic end-user at a discounted price at the end of a quarter (as he also confirmed was a common end-user tactic). Amongst other advantages in the use of the VAR sales channel claimed by the Defendants were that (a) in theory at any rate, it permitted the supplier to reduce the costs of its sales function (as Mr Apotheker agreed in cross-examination) and (b) using resellers potentially gave Autonomy access to more customers (as Mr Egan agreed).

Mr Egan also confirmed that from the VAR ’s perspective, there was an incentive (in addition to any commission) for doing such a deal in that it would help the VAR ’s market presence and help build its own relationship with the end-user.

For the Defendants, the underlying flaw in the Claimants ’ approach was that it lost sight of both the contract (which the Defendants suggested was also the best evidence of the parties ’ intentions) and the point that, for the purposes of IFRS accounting and IAS 18.14, the VAR was the customer, and the status and prospects of any future deal did not impact upon recognition of revenue from the sale to the VAR. The Defendants suggested that the latter mistake originated in a different approach taken in US GAAP with which, of course, HP and its management were more familiar.

The Defendants relied also on the fact that the VAR transactions and their salient features were well-known to Deloitte, who gave each of the impugned VAR sales a clean bill of health. Thus, for example, Mr Welham confirmed that Deloitte were aware that Autonomy would regularly resort to a VAR sale, so that it could recognise revenue from that sale, at the very last moment before the end of a quarter, since it might only be then that it would become clear that the end-user sale would be delayed. He also confirmed in cross-examination that it would not be unusual for the VAR to have had no contact with the end-user, nor any familiarity with the software to be on-sold under licence.

The Defendants rejected any suggestion of improper reversals or other impropriety in their treatment of uncompleted VAR transactions in the “Dark Period”. They insisted that everything that was done was typical of the usual reconciliation and clearing up process following an acquisition. They relied also on the evidence of Ms Harris in her second witness evidence in support of this ^{^[255]} .

Number and value, and the two categories, of impugned VAR sales

The 37 impugned VAR sales are listed and briefly described in Schedule 3 to the RRAPoC, which is headed “Contrived VAR Transactions” . Schedule 3 is 65 pages long and contains a brief summary of the Claimants ’ allegations in respect of each of the 37 transactions listed.

Of the 37 allegedly “Contrived VAR Transactions” (which in the Claimants ’ written submissions were referred to as “the impugned VAR transactions”), 30 were transactions in respect of which the Claimants alleged there was a side agreement or understanding which negated their economic substance, which resulted in the VARs ^{^[256]} being in every case no more than a ‘placeholder ’ and precluded the recognition of revenue from them.

5 of the remaining 7 transactions (VT9, VT17, VT19, VT22 and VT29) were ‘Collectability VARs ’ . In those cases, no such side agreement or understanding was alleged. The basis of the Claimants ’ claim was the allegation that there was never any realistic prospect of the VAR in question actually meeting its payment obligation, and it was never in truth envisaged that any debt owing by the VAR would be called in. Although there was no side agreement asserted, both motive and effect were similar: to generate revenue from what was in effect a contrivance. I deal in more detail with the Collectability VARs in the Schedule of Impugned VAR Transactions that accompanies this judgment. The 2 remaining transactions (VT5 and VT14) are dealt with on their own particular facts in that same Schedule.

The Claimants calculated that the impugned VAR sales accounted for 88% by value of Autonomy ’s VAR transactions above $1 million in the Relevant Period. As to their chronological spread, it is not disputed that:

(1) In Q4 2009, Autonomy made seven impugned VAR transactions each in excess of $1 million and which collectively accounted for $27.6 million in revenue;

(2) In Q1 2010, Autonomy made a total of five impugned VAR sales which collectively accounted for $28.4 million in revenue;

(3) After a lull in Q2 2010 when there was only one impugned VAR sale accounting for revenue of nearly $2 million, occasioned (the Claimants suggested) by Mr Hogenson ’s enquiries, Autonomy made four impugned VAR sales in Q3 2010 accounting for revenue of over $23 million (and two ‘ reciprocal’ transactions yielding over $7 million);

(4) Seven impugned VAR transactions followed in Q4 2010 accounting for revenue of $25.4 million;

(5) In Q1 2011, Autonomy made seven impugned VAR sales accounting for revenue of $22.9 million; and

(6) In Q2 2011 (the last reported quarter before HP ’s acquisition of Autonomy was announced in August 2011), it made a further four such sales, which accounted for revenue of some $28.6 million.

The amount of revenue recognised in respect of the impugned VAR sales in each quarter was alleged to be sufficient materially to distort Autonomy ’s quarterly report. That was so even though what is in issue, where (as in 19 out of 37 cases) an end-user sale eventuated, is a timing difference, because when the end-user sale was completed, revenue actually was received in most cases, albeit some time after it had been booked.

It was not suggested that the 37 impugned VAR sales were indicative or exemplars of some wider misuse of VAR sales by Autonomy. In other words, the Claimants confined their claims to the 37 impugned VAR sales, and did not challenge Dr Lynch ’s estimation in his first witness statement that Autonomy entered into thousands of transactions with resellers in the Relevant Period. I infer that the other VAR sales which were not counted were unobjectionable.

As previously indicated, the Defendants did not adduce evidence in respect of any such VAR sales to demonstrate similar but apparently unexceptional features. I did not therefore have the opportunity to assess whether the mass of VAR transactions which were not impugned shared features said by the Claimants to be objectionable in the context of the 37 impugned transactions, which might have confirmed the Defendants’ case that the features relied on by the Claimants (such as the fact that VAR sales were entered into on the eve of quarter-end, and that VARs were excluded from any involvement in the end-user negotiations, as elaborated later) were unexceptional.

Reciprocal VARs

Linked with both varieties of impugned VAR transactions were certain transactions, referred to by the Claimants as “Reciprocal VAR transactions”, which (somewhat confusingly) the Claimants lumped together with what the Claimants termed “Reciprocal Transactions” (as listed in Schedule 5 to the RRAPoC).

So-called Reciprocal VAR transactions were not really “reciprocal transactions” at all, since they did not involve any linked sale to and purchase from a counterparty. But what they were said to have in common was that they involved a purchase or purchases of a product (or services) which Autonomy allegedly did not “need” and were of “ no discernible value”. In other words, both varieties were put in place because of, and achieved, the same purpose.

In the case of ‘Reciprocal VAR transactions ’ , all of which were with ‘friendly ’ VARs, the Claimants ’ case was that the purchase from the VAR was alleged to have been a device to put the VAR in funds to enable it to repay what it owed in respect of an earlier purchase by it of a product from Autonomy under an impugned VAR transaction. The Claimants relied on these transactions as evidence that the VAR was not in reality on risk in respect of earlier sales giving rise to the debts, thereby showing that the earlier sales lacked substance. I deal with these ‘Reciprocal VAR transactions ’ later in this Chapter.

The Claimants ’ legal causes of action in respect of the impugned VAR transactions

The Claimants ’ primary and largest claim in this context was that because they were wrongly treated as real sales generating revenue which could be recognised at the point of sale, the way the impugned VAR sales were accounted for was improper and resulted in a false picture in Autonomy ’s accounts and published information (on which they maintained they had relied). This is thus in the nature of a claim for false accounting and untrue or misleading published information brought under FSMA, the relevant accounting standard said to have been wrongly applied being IAS 18.

The Claimants also made a claim for alleged “transaction-based losses”; the losses consisting of fees paid to VARs in the context of the impugned VAR transactions and what the Claimants described as “foregone receipts on improper transactions.” These claims for compensation for the Defendants ’ breaches of duty seek recovery of amounts which are relatively small, with a value amounting to some US$8.8 million.

The two hurdles for the Claimants in their FSMA claims in respect of impugned VAR sales

To succeed in their VAR claims under FSMA, the Claimants had to demonstrate (in addition to reliance and loss) both (a) that the recognition of revenue from the impugned VAR sales lacked economic substance and/or did not satisfy the criteria for recognition of revenue prescribed by applicable Accounting Standards, and that therefore Autonomy ’s published information showing it as recognised revenue was untrue or misleading, and (b) that the Defendants knew that (or were reckless in that regard).

As to (a), the Claimants put their case on two complementary bases. One basis was that in substance the impugned VAR transactions were not sales at all, and thus could not give rise to recognised revenue. The second basis was that the impugned VAR transactions did not satisfy one or more of the criteria stipulated by IAS 18.14 for the recognition of revenue from a sale of goods.

As the case progressed, the Claimants seemed to me increasingly to emphasise the first limb of their case in this regard (that the transactions were not in reality transactions of sale at all, and that this was conclusive without further need to consider the detailed application of the specific wording or interpretation of IAS 18.14). This had the advantage, from the Claimants ’ point of view, of reorienting the points in issue away from any dispute as to the proper interpretation and application of IAS 18 and towards a fact-based assessment, and also thereby minimising the role of Deloitte.

Although this latterly adopted approach is in some ways the more straightforward, it seems to me that it is important to understand the detailed provisions of IAS 18, both because that is the way the case was originally put, and because that was the apparent approach both of Autonomy and Deloitte.

In any event, both ways of putting the Claimants ’ case ultimately depend on a careful factual analysis of the actual conduct of the parties to the impugned VAR transactions. The Claimants maintained that this revealed a clear pattern consistent only with the conclusion that, though contractually enforceable, the VAR sales were not intended to have any real economic effect (save for the payment of commission to the benefit of the VAR as the price of its compliance in what the Claimants presented as (though not a sham in the legal sense)^{^[257]} a charade).

In a nutshell, the Claimants ’ case is that this pattern, especially when repeated over a series of transactions with ‘friendly ’ VARs, made clear that in every case, the VAR was and was always intended to be both insulated from risk and side-lined from any participation. That, the Claimants submitted, belied any true sale and demonstrated the VAR ’s purely nominal role.

As to (b) and the second hurdle, S.90A and Sch 10A of FSMA require proof of what I have referred to as (see the Introduction to this judgment at paragraphs 467 to 477) “guilty knowledge”. The Claimants thus had also to prove that the Defendants knew that the accounting treatment adopted by Autonomy, and sanctioned by Deloitte, in recognising revenue as arising from the impugned VAR transactions, was wrong and rendered Autonomy ’s financial statements untrue or misleading.^{^[258]}

That essentially depended on (i) what, if anything, the Defendants knew about the assurances given by Mr Egan and/or the way the VAR transactions were routinely acted upon, (ii) what basis the Defendants were satisfied that the criteria of IAS 18.14 were fulfilled in each case, and (iii) what comfort they legitimately drew from the fact that Deloitte reviewed and approved the accounting treatment of every one of the impugned VAR transactions. The latter was much relied on by the Defendants, on the basis that even if Deloitte were in error, it is difficult to ascribe greater and guilty knowledge to the Defendants, unless they knew something (including their own intentions or the undisclosed intentions of the parties) that Deloitte did not, and/or misled Deloitte into giving their approval of the presentation and accounts on a factually incomplete or misleading basis.

Structure of more detailed analysis of the claims based on false accounting

In light of the introductory preview above of a very detailed claim expounded by both the Claimants and the Defendants over the course of many hundreds of pages of submissions and many days of cross-examination of a variety of witnesses, I turn to address the following in more detail:

(1) the usual provisions of the contracts relating to the impugned VAR sales;

(2) the terms of and the proper approach to the interpretation of Accounting Standards, and in particular, IAS 18 para. 14 (“IAS 18.14”), which at the relevant time^{^[259]} governed the recognition of revenue from sales of goods (including software);

(3) how the expert evidence as to the application of IAS 18.14 affects the analysis;

(4) how the parties to the VAR sales actually conducted themselves, and in the light of that, whether in the case of the impugned VAR transactions the overall relationship between the contracting parties satisfied the criteria stipulated by IAS 18.14 for the recognition of revenue; and/or

(5) whether (as the Claimants submit) those transactions were enforceable but never really intended to be enforced or acted upon so as not in real economic terms to be sales of goods at all;

(6) what the Defendants knew and whether it amounted to “guilty knowledge”.

(1) Usual provisions of VAR sales contracts

Taking as an example the master agreement made between Autonomy Inc and Capax Discovery LLC dated 30 June 2009, it was provided that:

a) “…Once the Autonomy products on a purchase order have been shipped by Autonomy^{^[260]} , VAR may not cancel or amend the purchase order without the prior written consent of Autonomy.”

b) “…VAR shall not be relieved of its obligations to pay fees owed to Autonomy hereunder by the non-payment of fees by an End-User.”

c) “…VAR shall: (a) use commercially reasonable efforts to promote, market and sell Autonomy Products and Services to End-Users…(b) maintain a sufficient number of VAR ’ s personnel…trained in the features and functions of the current version of the Autonomy Products to (i) solicit orders for the Autonomy Products from End-Users; (ii) properly inform and demonstrate to End-Users the features and capabilities of the Autonomy Products…”

d) “Except as otherwise provided in this Agreement, VAR shall assume all responsibility and liability to End-

e) Users with respect to the Autonomy Products…”

f) “…VAR is free to determine its own prices and per copy fees to end-Users. Autonomy may from time to time publish suggested wholesale or retail prices and per copy fees, provided, however, that such prices and fees are suggestions only and VAR is and shall be free to determine the actual prices and per copy fees at which the Autonomy Product will be licensed to its End-Users”

g) “ ENTIRE AGREEMENT: AMENDMENT . This Agreement sets forth the complete and exclusive agreement between the parties with respect to its subject matter and supersedes any and all other written or oral agreements previously existing between the parties with respect to such subject matter. No alterations, modifications or additions to this Agreement shall be valid unless made in writing and signed by a Director or Officer of each party. The terms of any purchase orders or the like submitted by the VAR which conflict with any terms in this Agreement whether or not countersigned as accepted by Autonomy shall not be binding on Autonomy, regardless of Autonomy ’s failure to object to such terms.”

It is also of some interest, that in the definitions clause it was provided:

a) “End-User shall mean third party entities to whom VAR provides Services in respect of Autonomy Products” ; and

b) “Services shall mean services provided to an End-User by which VAR may include demonstration, pre-sales support, installation, customisation, training, maintenance and support services, and other consulting or integration of the Autonomy Products.”

I set out the terms of Autonomy ’s VAR agreements with the other ‘friendly ’ VARs (especially, MicroLink, MicroTech, and DiscoverTech)^{^[261]} , when addressing impugned VAR transactions with each of them. The master agreements with MicroLink and MicroTech (each described as a ‘Government reseller ’ ) were in similar form and materially the same as the agreement with Capax Discovery. The individual agreements with DiscoverTech were shorter but in each case they contained at least (a) a provision binding the reseller irrevocably to the purchase upon delivery and making clear the unconditional nature of the payment obligation and (b) a similar entire agreement clause.

Under all of Autonomy ’s agreements with ‘friendly ’ VARs, the reseller was also required to sign debtor confirmation letters in respect of the debt arising under the impugned deals. The debtor confirmations are identified further below by reference to the individual transactions. An example (which related to deals with Capax including Amgen) stated:

“The items listed above were properly charged to our account and were unpaid as of 30th September 2010 and there are no side letters or other agreements in respect of the subject matter of this request, except as noted below:

[Nothing was noted.]

We acknowledge that Autonomy Corporation plc retains no continuing managerial involvement in the delivery of this product or service, other than stipulated in the licence agreement.”

(2) Terms and interpretation of IAS 18

There is no dispute that Autonomy was subject, throughout the Relevant Period, to the requirements of IAS 18, and in particular IAS 18.14, and that revenue could not properly be recognised on sales of software (or hardware) unless its requirements were met. It was common ground also that if those requirements were met, recognition of revenue was mandatory, not discretionary: in other words, if all the criteria were satisfied, the revenue not only could be, but was required to be, recognised.

Terms of IAS 18

IAS 18 prescribed criteria intended to adumbrate the qualities of a true and unconditional sale of goods (including software and software licences) from which revenue arising must ^{^[262]} properly be recognised. Their application is not always easy, especially in the software context, where the goods are intangible assets that can be infinitely reproduced. It is of some note that IAS 18.14 has been replaced to clarify revenue recognition rules, especially for more complex transactions: IFRS 15 (which was first issued in April 2016) superseded it with effect for accounting periods beginning on or after 1 January 2018.

IAS 18.14 provided as follows:

“Revenue from the sale of goods shall be recognised when all the following conditions have been satisfied:

(a) the entity has transferred to the buyer the significant risks and rewards of ownership of the goods;

(b) the entity retains neither continuing managerial involvement to the degree usually associated with ownership nor effective control over the goods sold;

(c) the amount of revenue can be measured reliably;

(d) it is probable that the economic benefits associated with the transaction will flow to the entity; and

(e) the costs incurred or to be incurred in respect of the transaction can be measured reliably.”

IAS 18.15 elaborated on the test in IAS 18.14(a) as follows:

“The assessment of when an entity has transferred the significant risks and rewards of ownership to the buyer requires an examination of the circumstances of the transaction. In most cases, the transfer of the risks and rewards of ownership coincides with the transfer of the legal title or the passing of possession to the buyer. This is the case for most retail sales. In other cases, the transfer of risks and rewards of ownership occurs at a different time from the transfer of legal title or the passing of possession. …”

IAS 18.16 explained that the seller may retain “a significant risk of ownership” in a number of ways and provides examples of various well-recognised situations where that may be so, as follows:

(1) “when the entity retains an obligation for unsatisfactory performance not covered by normal warranty provisions”;

(2) “when the receipt of revenue is contingent on the derivation of revenue by the buyer from its sale of the goods”;

(3) “when the goods are shipped subject to installation, and the installation is a significant part of the contract which has not yet been completed by the entity”;

(4) “when the buyer has the right to rescind the purchase for a reason specified in the sales contract and the entity is uncertain about the probability of return.”

IAS 18.17 contained a saving provision as follows:

“If an entity retains only an insignificant risk of ownership, the transaction is a sale and revenue is recognised. For example, a seller may retain the legal title to the goods solely to protect the collectability of the amount due. In such a case, if the entity has transferred the significant risks and rewards of ownership, the transaction is a sale and the revenue is recognised. Another example of an entity retaining only an insignificant risk of ownership may be a retail sale when a refund is offered if the customer is not satisfied. Revenue in such cases is recognised at the time of sale provided the seller can reliably estimate future returns and recognises a liability for returns based on previous experience and other factors.”

Dispute as to interpretation of the Accounting Standards

These provisions resonate with, and indeed seem to me likely to have been informed by, legal tests for determining the transfer of risk, reward and managerial control in the ordinary context of a sale of goods. They identify legal characteristics of an unconditional sale. They provide examples to delineate when risk is assumed to pass (and when not) which appear to be borrowed from legal concepts and devices (such as warranties which in effect introduce conditionality, contingent payment provisions and terms enabling rescission which are usually referred to as Romalpa clauses). The way these matters are dealt with appears to encourage, especially in a lawyer, a legalistic approach and primary focus on any contract governing the transaction in determining the nature and extent of the bargain. My own initial tendency to adopt such an approach fuelled a dispute between the parties as to the extent to which the VAR contracts were properly treated as determinative of the true relationship between the parties to them (Autonomy and the relevant VAR).

Mr Miles, supported by Mr Casey, embraced and encouraged an approach which put the contract at the centre of the enquiry as to the substance of the impugned VAR sales and focused on the legal relationship between the parties as defined there and in any legally enforceable arrangements ancillary to it. Although they accepted that “revenue recognition, like other accounting requirements, involves consideration of substance over form”, Mr Miles submitted that “the substance of the relationship is to be found in the entirety of the parties ’ obligations” . He included in this “the entire commercial package of rights” (some of which might be collateral to the main contract) but not “arrangements of some kind which are not of legal effect”. He submitted that the “guidance does not involve a quest for a nebulous “understanding” separate from the legal rights and obligations which, taken as a whole, comprise the parties ’ deal” .

In the present case, the Defendants submitted that where, as in every one of the impugned VAR transactions in this case , being contracts for the sale of goods and not in the nature of complicated contracts such as a finance lease, the contract expressly provided that the entire agreement of the parties was to be found within the four walls of the contract, and nowhere else, the substance of the transaction was to be (in most cases at least) determined from the contract itself.^{^[263]} He submitted that:

“in practical terms, once you ’ve accepted that the transactions are real and not shams, then you can see what the substance is. It ’s the same thing as their legal form”.

He explained this further as follows:

“generally speaking, commercial contracts are very good evidence indeed of the intentions of the parties, indeed of the economic transaction between them, and suppose that an accountant were to want to understand what the transaction between two parties consists of, the contract generally provides the answer.”

Mr Miles submitted accordingly that the Claimants ’ concession that the VAR agreements were in every case binding contracts “carries with it an acceptance that the contracts embodied the true intentions of the parties.” He elaborated on this in his oral submissions in closing as follows:

“So what did those contracts do? Although there were some differences, there were also common features. The contracts were licences of goods for relicensing to an end-user. As I said earlier on, they contain unconditional obligations to pay, stating in terms that the reseller could not avoid payment by reason of a non-payment by the end-user. They governed the question of when title passed. It passed on delivery and that was by the software being made available electronically. It wasn ’t necessary for the reseller actually to download the goods. Title passes on it being made available.

The contracts also said that the contract is the whole contract that supersedes all agreements with respect to it and that no changes would be valid unless signed in writing. That is very strong evidence, we suggest, of the substance of the transactions for accounting purposes, put at its lowest.

Now, it is also worth noting that the Claimants accepted that nothing they could rely on would have amounted to a deferment of the debt in law or estoppel. They said that on day 84 at pages 136 to 137 [TS84/136-137]. Again, that is an important concession and not, we suggest, the way that the case was pleaded. Whether it was or not does not matter. It is an important point that, in law, the contract has the effect that it has.”

Thus, Mr Miles concluded, it follows that the VAR agreements are to be interpreted as both binding in accordance with their terms and demonstrative of the true intentions of the parties, including the intention that the obligations specified by their agreement should not be modified or re-characterised by reference to ancillary arrangements that they had agreed should not be made. There was a true sale established by an enforceable contract, and the VAR would have no defence in law to enforcement of its terms (including the payment obligation it contained). Even if assurances were given as to the likelihood that Autonomy would never look to its legal rights, and would do all that it could to see that the VAR was not left “on the hook” , the fact remained that the VAR was in every case on the legal hook, and the assurances provided no defence or answer. They amounted to no more than “ warm words” intended to encourage the VAR to participate, but not intended to constitute any part of the substantive deal or (in particular) remove from the VAR the risk it had contractually assumed.

Mr Miles also warned against the use of hindsight. He submitted that revenue recognition needed to be assessed at the date of the accounts, and hindsight should be avoided. Mr Welham accepted this when he was cross-examined as “probably a fair statement to make” , and both experts appeared to agree that it was not permissible. On that basis, Mr Miles submitted that post-transaction events, including (for example) a decision by the end-user to deal and contract directly with Autonomy and Autonomy ’s decision in those circumstances to forgive the VAR ’s debt, should not affect the analysis if the VAR was on risk at the time of the contract between itself and Autonomy. The Defendants pointed out that this was indeed Deloitte ’s approach.

Mr Rabinowitz rejected that approach and cautioned me against what he was concerned might be the natural inclination of a lawyer to adopt it. His case was that the focus of IAS 18.14, and of the practice of accountants in applying it, is on the economic reality of the transaction in the round, and whether the arrangement is one from which the revenue intended to be recognised can actually be expected. The audience is not a court applying the law: it is the market, seeking to assess how the company is performing.

The Claimants ’ position was that the standard requires consideration of whether the factual circumstances demonstrate any different economic reality than is revealed by legal analysis of the contractual relationship said to constitute an unconditional sale. As it seems to me, the essence of their position was that it is the economic substance which must determine the accounting treatment: economic reality trumps (as it were) any legal interpretation of the effect of the contractual arrangements agreed by the parties. Accounting for transactions by reference only to legally enforceable commitments will not provide a true and fair view if in fact the parties ’ shared intention is to ignore the contractual elements and proceed on the basis of a different understanding.

Contrary to the Defendants’ submission, the Claimants submitted that this is not inconsistent with their disavowal of any plea of sham. They accepted the validity of the contract, and that each of the transactions was valid and had legal effect and could be enforced by the parties against each other in accordance with their terms. Their point was that the shared understanding and intention of the parties was that neither party would ever seek to rely on or enforce its legal rights against the other, and insofar as obligations such as the payment obligation had to be seen to be performed, such assistance as became necessary (for example, by providing funds via a reciprocal transaction) would be arranged.

The Claimants ’ position was to the effect that the VAR transaction was, in economic reality, a side-show: the VAR was interposed as a placeholder for the end-user, and the economically significant sale in every case where a sale resulted was the sale to the end-user. It was that transaction on which in reality Autonomy was focused, and which it regarded as the real sale which would result in actual payment and in due course receipt of revenue. The accounting significance and revenue recognition accorded to the VAR transaction was contrived and misleading accordingly.

This dispute as to the interpretation of IAS 18.14 and its application to the facts clearly raises issues about the interplay between legal concepts and the Accountancy Standards, including:

(1) Whether the substance of a transaction is to be assessed according to its legally enforceable incidents: or put another way, whether the contract is definitive of the parties’ intentions and relationship; and

(2) More particularly, whether the criteria specified by IAS 18.14, for example as to the transfer of “significant risk and rewards of ownership” , or as to what constitutes “continuing managerial involvement” , or as to what is meant in IAS 18.16 by the receipt of revenue being “contingent” on the derivation of revenue by a reseller of revenue from an onward sale, are to be answered according to the terms of the contract between the parties, or are to be tested and applied according to any different economic reality that the circumstances and conduct of the parties may reveal.

I have addressed the first of these issues in the Introduction to this judgment in paragraphs 592 to 606 above. As there explained, in my judgment:

(1) The objectives of IAS/IFRS accounting, being to achieve by adherence to carefully developed accounting statements a true and fair presentation and account of the substantive economic performance of the company, and to inform appraisals and investment decisions likely to be made in reliance on that account, require that the substance of a transaction be assessed by reference to its intended economic effect. This cannot safely be determined only by reference to its legal aspects. I cannot therefore accept the Defendant ’s argument in its purest form to the effect that a contract which is accepted to be valid and enforceable in accordance with its terms, and which is expressed to be the parties ’ entire agreement, must be treated as definitive of their intended relationship in respect of the subject matter of the contract. Regard must always be had to the shared intentions of the parties and, in the light of that, the real intended effect of the transaction in question.

(2) As the Claimants contended, the attenuated or narrower form of the Defendants ’ argument, to the effect that in this case the contractual provisions of each VAR transaction, and especially the entire agreement clause, render it unlikely that the parties had any intention which differed from that set out in such agreements, and the Court should be slow to find that they supplemented the written document by any further agreement or understanding, implicitly accepts that the totality of the evidence must be considered. Even if there is a presumption that the parties intended to set out their intention in the contract they signed, the presumption must yield to the evidence as a whole.

(3) The inclusion in the contractual provisions of an entire agreement clause may preclude the parties asserting any other agreement or understanding in vindication of or answer to a legal claim; but it cannot negate the facts or preclude enquiry as to the underlying realities in order to reach an accounting treatment of the transaction which provides a true and fair view of its true economic substance and effect. Indeed, the inclusion of an entire agreement clause may be revealed to be simply part of an intention to deflect attention from arrangements parallel to the contract which changed its economic effect.

As to the second issue, and with particular reference to IAS 18.14, in my judgment:

(1) The criteria for revenue recognition, though informed by legal precepts derived from the law on the Sale of Goods, are to be satisfied, not only in terms of the contractual effect, but also in economic reality: they require that the significant risks of ownership should actually, as well as nominally, be transferred, and that the seller should not in fact retain any continuing involvement to a degree usually associated with ownership, nor effective control of the goods sold;

(2) Thus, IAS 18.14 requires an assessment of the true intended effect of all the arrangements and understandings between the parties to the relevant VAR transaction; all the facts and circumstances which bear on that assessment must be taken into account in determining the real economic substance of the relevant transaction and in accounting for it so as to provide a true and fair view of it.

I accept that facts which emerge after the transaction are not to be taken into account because hindsight is not to be used. But in my judgment, the preclusion of the use of hindsight relates to after-discovered facts and circumstances, rather than to the proof of facts and circumstances already present at the time of the transaction. The true intentions and understandings of the parties at the time of the transaction are known to them at that time; and the proof of them to establish the real substance of the transaction by recourse to what occurs later is not prohibited.

(3) Expert evidence as to accountancy practice in applying IAS 18

In my view, the conclusions above expressed are consistent with accountancy practice as expounded by both experts (Mr Holgate for the Claimants, and Mr MacGregor for the Defendants), despite serious differences between them on various aspects of accountancy practice relating to IAS 18.

The experts agreed that as regards Accounting Standards in general:

(1) IAS/IFRS is less prescriptive than US GAAP in respect of all the matters in issue;

(2) The application of Accounting Standards to a specific company is a matter of judgement, and where the application of principles calls for judgement, there may be a range of possible views that a reasonable accountant may reach;

(3) Accounting decisions depend on the substance of the transaction and not merely its form; where the parties have agreed contractual terms they are highly material to, and likely to be strong evidence (or what Mr MacGregor referred to as “the driving force” ) of, the substance of the transaction, but they are not conclusive;

(4) Accounting judgements fall to be made on the basis of the facts and circumstances as they were at the accounting date and hindsight cannot legitimately be used.

In the context of the impugned VAR transactions, and as indicated above, the experts agreed that:

(1) The relevant contract to examine for the purposes of revenue recognition was that between Autonomy and the VAR: under both IAS and IFRS, unlike under US GAAP, there is no requirement for there to be sell-through to an end-user before revenue is recognised;

(2) It is not permissible to assume that all VAR transactions can be treated generically; the individual commercial facts relating to a given transaction at the time that it was effected are highly relevant to its accounting treatment;

(3) The revenue recognition principles in IAS 18 could be difficult to apply other than in simple transactions: the notion of the transfer of risk and reward was regarded as especially problematic, and IFRS 15 (which superseded IAS 18.14) replaced it with the question whether it is the supplier or the purchaser who in fact controls the goods or services in question;

(4) All of the IAS 18.14 criteria need to be satisfied in order for revenue to be recognised: but that once the criteria are satisfied, recognition of the revenue is mandatory;

(5) As to the criterion in 18.14(a), as expressly stated in 18.15, “in most cases the transfer of risks and rewards of ownership coincides with the transfer of legal title or the passing of possession to the buyer” and that will generally be determined by the contractual terms;

(6) The criterion in 18.14(a) does not require all the risks and rewards to have passed: the test is whether substantially all of them have, and IAS 18.16 sets out various well-recognised situations where, under the contractual arrangements, the seller may retain the significant risk and rewards of ownership;

(7) Generally (and though each situation should be considered individually), the criterion in 18.14(b) (relating to the need for transfer of managerial control) goes hand in hand with the risks and rewards of ownership, and it would be unusual for an entity to retain managerial involvement to the degree usually associated with ownership or effective control where risks and rewards of ownership have passed: it is hard not to satisfy (b) where (a) is satisfied;

(8) The third criterion (IAS 18.14(c)), which requires that the revenue can be measured reliably, is a matter of fact which is usually determined by the agreement, and in the case of a software sale is usually the sale price;

(9) The fourth criterion (IAS 18.14(d)), which is that it must be probable that the economic benefit associated with the transaction will flow to the entity, requires a judgement as to the probabilities of whether it is more likely than not;

(10) The fifth criterion (IAS 18.14(e)), which is that the costs incurred or to be incurred in respect of the transaction can be measured reliably, is a question of fact and is not disputed in this case.

Inevitably perhaps, given the difficulties inherent in the application of IAS 18.14, especially in the context of software transactions, there were in other contexts fundamental differences both in the approach of the two experts and in their evidence as to accountancy practice.

Mr Holgates ’ approach on VARs in his reports

Mr Holgate ’s views on the accountancy issues raised by the VAR claims were built around 13 Assumptions of fact that he was instructed to make in his report. These Assumptions reflected the Claimants’ pleaded case, and it is worth reciting them since they demonstrate both his approach and analysis and also what the Claimants sought to prove. As set out in Mr Holgate ’s first report , they were as follows:

(1) “Assumption 1: There had been no communication between Autonomy and the VAR relating to a transaction involving the identified end-user until immediately prior to the end of the relevant quarter;

(2) Assumption 2: There was no price negotiation between the VAR and Autonomy;

(3) Assumption 3: The VAR had made no prior efforts to sell such a licence to, and had no prior relationship or contact with, the identified end-user;

(4) Assumption 4: The VAR did not undertake or propose to provide any added value, or any service, to the end-user;

(5) Assumption 5: For VAR transactions involving the sale of a licence to use Digital Safe software (and software to be used with Digital Safe), the Digital Safe software could only be implemented and thereafter operated by Autonomy (and not by the VAR);

(6) Assumption 6: The VAR did not have the means to pay the Autonomy group company in the absence of an onward sale of the relevant licence to the identified end-user;

(7) Assumption 7: For sales to Capax Discovery as VAR: Capax Discovery was a newly incorporated company in March 2009 and therefore had no financial history at that time. Capax Discovery wrote to Autonomy in March 2009, providing financial information for Capax Global “on the express understanding that Capax Global is a separate and distinct entity from Capax Discovery. All contractual obligations will be between Capax Discovery and Autonomy only”;

(8) Assumption 8: The VAR did not, after the agreement between Autonomy and the VAR had been entered into, make any effort to sell a licence for the relevant software to the end-user. Instead, the Autonomy group company continued its own efforts to achieve a sale of the licence directly with the end-user (and without consultation with the VAR);

(9) Assumption 9: The purchase orders or sales agreements for the transactions between Autonomy and the VAR specified that the software was for onward licensing to the particular end-user;

(10) Assumption 10: There was an agreement or understanding (whether or not legally enforceable) between the Autonomy group company and the VAR, which was not apparent on the face of the written contractual documentation between the Autonomy group company and the VAR, to the effect that the VAR would not be required to satisfy any liability to Autonomy from its own resources;

(11) Assumption 11: The VAR was relieved of its ostensible liability to pay the price for Autonomy software licence it had purchased by one or more of the following means: (a) the purported sales agreement between Autonomy and the VAR being cancelled, (b) a credit note being issued to the VAR discharging its ostensible liability to pay the price, or (c) the VAR ’s debt being written off.”

(12) Assumption 12: Where the Autonomy group company subsequently achieved a direct sale to the end-user, the Autonomy group company arranged for the end-user to pay the VAR so that the VAR could then pay the relevant Autonomy group company;

(13) Assumption 13: An Autonomy group company was caused to make a payment to the VAR to purchase rights, goods or services that the Autonomy group company did not need (and which had no discernible value to it), but which had the purpose and effect of putting the VAR in funds which it then used to pay for the Autonomy software licence.”

Mr Holgate accepted that none of the 37 transactions featured all 13 Assumptions. However, he considered it clear that:

(1) A VAR sale to which Assumptions 1 to 4 (relating to the substance of the transaction) and 9 (which he described as showing that “…the sale to the VAR was not the end of the story as far as Autonomy was concerned; indeed it suggests that the sale to the VAR is not an important part of the story at all” ) were factually established to be applicable was “not genuine and lacked substance”;

(2) If Assumption 8 was shown also to apply, the VAR transaction would have been demonstrated not to “affect the role and responsibility of Autonomy” and thus would “fail to meet IAS 18 paragraphs 14(a) (risks and rewards) and 14(b) (managerial control)” ; he added that it would also show lack of substance; and

(3) If a side agreement (whether legally enforceable or not) making clear that in practice there would be no need for the VAR to pay Autonomy from its own resources so that Assumption 10 applied, that of itself would demonstrate that at best the sale was conditional, and in any event that the transaction lacked substance, so as to disqualify revenue recognition. He concluded that “the 30 transactions that feature assumption 10…lacked substance and so revenue should not have been recognised in respect of them” and that such transactions would also fail to satisfy each of paragraphs 14(a), (b) and (d) of IAS 18. Assumption 10 was thus the only one which was conclusive if it applied.

In Mr Holgate ’s assessment, Assumption 10 applied to all 30 of the impugned VAR sales where there was some form of side agreement; and in 26 out of 30 Assumptions 1 to 4 and 8 and 9 applied also. He considered the conclusion that those impugned VAR sales lacked substance and failed to meet the criteria was further reinforced when account was taken to what happened after the ‘sale ’, principally because:

(1) None proceeded to a sale by the VAR to the end-user;

(2) In none was the VAR required to pay from its own resources, the VAR being held harmless in each in various ways;

(3) 19 of the 30 impugned VAR sales were ultimately concluded between Autonomy directly with the end-user after negotiations between them in which the VAR played no part;

(4) 11 of the 30 impugned VAR sales never proceeded to any sale to an end-user, but the VAR was in these transactions also held harmless from loss in a variety of ways.

Furthermore, as regards Assumptions 5, 6, 7 and 11 to 13, all of which related directly to the revenue recognition criteria in IAS 18, Mr Holgate ’s approach and assessment were in summary to the effect that:

(1) If, as he was instructed to assume by Assumption 5, Digital Safe software was of no value to the end-user absent Autonomy ’s implementation and operational services, any VAR transaction which involved the licence sale of Digital Safe software without such services failed the criteria in paragraphs 14(a) and (b) of IAS 18 because Autonomy necessarily had to and did retain both significant managerial involvement/control and the risks and rewards of ownership. The relevant transaction also thereby lacked any real substance.

(2) If, as he was instructed to assume by Assumption 6, the VAR did not have the means to pay in the absence of an onward sale, then in a situation where the VAR was not in reality ever intended or intending itself to sell to an end-user, it could not be said that payment by the VAR was “probable”, so that the transaction failed to meet IAS paragraph 18.14(d). Mr Holgate ’s view was that this rendered revenue recognition improper in the case of 19 of the 36 impugned VAR sales.

(3) Likewise, if, as he was instructed to assume by Assumption 7, Capax Discovery was a newly incorporated company with no trading history, Mr Holgate considered that payment could not be said to be “probable” so that again, the criteria in IAS 18.14(d) was not satisfied.

(4) If, as he was instructed to assume by Assumption 11, the VAR was relieved of any liability to pay by dint of (a) cancellation of the sale or (b) the issue of a credit note or (c) the debt simply being written off in each case at the instance of Autonomy, Mr Holgate considered that this similarly resulted in failure to satisfy IAS 18.14(d), and also IAS 18.14(a). It would also reinforce the point about lack of substance.

(5) If, as he was instructed to assume by Assumption 12, where the Autonomy group company achieved a direct sale to the end-user, it would arrange for the end-user to pay the VAR so that the VAR could then pay the Autonomy group company, Mr Holgate considered that the VAR transaction would fail to meet the criteria in IAS 18.14(a), (b) and (d) and the test of substance.

(6) If, as he was instructed to assume by Assumption 13, the Autonomy group company was caused to make payments to the VAR to purchase goods or services that the Autonomy group company did not need and which had no discernible value, the payments would be in the nature of disguised gifts and, taking the two transactions together the Autonomy company would have made no net gain, and Mr Holgate considered that the transaction failed to meet IAS 18.14 (a) and (d).

(7) Assumptions 11, 12 and 13 could, in Mr Holgate ’s view as stated in his first expert report, be regarded as “different ways of putting assumption 10 into effect” : Mr Holgate considered that one or more of those three Assumptions applied to 36 (that is, all but one) of the impugned VAR sales. He accepted that all three related to events that occurred after the VAR sale date but considered that it was permissible and appropriate to take them into account as evidence of the understanding referred to in Assumption 10. In any event, after the first few such transactions the pattern would not any longer be hindsight: it would be plain to see before the sale date.

Mr Holgate did not come through unscathed from cross-examination. Mr Holgate ’s evidence in cross-examination was notably more equivocal than his views as expressed in his first report; and it seemed to me that this was principally because, when cross-examined, he had to take into account nuances, and the actual documentation, for which the Assumptions allowed no room. Indeed, he accepted that “in the real world, [the auditor] would find out additional information, of course” and that the conclusions he had reached in relation to particular transactions were the product of his Assumptions:

“If I was working with actual facts and seeking to establish what they were, then I agree I would need to look at the master agreement and as many other pieces of information as I could find. But if I ’m working with assumed facts as given to me, then that ’s the basis on which I ’ve been asked to provide my opinion”.

Thus, when asked to consider factual variations or practical nuances not comprehended in the Assumptions he frequently had to resort to the explanation that he had simply answered on the basis of the assumptions he had been instructed to make, taking each to have been established. He contrasted this with “the real world”, and accepted that in reality, the scope for (a) questioning, resulting in further information giving a more complete picture and revealing nuances not catered for in the assumptions, (b) valid commercial justification of which he was not aware in a field of which he had no experience ^{^[264]} , and (c) the fairly broad margins within which a proper judgment might be made, all could well result in a different and perfectly proper answer.

For example as regards Assumptions 1 to 9:

(1) Mr Holgate accepted that Assumptions 1 and 2 at most would have led to further enquiry, and even together did not undermine revenue recognition;

(2) He struggled to explain any rationale for thinking that Assumption 3 supported an unequivocal conclusion, and retreated into a weak and unconvincing assertion of gut feeling, which seemed contrary to the “sell in rather than sell through” approach of IAS 18, that:

“…the real customer is the end-user and not the VAR. The end-user is the party who needs the goods, who wants to buy them and use them…And that therefore feels like, to me, in terms of substance, feels like the real transaction rather than the sale to the intermediate party.”

(3) He sought to rely in the same context and more generally on the premise that the fact that the VAR may in certain cases have had no knowledge of and no relationship with the prospective end-user was “a very unusual situation”. But when asked on what basis of comparison he felt able to say this, he had to accept that he had provided no evidence of whether this was unusual by reference to Autonomy or the market but instead inferred it to be unusual because:

“if all transactions were like these impugned transactions, such that Autonomy sells to the reseller but then the reseller doesn ’t sell in any case on to the end-user, that ’s no business - that ’s no foundation for a business.”

(4) Mr Holgate regarded Assumption 4 (that the VAR did not undertake or propose to provide any added value or service to the end-user) as signifying either “irrational commercial behaviour” or simply a means for Autonomy to buy revenue recognition at the cost of margin which otherwise it could have retained for its own benefit; but he was constrained to accept that taken together, Assumptions 1 to 4, if established on the facts, would not have caused Deloitte ’s decision to approve revenue recognition to have been such that no reasonable accountant could have reached it (even if he himself might not have taken the same view).

(5) Similarly, in his first report, Mr Holgate stated that Assumption 9 (that the VAR agreement specified that the software licence sold to the VAR was for onward licencing only to a particular end-user), though on its own not strongly indicative of a lack of substance, strengthened his conclusion that he had already reached on the first four Assumptions that the VAR sales lacked substance and were artificial. But under cross-examination he accepted that “in the real world” (which he added “is different from an exercise based on assumptions given to me” ) this was simply another factor which would prompt the auditor to ask further questions, and he did not exclude the possibility that the outcome would substantiate the VAR transaction.

(6) Mr Holgate came to accept that Assumption 5 (concerning VAR transactions involving the sale of a licence to use Digital Safe software for operation by Autonomy and not the VAR, as to which see further below) would have no real significance for the purposes of revenue recognition.

(7) Even in respect of Assumption 6, relating to the VAR not having the means to pay in the absence of an end-user sale, Mr Holgate accepted that the issue was “ fact-sensitive” and that it could be quite permissible to take into account the contract receivable ultimately due from the end-user, though he sought to attenuate this response by suggesting (rather unconvincingly, to my mind) that if the VAR had very little to do with the end-user sale, the receivable might in some way (unexplained) be devalued. His evidence on this was as follows:

“Q. Would you agree that it would also, in making that assessment, be permissible to take into account the likelihood of the VAR deal closing with the end-user as a way of generating funds to pay Autonomy? I ’m not saying it would necessarily be the only source but do you accept that that ’s something that could permissibly be taken into account in reaching an overall assessment?

A. Yes, you can take that into account. Clearly if the VAR has pre-sold to the end-user, then that ’s very definitely helpful of course because it ’s a contractual receivable for them. If they - but if that ’s not the case, then the question is how certain is the sale by the VAR to the end-user and that ’s - well you find out as much as you can about that. But if, for example, Autonomy is pursuing the sale and the VAR isn ’t, then that will count against it, but if the VAR were very active in making the sale and pursuing the deal, then that would count towards it.

Q. So again it ’s quite fact-sensitive, is that fair?

A. Yes.”

(8) Similarly with Assumption 7, in relation to Capax Discovery ’s status as a newly formed entity with no financial history and which Mr Baiocco had expressly stressed was distinct and separate from Capax Global, Mr Holgate accepted that the extent to which Autonomy would properly be able to take comfort from the position of that company within the Capax Group was a matter of fact, prompting further enquiry and then a conclusion.

(9) As to Assumption 8, that the VAR did not, after its agreement with Autonomy, make any effort to sell on to the end-user and left that to Autonomy, Mr Holgate opined in his first report that it was:

“strong evidence that [Autonomy] had not transferred to the reseller the risks and rewards with respect to licences that the reseller purported to purchase”

and also that:

“a vendor who continued the sales effort would clearly have retained managerial involvement in and control over the goods in question.”

Under cross-examination, however, Mr Holgate had to accept that the Assumption may well have introduced inadmissible hindsight and that in any event, and albeit “a strange thing to do” , which would raise questions an accountant might wish to explore and take into account, it would not necessarily prevent revenue recognition, if, for example, it was explicable on the basis of Autonomy having the better relationship with the end-user, and thus the far stronger prospect of sealing the deal.

(10) Assumptions 11, 12 and 13 were similarly relied on by Mr Holgate in his first report as both reinforcing his conclusion as to lack of substance and also as demonstrating that the impugned VAR deals failed to meet IAS 18.14 (a) and (b) and probably also (c). However, under cross-examination, Mr Holgate accepted that all three Assumptions 11, 12 and 13 concerned later events and described an outcome which could not have been known to the accountants at the time and therefore could not have affected the accounting judgment. He sought to rationalise these as in effect supporting Assumption 10:

“I think all of those, to my mind, are methods of dealing with or implementing assumption 10, which we ’ve discussed, is a point for the time of the transaction; 11, 12 and 13 are the playing out of that subsequently…

…

…all of them, as I say, are the playing out of assumption 10 which was a feature at the time.”

Mr Holgate seemed ultimately to consider the force of the above Assumptions as being in their cumulative effect, and the picture it built up, even if (as eventually he accepted) none individually were conclusive. Further, his acceptance, under cross-examination, that Assumptions 1 to 4 and 9 did not, even taken together, necessarily cause the decision in this case to recognise revenue to be outside the ambit of what may constitute a reasonable judgement means that their proof cannot of itself avail the Claimants, though of course all must be weighed together in the balance. His concessions in cross-examination had the like consequence as regards Assumptions 6 to 8 and 11 to 13.

Even in relation to Assumption 10, which he regarded as the most important and the only potentially conclusive one of the 13 Assumptions (if established on the evidence), he was eventually driven to accept that the clash between the contractual stipulation against any side agreement, buttressed by the confirmation letters promising that there was none (neither of which, either the contracts or the letters, he had read at the time of his first report, which was most surprising), resulted in what he called “ an ambiguous situation” . In the course of his cross-examination, I asked him to clarify this for me:

“ MR JUSTICE HILDYARD: If the salesman said, ‘I don ’t think you ’ll have to pay for this if everything goes wrong but I want you to understand that whatever I say can never be enforced ’, you would be out on the whims of the law, what would the accountant, the auditor, say?

A. My Lord, it ’s a tricky one. If the salesman is saying both sides of those things, ‘I want you to understand you won ’t have to pay but that ’s not a legal…’ then the understanding that the customer gets is - well, is less –

MR JUSTICE HILDYARD: The understanding he gets is he ’s on a wing and a prayer.

A. Yes

MR JUSTICE HILDYARD: But the wing and the prayer will do, will it, to undo the transaction?

A. I think in practice you would look at other transactions and see what happened in the past, see how these things have turned out, that ’s what you would do in practice. But, yes, it ’s an ambiguous situation…

MR MILES: But in those circumstances, on the scenario that the judge has just posed, then although you said it ’s tricky, actually the truth is that the revenue would be recognised, wouldn ’ t it?

A. No.

Q. Do you accept that there ’s a range of possible views on that?

A. I accept that an accountant dealing with this in practice would do some digging to find out more surrounding facts, would look at past transactions to see if there had been similar circumstances to see what can be learned from that. Now, this might be a case where there is really quite a borderline difficult judgement and there may be different judgements reached reasonably by different accountants because I said earlier that, speaking in general terms about IFRS, there ’s a range of judgements that could be made and it ’s relatively narrow and not broad, it ’s not ‘anything goes ’. But what we ’re describing here I think is quite a narrow - it ’s quite a fine point. You could read it either way. It ’ s quite difficult, it ’s somewhere in the middle. So in this case I would agree with you that you could view that either way and still be…a reasonable judgement.”

This was obviously a significant concession, which also demonstrated a weakness in Mr Holgate ’s approach in his reports resulting from the reliance he was instructed to place on summary Assumptions, and the fact that in consequence he had not considered the actual facts (or even, as mentioned previously, the contracts or Deloitte ’s working papers). The Assumptions he was instructed to make were inevitably in synoptic form, which could not provide an accurate proxy for the complex and nuanced actual circumstances which both experts had agreed had to be considered in respect of individual transactions. Further problems include that such assumptions, especially when used to create a substitute or proxy universe, may box the expert in, or lead him or her, when asked to consider alternative factual hypotheses, to tend to struggle if they are in conflict with the assumptions given, and either retreat into them or seek to defend them as reasonable, which undermines their purpose. Also, for the same reason, or because they may also be couched in terms or even a particular sequence, they may encourage “ confirmation bias” and tend to suggest the conclusions sought by the party calling the evidence, which again may undermine the value and reliability of the opinion given.

In the end, although Mr Holgate had initially appeared to suggest that the criteria provided bright lines which could be applied to the Assumptions to provide a clear, almost scientific, judgement on whether revenue should have been recognised in respect of the impugned VAR sales, he accepted that he had not the information, and it was not really the function of an expert, to make such a binary assessment. Mr Holgate was led by the Assumptions into expressing conclusions of fact (which is not the province of an expert) by reference to contractual arrangements he had not studied and on information he accepted to be incomplete and insufficient.

It took him ultimately to accept that the issue came down to a judgement call as to whether, applying the criteria to the actual facts but also standing back to consider its true intended economic effect, the transaction had any real substance separate from any ultimate sale to an end-user. That would depend principally on (a) who was “the real customer” in the sense of turning the goods sold to use or account; (b) whether there was ever any intention for the reseller to negotiate and contract a sale to the end-user; (c) whether the parties truly intended the VAR to pay Autonomy from its own resources if no end-user deal eventuated, or whether instead they intended the end-user sale to be the source of funds to enable payment, failing which some alternative arrangement to remove or satisfy the VAR ’s outstanding legal obligation would have to be made. Put shortly: was the VAR intended to be a real or a nominal purchaser?

Despite the appearance of differences between them, it seemed to me that this chimed with Mr MacGregor ’s approach, to which I now turn.

Mr MacGregor ’s approach on VARs in his reports

Mr MacGregor (who was instructed by the First Defendant) had more practical experience of actual auditing (though latterly his experience was not field work but forensic accounting) than Mr Holgate, but his evidence had less analytical coherence. Especially in his answers when cross-examined, he was not always consistent. He often found the varying assumptions he was asked to make somewhat confusing, especially (as is understandable) when complex and cumulative, and he tended to complain in general terms or express confusion, rather than gather his thoughts and ask for clarification. Mr Miles interceded when absolutely necessary; but there were nevertheless times when it was not altogether clear to me what the emerging evidence was. Both in his appreciation of the question asked, and in his answers, Mr MacGregor tended to lack precision. He also occasionally appeared to lose his footing, and gave the impression of trying to remember what he had said in his report, rather than focusing on his answer to the question asked.

Mr MacGregor had, however, the advantage of having read the VAR agreements, and Deloitte ’s working papers. His approach was thus much closer to replicating how Deloitte did in fact approach the matter: it was closer also to what Mr Holgate accepted was the “real world” (distinguishing that from the exercise he was instructed to undertake himself). He was careful to distinguish between, on the one hand, matters of accountancy practice and approach, and on the other hand, issues of fact or legal interpretation or effect, which he left to the court to decide.

He considered that the effect of an entire agreement clause was a matter of law, and that “such a clause suggests that there would be no validity to any agreement which was not contained in the contractual documentation itself” . He appeared to accept that the law might altogether resolve the question, and he also acknowledged that in any event:

“while it is important to consider the substance of a transaction as well as its legal form, the contract terms can drive the accounting treatment and should not be ignored in determining the point at which revenue should be recognised and the measurement of revenue”.

However, Mr MacGregor did not suggest that there was any settled approach to the issue as to the interplay between contract and informal assurances in accountancy practice; and subject to the law having preclusive and conclusive effect (which he left open), he appeared to consider that assurances given outside the contract could and should be taken into account in reaching a view as to the true substance of the transaction, at least as to whether there had been any true sale (which he regarded as the real question, the provisions of IAS 18.14 being the means of reaching an answer), and that this was the case even if such assurances would be unenforceable in law.

Subject to his caveat as to whether in law the entire agreement clause was conclusive, Mr MacGregor appeared to accept that the recognition of revenue from the impugned VAR transactions was improper if, notwithstanding what the contract appeared to provide, there was no true sale in economic terms because neither party intended the fundamental aspects of a true sale (payment of the purchase price and transfer of management and control of the goods sold) actually to take place. In such circumstances, assuming that the true intent was that Autonomy should close a direct deal with the end-user and obtain actual payment from the end-user only at the point of that sale, Mr MacGregor regarded the only true sale in economic terms as being to the end-user, and revenue should only be recognised on completion of the sale to the end-user.

Mr MacGregor regarded the specific criteria set out in IAS 18.14 as indicia of an unconditional sale: a purported sale which did not satisfy those criteria could not be treated as a sale from which revenue could be recognised. But for him the specified criteria were not the only means of determining the question: if the economic reality was that the parties never intended the ‘sale ’ to have any economic consequences and was simply a nominal arrangement put in place pending a true sale to an end-user which the parties intended to be (a) negotiated by the original seller (Autonomy) and (b) the true source of payment for the goods, there was no need to look further. This echoed Mr Holgate ’s point that a condition of recognising revenue was that the transaction said to generate it should have economic substance. But Mr MacGregor ’s test of substance was more obviously based on the intention of the parties: if the parties did not intend to fulfil the obligations of a real sale, no revenue recognition could be permitted.

That focus on the real intentions of the parties was, as I understood his evidence, Mr MacGregor ’s central point. He did not regard this as conclusively determined by their contract, although what they had contractually agreed would be a very important factor in determining their overall intentions (for substantially the obvious reasons accepted by Mr Holgate). The following passage from Mr MacGregor ’s cross-examination seems to me to capture his approach (and a flavour of the way he explained it):

“ Q. Well, let ’s take a situation, Mr Welham in his evidence has told my Lord that even if it isn ’t legally enforceable, if in fact there is an oral agreement which stands next to the contract, you take that into account and Mr Holgate has said the same thing. You don ’t disagree with that, do you?

A. You would think about it and you would take it into account. Whether it has ultimately any relevance to the accounting is always going to be a matter - potentially a matter of judgement but potentially not a matter of judgement.

Q. You would think about it and it may well form part of the substance, regardless of whether it is legally enforceable?

A. If it ’s not legally enforceable, it is probably going to have - it ’s probably going to have limited use. I mean, as far as the oral side of it is concerned, I mean in the hierarchy that auditors, for example, use when trying to assess evidence, it ’s well known: information from the parties is more reliable. It ’s well known: information that is written rather than oral is much more reliable.

Q. If in fact there is an understanding and an agreement, it may not be legally enforceable but it exists and both parties know it exists and both parties intend actually to give effect to it in any event, regardless of whether it ’s legally enforceable, you would not disregard that when you ’re looking at the substance of the transaction?

A. Well, I think the keyword there is “ intention” and if at the time a contract is entered into and there is an intention to do one or more things and that intention is carried out, such that it overrides the terms of the agreement, then I can well see that that would be the case. In fact, I do say that in my report. The intention at the time of the contract has to be thought of as relevant there.

Q. And that is so, regardless of whether that is legally binding, that is to say you can come to a court and enforce that?

A. If one has a debt that one doesn ’t intend to collect, doesn ’t intend to collect, forget “I ’m going to give somebody time to pay”, just doesn ’t intend to collect, then that fails the definition of an asset because there is no future economic benefit going to come as a result of that debt. So that fails - that would fail at the first hurdle.”

Likewise, as regards the effect of all negotiations for the end-user sale being undertaken by Autonomy, without any involvement on the part of the VAR nor any suggestion of agency, Mr MacGregor stated in an Appendix to his first report:

“…if at the time of the sale to the reseller, Autonomy intended to continue to attempt to sell direct to the end-user, and if it intended to cancel the sale to the reseller (or otherwise relieve the reseller of the debt) on a subsequent successful direct sale (or no sale) then no revenue should be recognised in the income statement until such time as, for example, a sale to the end-user made probable the flow of economic benefits to Autonomy.”

He clarified in oral evidence that it was not sufficient for revenue recognition to be impermissible that ultimately Autonomy did deal and contract with the end-user: what did “invalidate the original sale” was “the intention to deal with the end-user, come what may…” . In a later passage in answer to a question from me he elaborated:

“…if it ’s the intention at the time of the sale to the VAR takes place, the VAR is going to do nothing, the VAR is just going to sit there and Autonomy is going to do as it was always doing and go in there, you know, continue to negotiate the price and the amount of software and all those sorts of things, then there ’s no sale. That ’s not a sale that Autonomy would be entitled to recognise.

However, if it ’s not the intention, if it ’s something less than that, then one would need to look at all the facts and that would be a consideration, possibly a material consideration…It ’s clear cut to me when there is a definite and clear intention; it becomes less clear when it ’s just an understanding.”

The distinction which Mr MacGregor appeared to draw between a settled intent never to require payment (leading to the conclusion that in substance there was no sale) and some understanding about deferral, falling short of a contingent contract, which Mr MacGregor considered to be consistent with a sale in substance and, therefore, to be “ fine” , was also clear from a later passage of his cross-examination when Mr Rabinowitz pressed him as follows:

“Q. If the side agreement or understanding and intention at the time of the contract, again to ensure that the VAR would not have to make a payment out of its own resources, was simply that if the VAR could not achieve a sale to a specified end-user, Autonomy would find and facilitate a sale by the VAR to some alternative end-user enabling the VAR to use the money so obtained to pay Autonomy, again you would say that transaction, when arranged or understood or agreed, simply lacked economic substance, correct?

A. No, I don ’t agree with that. I don ’t agree with that.

Q. Why not? Because?

A. For this simple reason. If I sell something to somebody and they don ’t currently have the means to pay but expect to have the means to pay in the future, I can recognise a sale. In this case, if I sell something to somebody with the intention that they sell on to the end-user and thereby will have the money to pay for the purchase, then that ’s fine. And if they fail to make that first purchase but Autonomy then arranges, manages to find a further, a second end-user they can sell to, I don ’t see what the problem is with that being a sale.

Q. I think you may have misunderstood the assumption . The assumption is that the nature of the arrangement made is that Autonomy says to the VAR, “All right, here ’s software to sell to that end-user but don ’t worry about it if you can ’t sell to that end-user, you won ’t have to pay us until - we will find you another end-user with whom there is a contract to sell, sell to that other end-user and then you can use the money from that other sale to pay us for the first sale” ?

A. No, I did understand that one and I think that seems to me to be fine.

Q. Fine?

A. Yes.

Q. How is that not the same as a situation in which the parties have effectively understood and intended that the reseller would not have to satisfy any liability to Autonomy from its own resources?

A. Because you didn ’t preface that example with the side arrangement that there is no intention to - on the seller to collect…”

What I took from all this was that neither forbearance after the event in pressing for payment, nor intervention to facilitate a sale to the end-user which was not premeditated, would undo the substance of the sale; but intention from the beginning to the same effect would. As it seemed to me, on Mr MacGregor ’s approach, two primary questions had to be raised and answered by reference to the detailed facts:

(1) which was, in reality, the transaction which at the time of the VAR ‘sale ’ Autonomy intended should generate its actual receipt of revenue: was it the VAR sale or the sale to the end-user?

(2) who, at the time of the VAR sale, did Autonomy intend should in fact negotiate and conclude the end-user sale?

If the answer to these questions was that Autonomy intended that any money it received would come from the end-user sale, rather than the VAR sale, and/or if Autonomy intended that it, rather than the VAR, would be the entity which would in fact negotiate and conclude the end-user sale, the conclusion would follow that the VAR sale lacked substance and the criteria in IAS 18.14 were not satisfied, so that no revenue could permissibly be recognised from the VAR sale.

In my view, these tests are in substance the same as the tests which Mr Holgate considered should be adopted; and the experts, albeit in different language and with different emphases, and with Mr MacGregor preferring the test of intention to that of objective substance, reached the same conclusion that (a) there would be no true ‘sale ’ and (b) the criteria in IAS 18.14 would not be fulfilled if those committing the parties to the transaction as a matter law intended at that time that:

(1) the VAR would not play any part in negotiations with the end-user;

(2) Autonomy was to negotiate and close a direct deal between itself and the end-user with a view to the VAR transaction being dissolved if that eventuated, subject to payment of a fee (a MAF) to the VAR for its trouble;

(3) the VAR would never be required actually to make any payment to Autonomy for the software ‘sold ’ whatever the contract might say unless and until funded by the proceeds of a sale to the end-user or credited by Autonomy in the same amount, or somehow put in funds by Autonomy itself.

(4) Factual analysis

As summarised earlier, the Claimants contended that the court should find that the arrangements between Autonomy and the VARs in all the impugned transactions followed a tell-tale ‘ pattern ’, which demonstrated the true intentions of the parties and the real economic substance and intended effect of the transactions.

According to the Claimants, this ‘ pattern ’ involved one or more of the following features which (they submitted) caused the economic substance and intended effect of the transaction to be very different than appeared from the unconditional sale (and transfer to the VAR of risks, rewards and managerial control) legally provided for by contract:

(1) the VAR made no real attempt to sell the software license on to the end-user, such efforts being a matter exclusively for Autonomy;

(2) the VAR was never required or pressed to pay for the software license from the VAR ’s own resources.

These two matters were presented by the Claimants as complementary. As it was put in their written closing submissions:

“…if the Court finds…that it was intended at the time of the VAR transaction that the VAR would do nothing, that fact is - in itself - a reason to infer that the VAR was also not intended to be required to pay Autonomy from its own resources.

The reason is simple. If the VAR was genuinely expected to pay Autonomy from its own resources, the most obvious way for the VAR to put itself in funds would be by working hard to secure an end-user sale.

Contrariwise, it would be most unnatural for a VAR - which knows that it is going to be compelled to make payment for a software licence imminently - to do absolutely nothing to sell that software licence on to the intended end-user and thereby obtain the funds with which to pay Autonomy…”

As to proof of the pattern and the intention of the parties thereby revealed, amongst the features characteristic of the impugned sales were the following:

(1) The sale to the VAR took place on the last day of a quarter.

(2) The VAR had had no previous contact with the contemplated end-user nor any information or knowledge about its intentions or its financial position and reliability from which it could gauge the prospect of an onward sale.

(3) The VAR never had any contact, still less undertook any process of negotiation, with the end-user.

(4) There was never any sale by the VAR to an end-user, and instead

(5) Either Autonomy eventually sold directly the same software (plus or minus extensions) to the end-user, or there was no end-user sale at all; and then (one way or another).

(6) Autonomy (whether expressly or implicitly) released the VAR from any further obligation or found some way of funding it to take the VAR off the legal hook.

(7) Although there was nothing in any of the written contracts entered into between Autonomy and the VAR which conferred any entitlement or expectation of a MAF (or any fee) Autonomy paid a MAF in every case in which an end-user deal eventuated, illustrating again that the parties paid only lip service (at most) to the provisions in the written agreement prohibiting side agreements and providing that the written contract should comprise the extent and entirety of their agreement.

On the basis of my view of the expert evidence and the proper approach to the Accounting Standards, these features and the pattern they reveal, if proven and shown to demonstrate the true intentions of the parties to the impugned VAR sale in question, would disqualify from revenue recognition any impugned VAR sales to which they apply.

The Claimants ’ case that in reality the VARs were never at risk and Autonomy retained managerial control

I turn then to analyse and assess, grouped by reference to particular resellers or VARs, in respect of each of the impugned VAR transactions, (i) the evidence of the persons who negotiated the impugned VAR transactions and thereafter (ii) how the Defendants depicted the individual transactions to Deloitte and thereafter to the FRRP and finally, (iii) the allegations that each of the Defendants had ‘guilty knowledge ’ of the impropriety of recognising revenue from the impugned VAR transactions.

Before addressing the factual evidence relied on by the Claimants in respect of each of the impugned 30 VAR sales (in respect of which the Claimants alleged there was a side agreement or understanding which negated their economic substance), it may assist to provide a brief overview and rough chronology of the arrangements concerned, the parties to them and the personalities involved.

The counterparties to the impugned VAR sales

Autonomy ’s principal counterparties in those impugned VAR transactions were as follows:

(1) MicroLink , which had been a reseller for Autonomy for many years prior to the beginning of the “ Relevant Period” in Q1 2009, and had entered into many VAR transactions with Autonomy prior to 2009, none of which are impugned. MicroLink had US Federal Government security clearance and by December 2008 had become Autonomy ’s primary reseller to the US Federal Government. MicroLink was the VAR in a series of reseller transactions in 2008 and 2009, of which the Claimants have impugned 11, with a variety of prospective end-users, including IBM-Ameriprise and the NSA (part of the US Federal Government). The Claimants referred to this series of transactions together as “VT1”. MicroLink was acquired by Autonomy at the end of 2009. The Claimants alleged that Autonomy ’s purpose in making this acquisition was to enable MicroLink ’s indebtedness “to be reclassified as inter-company debt and effectively washed away.” After that acquisition, Autonomy used MicroTech, which it had also used as a VAR since 2006, in its place: see paragraph 1963(3) below.

(2) Capax Discovery LLC , usually referred to as “Capax Discovery”, which Autonomy used as a VAR for 10 of the impugned VAR sales (and also another transaction said to be an improper linked or “ reciprocal” transaction, a category to which I return later). Capax Discovery LLC was established in early 2009, as a subsidiary of Capax Global LLC when Mr Stephen Williams, who had joined Capax Discovery in 2008 from Autonomy ’s e-Discovery division, persuaded Capax Global ’s management to expand into e-Discovery or electronic data discovery ( “EDD”) business. Both companies were usually referred to simply as “ Capax”. Capax Discovery LLC was also party to the EDD transaction which was impugned by the Claimants to which I also return later. The Capax Group was, by 2009, the largest professional service provider to Autonomy, and one of the largest professional service providers to Microsoft. Its business had largely been built around Microsoft, but Autonomy was its second largest partner. The 10 impugned VAR transactions involving Capax Discovery in the Relevant Period (Q1 2009 to Q2 2011), each of which took place under a VAR agreement between Autonomy and Capax Discovery dated 30 June 2009, were:

i. VT2, in Q2 2009, in respect of which the prospective end-user was a Texan electricity supplier called TXU Energy Retail (“TXU”), and for which the software licence fee was $783,086 (plus $78,309 for one year’s support) payable in three instalments. This was followed by a further transaction for a fee of $61,652 plus additional fees of $6,165 for support and maintenance and $395,023 for hardware.

ii. VT3, in Q3 2009, in respect of which the prospective end-user was Kraft (a well-known US grocery manufacturing and processing conglomerate, and by then a long-standing customer of Autonomy) and for which the VAR sale price was $4,000,000 and a further support and maintenance fee of $200,000.

iii. VT4, in Q4 2009, in respect of which the prospective end-user was Eli Lilly & Co (a large pharmaceutical company) and for which the VAR sale price was $5,986,827 and a further $299,342 support fee.

iv. VT10, in Q1 2010, in respect of which the prospective end-user was the UK Financial Services Authority (“the FSA”) and for which the VAR sale price was $4,285,714 plus a support and maintenance fee of $214,286.

v. VT16, in Q3 2010, in respect of which the prospective end-user was Amgen Inc (“Amgen”), a pharmaceutical company and for which the VAR sale price was $9 million and a further one-year support fee of $450,000.

vi. VT20, in Q4 2010, in respect of which the prospective end-user was Defense Knowledge Online (“DKO”) for the US Department of the Army, and for which the VAR sale price was $1,950,197, plus a support and maintenance fee of $292,530, payable in three equal instalments (in March, June and September 2011).

vii. VT21, in Q4 2010, in respect of which the prospective end-user was Merrill Lynch, by then a subsidiary of Bank of America and for which the VAR sale price was $1,830,600, including two years’ support and maintenance, payable in two equal instalments of $915,300, one within 90 days and the other within 180 days.

viii. VT27, in Q1 2011, in respect of which the prospective end-user was McAfee, and for which the VAR sale price was $5,000,000 plus a support and maintenance fee of $250,000, payable in two equal instalments (in July and September 2011).

ix. VT28, in Q1 2011, in respect of which the prospective end-user was UBS, the well-known bank and for which the VAR sale price was $8,000,000, plus an annual support and maintenance fee of $400,000, payable in two equal instalments (in July and end of July 2011).

x. VT34, in Q2 2011 (though allegedly not signed until 1 July 2011, the first day of the next quarter) in respect of which the prospective end-user was again UBS, and for which the VAR sale price was $7,664,132, plus an annual support and maintenance fee of $383,207.

(3) MicroTech , which Autonomy used as a VAR in eight of the 30 impugned VAR sales (in respect of which the Claimants’ alleged that there was a side agreement or understanding) and in one further impugned VAR transaction (VT5) where no side agreement or understanding is alleged, was (like MicroLink) used by Autonomy long before the Relevant Period. Also like MicroLink, MicroTech was a US Federal Government-approved “8A” reseller. The nine impugned MicroTech VAR transactions (including VT5 where no side agreement or understanding is alleged), all governed by a June 2006 MicroTech Master Agreement, were:

i. VT5 , in Q4 2009, in respect of which the end-user was DiscoverTech, and for which the VAR sale price was $9,523,810 plus a first-year support fee of $476,190. However, it is to be noted that no side agreement or understanding was asserted, and the impropriety alleged turns on what the Claimants termed “its own peculiar facts” relating to (a) the connection between VAR and end-user since MicroTech and DiscoverTech (and MicroLink) were “Truitt-related companies” and (b) a suggestion that VT5 was simply a means of returning to Autonomy $10 million of the purchase price of MicroLink.

ii. VT6, in Q4 2009, in respect of which the prospective end-user was Honeywell Aerospace (“Honeywell”) and the VAR sale price was $1,800,000 plus a first-year support fee of $90,000. Ultimately Honeywell did not conclude any end-user sale with either MicroTech or Autonomy.

iii. VT7, also in Q4 2009, in respect of which the prospective end-user was Manufacturers Life Insurance company (“ManuLife”) and the VAR sale price was $1,080,000 plus a first-year support fee of $104,000. Autonomy issued a credit note for those amounts and paid a MAF to MicroTech when, shortly afterwards (in March 2010), it entered into a larger, direct deal with ManuLife.

iv. VT8, also in Q4 2009, in respect of which the prospective end-user was Morgan Stanley and the VAR sale price was $4,888,800 ($4,656,000 plus a fee of $232,200 for support and maintenance). As in VT7, Autonomy issued a credit note to MicroTech and paid a MAF when subsequently (in March 2010) it entered into a direct agreement with Morgan Stanley for a higher amount.

v. VT13, in Q1 2010, in respect of which the prospective end-user was the Vatican Library and the (initial) VAR sale price was $11,000,000 plus a fee of $550,000 for support and maintenance. The end-user was planning to use IDOL to digitise the Vatican Library’s manuscript collection of over 80,000 manuscripts and over 40 million pages of documents, and this was a very prestigious and enormous project, which would (to quote Dr Lynch) “certainly have been a contender for” the biggest single deal ever done by Autonomy.

vi. VT25, in Q4 2010, in respect of which the prospective end-user was the US Department of Interior (“DoI”) and the VAR sale price was $4,000,000 plus $200,000 for support and maintenance. No end-user deal was ultimately achieved, nor did MicroTech make any payment either.

vii. VT32, in Q1 2011, in respect of which the end-user was Bank of Montreal and the VAR sale price was $2,880,000 plus $144,000 for annual maintenance and $50,000 for annual premium support. In June 2011, Autonomy closed a direct deal with Bank of Montreal for a licence fee of $2,800,000 and in August 2011 it issued a credit note to MicroTech for the amounts it owed under the VAR transaction.

viii. VT33, also in Q1 2011, in respect of which the prospective end-user was Xerox, and the VAR sale price was $1,170,000 plus $58,500 for support and maintenance. Ultimately, on 29 July 2011, an Autonomy group company called Verity Inc entered into a direct deal with Xerox for an amount of $1,300,000 including support and maintenance and a further fee of $14,175 for an additional Spanish module. The end-user direct deal with Xerox provided for payment to MicroTech as Autonomy’s designated payee and MicroTech then paid on to Autonomy but having deducted $85,675 as a fee.

ix. VT37, in Q2 2011, in respect of which the prospective end-user was HP (which hoped to provide the technology services to the United States Postal Service “USPS”) and the VAR sale price was $7,000,000 plus $350,000 for one year’s maintenance. In the event, no end-user deal was closed. MicroTech paid the full amount due under the VAR transaction in August 2011.

(4) DiscoverTech , which Autonomy used as a VAR in eight of the impugned VAR transactions, was newly formed by Mr David Truitt (Mr Steve Truitt’s brother) in 2009 as a vehicle to carry on the software business comprising an advanced information discovery and social networking product which was spun out of MicroLink just before Autonomy acquired MicroLink. (That acquisition and the prior spin-off are both also matters impugned by the Claimants and dealt with elsewhere in this judgment: my present focus is on DiscoverTech’s two subsequent VAR transactions with Autonomy which are also impugned.) The eight impugned VAR transactions between it and Autonomy were each governed by an individual reseller agreement, but the agreements were in materially the same terms with, in every instance, wide ‘entire agreement’ clauses. The transactions were:

i. VT11, in Q1 2010, in respect of which the prospective end-user was Citigroup (“Citi”, which was one of Autonomy’s large, long-standing clients), and the VAR sale price was $5,500,000, plus a first-year support fee of $275,000. The VAR sale comprised software only, DiscoverTech paid Autonomy 20% (over $2,000,000) upfront. It was intended that DiscoverTech should on-sell to Citi; but there were difficulties and delays in getting on to Citi’s approved vendor list. Ultimately, a tri-partite agreement was made, under which Autonomy was deemed to be receiving payment in respect of the software from Citi as agent for DiscoverTech (to which Autonomy also paid a MAF). However, a further wrinkle was that the direct sale was in fact of storage cells (Zantaz Digital Safe Smart Cells with uploaded IDOL software) rather than software alone: I address this in greater detail later.^{^[265]}

ii. VT12, in Q1 2010, in respect of which the prospective end-user was Philip Morris International (“PMI”), the tobacco company, and the VAR sale price was $4,185,000, plus a first-year support fee of $209,250. Initially, PMI wanted to use its partner SHI for the deal and Discover Tech expected a purchase order from SHI. But PMI then changed their mind; and ultimately, a direct deal was made between Autonomy and PMI, following which Autonomy issued a credit note to DiscoverTech and paid it a MAF reflecting the margin DiscoverTech would have got had the end-user signed with them.

iii. VT23, in Q4 2010, in respect of which the prospective end-user was Bank of America (“BofA”), and for which the licence fee was $3,500,000 plus a support and maintenance of $175,000. The deal ran in parallel with VT24 (see paragraph 1963(4)(iv) below). Both were part of a larger overall deal for end-users BofA/Amgen/Merrill Lynch (the last two being BofA group companies), with a total licence value of $21,330,600 which was parcelled out to various resellers (MicroTech, Capax Discovery and DiscoverTech) in impugned transactions VT16, 21, 23 and 24). Ultimately, Autonomy made a direct deal with the end-user for a total licence fee of $19,500,000 in respect of all four transactions, with MicroTech as designated payee, and arrangements for MicroTech to account to Discover Tech and Capax Discovery for their relevant parts.

iv. VT24, in Q4 2010, in respect of which the prospective end-user was also BofA, and for which the license fee was $7,000,000, plus a support and maintenance fee of $350,000 (which was part of the larger overall deals referred to at paragraph 1963(4)(iii) above).

v. VT30, in Q1 2011, in respect of which the prospective end-user was Prisa, a Spanish and Portuguese language media group and an existing Autonomy customer before the VAR transaction. The VAR transaction was for a licence fee of $3,600,000 plus a first-year support fee of $200,000. Ultimately, no end-user deal was concluded with Prisa, either by DiscoverTech or Autonomy. The Claimants accepted that DiscoverTech satisfied its obligations but allege that this was only because it was enabled to do so by a purchase from DiscoverTech by Autonomy of a product called DiscoverEngine. This product was, amongst other things, a connector to Microsoft’s SharePoint product (which is a content management system in wide usage in enterprise environments) and had been developed to support and enhance Autonomy’s IDOL software for IDOL customers who were also SharePoint users. The Claimants make no claim that the transaction was reciprocal or linked, and no FSMA breach or misrepresentation is asserted, nor is it suggested that the transaction resulted in losses. However, the Claimants suggested that Autonomy had another similar product already and that it had no use for DiscoverEngine: the Defendants rejected that.

vi. VT31, in Q1 2011, in respect of which the prospective end-user was ThinkTech Inc. (“ThinkTech”), an associate of the brokerage firm TD Ameritrade, and the VAR licence fee (for an amendment to ThinkTech’s existing Digital Safe hosting arrangement) was $1,800,000 plus a first-year support fee of $180,000. No end-user sale eventuated, as the Claimants maintained was inevitable since TD Ameritrade had even before the VAR transaction indicated that it had no intention of proceeding; and the way in which the VAR deal was (to quote the Claimants) “unravelled” is also contentious.

vii. VT35, in Q2 2011, in respect of which the prospective end-user was Abbott Laboratories (“Abbott”), a healthcare company, and the VAR licence fee was $8,611,011.07, plus $388,988.93 in respect of hosting and a first-year support and maintenance fee. A small direct deal was concluded between Autonomy and Abbott in July 2011 for a fee of $600,000; but the large end-user deal never eventuated. The Claimants alleged that Abbott’s General Counsel had vetoed the transaction and that there was never any real prospect of an end-user sale.

viii. VT36, in Q2 2011, in respect of which the prospective end-user was Hyatt and the licence fee was $5,333,914 plus a first-year support fee of $266,696. Under the deal, DiscoverTech was to sub-licence to Dell, which in turn would sub-licence to Hyatt. Again, no end-user deal was concluded; again, the Claimants claimed that this was all but inevitable, since before the VAR transaction Dell/Hyatt had “said no”, VT 36 was also one of the deals which the Claimants claimed was “unravelled” and the debt written off to protect DiscoverTech and honour the alleged side agreement or understanding.

(5) FileTek, which specialized in the archiving of structured data, was the developer of the StorHouse and Trusted Edge software. It entered into a single VAR transaction with Autonomy (VT18). The VAR licence fee was $10,000,000, plus a first-year support fee of $500,000, and the prospective end-user was the United States Department of Veterans Administration Authority (“USDVA”). USDVA were a repeat customer of Autonomy’s and had been working with Autonomy on an email archiving system. In the event, no end-user deal was concluded with USDVA by either FileTek or Autonomy. FileTek paid Autonomy in tranches, including $500,000 on the same date that the VAR transaction was concluded, with further payments of $1.5m in March 2011, $1m in April 2011 $1.5m, in June 2011 and the remaining $6m in August 2011. However, the Claimants alleged that these payments were funded out of the funds paid by Autonomy to acquire StorHouse, which they further alleged was a ‘reciprocal’ transaction for which Autonomy had no need, as I elaborate later.

The Claimants provided a useful summary of the quarterly distribution of the impugned VAR transactions in the “Relevant Period” (Q1 2009 to Q2 2011) and the licence revenue attributed to those transactions in aggregate in each quarter: I append that as Appendix 3.

Summary of the Claimants’ overall approach in presenting the factual evidence

There were six principal elements in the Claimants presentation of the factual evidence:

(1) Witness evidence (what the Claimants called “direct evidence” ) on the key points as to (a) the alleged understanding or intention that the VAR should not be involved in any way in the end-user sale envisaged by the VAR transaction and (b) the alleged understanding that the VAR should not be required to pay from its own resources;

(2) Circumstantial evidence of a characteristic ‘ pattern ’ in the case of the impugned VAR transactions (other than the ‘Collectability VARs ’ ), alleged to be commercially and rationally explicable only by reference to a shared understanding or intention that (a) only the seller (Autonomy) and not the VAR was to be involved in any way in the end-user sale envisaged by the VAR transaction and (b) the VAR should not be required to pay from its own resources, and only pay if and when paid by the end-user;

(3) Evidence as to the payment of MAFs in circumstances where (a) the VAR had had no involvement in marketing and (b) the VAR agreements made no contractual provision for the payment of MAFs;

(4) Evidence of efforts to suppress the truth as to the ‘ pattern ’, what the VARs had been given to understand, and the basis for the payment of MAFs, including alleged efforts to mislead both (a) Deloitte and (b) the the FRRP/FRC;

(5) Alleged inconsistencies in Dr Lynch ’s attempts to resist such inferences; and

(6) Evidence of the involvement and true objectives of both Mr Hussain and Dr Lynch and their ‘ guilty knowledge ’.

I turn to discuss in turn each of these elements, and the Defendants ’ case in relation to them, in general terms.

“Direct evidence” of the alleged side agreements or understandings

As to (1) in paragraph 1965 above, the Claimants relied on “direct” or witness evidence in relation to the central allegation that there were consensual side agreements in respect of each of the impugned VAR transactions (except the ‘pure collectability VARs ’ ) which were such as to undermine “the legal obligation to cough up” (to adopt Mr Miles’ phrase in his opening) and to reserve to Autonomy the exclusive role in negotiating and closing an end-user deal which would in fact be its real source of revenue.

“Direct evidence”: the relevant witnesses

The Claimants relied primarily on what they depicted in the case of each of the impugned VAR transactions (other than the ‘pure Collectability VARs ’, (VT9, VT17, VT19, VT22 and VT29 where the issue is only as to the credit-worthiness of the VAR) as:

“a wealth of direct evidence to the effect that, at the time the VAR transaction was entered into, no-one - Autonomy included - intended that the VAR should be involved in any attempt to on-sell the software licence to the end-user.”

The Claimants submitted that much of the witness evidence (and particularly that of Mr Egan and Mr Baiocco) to the effect that the VAR was often, indeed usually, told that it was not expected and not encouraged to participate in, or have anything at all to do with, Autonomy ’s sales efforts was unchallenged.

The Claimants sought to rely on this also, as well as direct oral evidence, to support what they presented as the second limb of their case, which was that it was orally agreed or understood, and in any event undoubtedly intended between Autonomy and the VAR, that Autonomy would not require the VAR to pay the licence fee from its own resources if no end-user deal eventuated.

The order of these points is noteworthy. Both experts regarded the primary question in relation to revenue recognition under IAS 18.14 as being whether the seller (Autonomy) had transferred to the buyer (the relevant VAR) “the significant risks and rewards of ownership of the goods” (IAS 18.14(a)). They regarded the further criterion in IAS 18.14(b), that the seller should have retained “neither continuing managerial involvement to the degree usually associated with ownership nor effective control over the goods sold”, as being a subsidiary matter which was almost invariably determined by the answer to the first question. However, the Claimants reversed that sequence in their presentation of the evidence. The Claimants also submitted that a finding that it was never intended that the VAR, but only Autonomy, should make any attempt to on-sell the goods to the end-user, would “by itself be sufficient to make good the Claimants ’ case that no revenue should have been recognised at the point of sale to the VAR.”

The Claimants described the further finding that they sought, that it was orally agreed between Autonomy and the VAR that Autonomy would not require the VAR to pay the licence fee from its own resources, as “more controversial as between the parties”. But they submitted that the direct evidence was in this context also “ compelling”.

Overview of the witness evidence put forward by the Claimants

The witnesses who gave oral evidence in relation to one or more of the impugned VAR sales were:

(1) As the person said to have reached on behalf of Autonomy the side agreements or understandings with the VARs: Mr Egan (Autonomy), who gave evidence by witness statement, on which he was cross-examined by video-link;

(2) On behalf of various VARs: Mr Baiocco (Capax Discovery) and Mr Szukalski (FileTek), both of whom were cross-examined at the hearing on their respective witness statements; and

(3) As regards the information provided to and the understanding of Deloitte: Mr Welham of Deloitte (no one else from Deloitte gave evidence at this trial).

In addition, the Claimants relied on:

(1) The evidence of Mr Loomis of FileTek in the US criminal trial, introduced by the Claimants under a hearsay notice;

(2) The evidence of Mr Steve Truitt of MicroTech in the US criminal trial and in civil proceedings brought by MicroTech against Autonomy in the US (the “MicroTech litigation”) introduced by the Claimants under a hearsay notice;

(3) The evidence of Mr Tomas Esterrich (MicroTech ’s CFO) in the MicroTech litigation in the US, also introduced by the Claimants by hearsay notice; and

(4) The evidence of Mr David Truitt of DiscoverTech and MicroLink in the US criminal trial, and in the MicroTech litigation, likewise introduced by the Claimants by hearsay notice in the same way.

Both the Claimants and the Defendants in their written closings submissions provided first, a general overview of the witness evidence (both oral and hearsay), before thereafter describing in more detail the series of impugned VAR transactions and the evidence of the actors in relation to them. I adopt the same approach below.

Overview of Mr Egan ’s evidence as to what he told the VARs in every impugned VAR transaction

Mr Egan was presented as the Claimants ’ main witness on this aspect of their case and his evidence covered most of the impugned VAR transactions.

I have made an assessment above (see paragraph 426(5) to 426(7) above) of Mr Egan ’s reliability as a witness, and explained why I consider that he was in a difficult predicament: having entered into a Deferred Prosecution Agreement ( “DPA”) with the US DoJ) which committed him to prescribed evidence, honed after numerous interviews with the US DoJ, HP ’s lawyers and his own, his freedom depended on adherence to it; he was now being cross-examined in these proceedings, contrary to his expectation, but on video-link since he refused to attend in person; and he could not be required to attend in person, being “beyond the seas”. No doubt his oath weighed with him; but whilst it was at least unlikely that he could be subjected to proceedings here for perjury, departure from his script brought the real prospect of losing his freedom in the USA.

In his closing argument, Mr Miles remarked of his evidence:

“… The impression one gets - because he was reasonably candid a good deal of the time in cross examination –… is that this statement has been lawyered up and put in front of him and he was prepared to sign it because he was under the impression that he would not have to give evidence.”

I broadly agree with this assessment. Thus, for example, although Mr Egan gave the impression in his witness statement made in September 2018 that he could remember quite clearly what had happened in the reseller deals in which he was principally involved, a rather different picture emerged from his oral testimony given by video link. Under cross-examination it emerged that he could remember very little about the details of individual deals. Similarly, in the case of the MicroTech deals and the ATIC deal, it became clear that his witness statement was comprised of an orderly recollection of events carefully crafted by his lawyers which he could not in reality support from his own recollection.

Further, whereas in his witness statement the impression given was of frequent contact both with Dr Lynch and Mr Hussain at Autonomy and Messrs Steve and Dave Truitt, when cross-examined he had to accept that at least as regards Dr Lynch and Mr Steve Truitt his contacts were minimal:

(1) Mr Egan specifically corroborated Mr Steve Truitt ’s testimony to the effect that they had had only very occasional contact. He thought that one of the possible occasions was in the course of the Vatican deal, but even then Mr Egan did not feel “comfortable saying that absolutely” .

(2) He also confirmed that he mostly spoke to Mr David Truitt. However, Mr David Truitt was not the decision-maker: he was passing on information to Mr Steve Truitt and Mr Jimenez.

(3) As regards Dr Lynch, the impression he gave in his witness statement was of a line of direct report, with the natural inference that he had told Dr Lynch the various things that he knew; in cross examination he withdrew the very serious allegations he had made against Dr Lynch in that context and, for example, was unable to offer any real recollection or evidence at all against him in relation to the repeated allegations of “pre-textual emails” . (Mr Miles made the inevitable submission in this connection that all this raised “a real question mark over what on earth they were doing in his witness statement in the first place. It looks, I ’m afraid to say this, very much like something that was inserted by lawyers.” )

Nevertheless, in cross-examination, and as Mr Miles acknowledged, Mr Egan appeared to me to do his best to tell the truth, even if warily in respect of matters referred to in the statement he had agreed with the DoJ on which depended his freedom; and his depiction of the way he conducted Autonomy ’s VAR business in general terms struck me as reliable.

In that regard, Mr Egan confirmed that he had said substantially the same thing to the relevant VAR in each transaction, in accordance with instructions given to him by Mr Hussain, whom he described as “a very detail-oriented, hands-on manager” who “often tracked deals in great detail [and]… kept in touch with members of the sales force, directing which deals to pursue and what prices to negotiate” . Thus, he repeatedly cross-referred to what he described as “the guidance that Mr Hussain gave me as described in paragraphs 28 and 29 above” (of his witness statement) as a shorthand for describing what he had said.

In his witness statement, Mr Egan summarised in the following terms what he said Mr Hussain, as the person he identified as the architect of the impugned VAR transactions from whom he took his instructions (which was not challenged) , had told him he (Mr Egan) should tell the VARs:

“Mr Hussain defined the parameters of the practice; I implemented it. Mr Hussain ultimately determined which deals, and in what amounts, would be taken to a VAR, and which VAR to approach. In view of the significant financial cost of these deals (i.e. the fees paid to the VAR for taking them on), no one else (apart from Dr Lynch) had the authority to make that decision. Mr Hussain described the VAR deals to me as “acceleration deals”…

Mr Hussain gave me specific instructions to follow so that these deals would be accepted by Autonomy ’s auditors (Deloitte)…

Mr Hussain provided guidance to me regarding what was, and was not, acceptable to communicate in my conversations with VARs. He laid out explicit rules about what could be offered as incentive to the VARs, what was required of the VARs, and what could not be part of any deal. He instructed me to tell the VAR that in order for Autonomy to be able to recognise revenue, the VAR would have to sign a document that stated a binding obligation to pay for the software, that Autonomy would deliver the software to the VAR before the end of the quarter, and that there had to be sufficient evidence that the VAR could pay for the software regardless of whether a sale was made to the end-user. He also emphasized that it was vital that the VAR confirm to Deloitte that the VAR owed the money to Autonomy and intended to pay. I was also instructed by Mr Hussain to say, and did say, to the VAR orally that Autonomy would continue its efforts to sell to the end-user; often that the VAR was not expected to participate in those sale efforts; and, importantly, that Autonomy would do everything in its power to help ensure that the VAR would not be left “holding the bag”. On some occasions, Autonomy completed the sale to the end-user and caused the end-user to pay the VAR, which, in turn, allowed the VAR to pay Autonomy. On other occasions, if the end-user paid Autonomy directly, the VAR was relieved of its payment obligation.

…

At Mr Hussain ’s direction, I assured the VARs that if Autonomy was ultimately unable to close a deal with an end-user, there were various options that Autonomy had to “fix” the situation for the VAR so that it would not end up having to pay for the software from its own resources. Over time, this happened and Mr Hussain came up with a number of different ways of handling this, including buying products that Autonomy did not need from the VARs to offset losses from deals that we did not manage to sell through to the end-users. Mr Hussain made it clear that our assurances about the VAR not being left holding the bag could not be put in writing. We realized, of course, that if we left the VAR “holding the bag,” we would be unable to do future transactions with that VAR as it would probably ruin the relationship.”

However, Mr Egan gave little or no detail of any actual occasion, or exchange with the VARs, at which he allegedly communicated the assurances referred to above in accordance with Mr Hussain ’s instructions. He simply asserted that these were the assurances always given; and in the case of each transaction, he referred back to the general description as his evidence of the side agreement relied on.

As to the nature, rationale and perceived disruption of these arrangements, Mr Egan ’s evidence in his witness statement was as follows:

“The incentive for the VAR to take the license [sic] in order to help Autonomy reach its revenue goal for the quarter was that the VAR would be paid a “margin” or fee on the deal, which, typically, was 10% of the deal price. For its part the VAR had to sign a purchase order relating to individual deals and, if asked, to confirm to Deloitte that the VAR remained responsible to Autonomy to pay for the deal, and that there were no side letters or other agreements in place between Autonomy and the VAR. Therefore, although referred to as “ at risk” deals (because no deal had yet been concluded with the end-user), the VAR was not truly at risk of incurring any loss on the transaction given our pledge that Autonomy would do everything in its power to help ensure that the VAR would not be left holding the bag. I felt very personally obligated to make sure that the VAR would be made whole on the purchase. Any concern the VAR might have had at the outset about the possibility of Autonomy reneging on its pledge would have been dispelled over time by our actual practice of never in fact allowing a VAR to suffer a loss.

On several occasions, no sale was able to be made to the prospective end-user. I believe that each time this occurred, we found a “fix” for the deal so that the VAR did not have to pay for the software with its own funds…

The deals with the VARs were almost always entered into right at the end of each quarter, after Mr Hussain had determined that a sale to a particular end-user could not be completed in that quarter and when he was able to determine the size of the gap between that quarter ’s revenue target and the sales that had been made, or would be made to non-VAR customers before the end of the quarter. In my opinion, the practical effect of VAR deals of this type was to accelerate into the current quarter revenue that would otherwise have been recognised in a later quarter, assuming that a sale to the end-user could be made in a later quarter. This helped Autonomy to meet its revenue target for the current quarter. However, the problem this created for me was that, in the following quarter, I had to spend time and effort attempting to close the end-user transaction so that the VAR would be protected. I also had to make all of the sales in the following quarter that we otherwise needed in order to meet the following quarter ’s sales goal. In concept, we were borrowing revenue from a future quarter to achieve the revenue goal in the current quarter. However, in the following quarter, we had to close end-user deals just to enable the VAR to “pay back” the “ debt” incurred in the prior quarter.

This practice started on a relatively small scale involving a small number of end-user deals that were very likely to close in the near future. Over time, the number and size of the deals increased, and, in certain cases, the probability that the end-user deal would be completed decreased. The hole in which we started each new quarter - the implied obligation to find a way to complete the old end-user deals (or find some other solution) - grew larger and larger. At the same time, the revenue targets that had to be satisfied with new deals continued to increase. The ever-increasing revenue targets, in turn, created the need for yet more transactions of the type that I have just described and other revenue-generating tactics that I will describe…In my view, this pattern ultimately became unsustainable.”

Mr Rabinowitz submitted that important parts of Mr Egan ’s witness statement were not challenged in cross-examination. He instanced especially Mr Egan ’s evidence that, at Mr Hussain ’s direction, he often told the VAR that the VAR was not expected to participate in Autonomy ’s sales efforts. He also referred to and relied on Mr Egan ’s elaboration of his evidence in relation to particular VAR deals with favoured VARs following the prescribed ‘ pattern ’, including especially:

(1) Mr Egan ’s reference to a series of impugned VAR transactions with MicroLink which Mr Egan described as in each case being a “paper transaction”, in which MicroLink was simply required to submit a purchase order and only Autonomy was to be involved in pursuing and closing a sale to the end-user, whom Mr Egan described as Autonomy ’s “ true customer”;

(2) Mr Egan ’s evidence that in a VAR deal with Capax Discovery in anticipation of an end-user deal with Kraft (VT3), again all Capax Discovery had to do in return for a MAF of 10% was to issue a purchase order, sit back whilst Autonomy continued its efforts to close a deal with Kraft, and then charge a 10% fee;

(3) Mr Egan ’s evidence that at the end of Q4 2009 he entered into a number of supposedly “ at risk” deals with MicroTech (VT5, VT6, VT7 and VT8) “to get the revenue associated with the corresponding prospective end-user deals into the fourth quarter of 2009” in which MicroTech would be paid a fee for playing no active part in any negotiations with the end-user;

(4) Mr Egan ’s evidence that the like ‘ pattern’ was followed in the deal on 31 December 2010 with DiscoverTech in respect of a proposed end-user deal with BofA (VT23/24), in which again DiscoverTech was simply required, for its fee, to sit back whilst Autonomy sought to close the deal directly.

The basis for Mr Rabinowitz ’s submission that much of Mr Egan ’s evidence was not challenged is that he was not separately cross-examined on each of these transactions. However, Mr Egan confirmed in cross-examination that he said the same kind of thing to each of the VARs with whom he negotiated deals, though of course there would be different amounts of money, differences in software and license or other user terms, and different information as to the state of play with the prospective end-user, according to the particular facts.

I was therefore invited by Mr Miles to treat his explanation as applicable to all the impugned VAR deals (with appropriate variation according to the circumstances) and his references to specific VAR transactions as exemplars. On that basis, Mr Miles submitted that his cross-examination of Mr Egan was therefore appropriately confined to testing the assertion in respect of one transaction and treating that as a challenge to all other transactions where the same basic allegation was made that the impugned VAR transaction in question was (a) a paper exercise in which the VAR would (b) subscribe to an agreement for the sake of form but (c) take no part in the negotiation or closing of any end-user transaction and would (d) notionally, but not actually, take on risk. I accept that: Mr Egan had confirmed that he had given like assurances in every case, and a challenge to one was a challenge to all.

The focus of that cross-examination was on whether any of the assurances, if given, were intended to affect the legal obligation and indebtedness of the VAR. As to that, as Mr Miles emphasised in his closing submissions, Mr Egan considered that in every case the VAR in question was fully on risk, and nothing he said to individual VARs affected this. But this was addressing the legal position, which the Claimants did not dispute. It did not address what the parties intended would actually be their economic relationship, which was the focus of the Claimants ’ case.

In that context, Mr Egan ’s evidence was to the effect that he did indeed give assurances to the VAR that Autonomy would do its best to rescue them if the anticipated end-user deal failed, and indicated that Autonomy would forbear from enforcing the payment obligation in the meantime. In making clear that their legal risk remained, and that there was no binding commitment on Autonomy ’s part except as set out in the contract, the impression of the sub-text was unmistakeably that whilst the legal position had to be emphasised, of course, and lip-service must be seen to be paid, the reality was that Autonomy needed the deal and the VAR wanted the immediate business and the prospect of more. As the saying goes, ‘a nod ’s as good as a wink to a blind horse’. It seems to me to be clear that all those concerned understood, and the VARs proceeded on the basis, that the economic realities would have little to do with the definition of their nominal legal relationship, and Autonomy would not leave the VAR to shoulder any uncovered liability for acting in its nominal role.

Thus, when cross-examined, Mr Egan elaborated on what he meant with particular reference to the FileTek/USDVA transaction (VT18) in Q3 2010 , as follows:

“Q. Is this the position: that you made clear to FileTek that they would be on risk but that Autonomy would do what it could to make sure that in the end they would get paid?

A. They would be fully at risk but that we would use every effort to backfill that deal if for some reason that deal did not happen and not leave them holding the bag effectively.

Q. That idea of not leaving them holding the bag, that didn ’t affect, as you understood it from your discussion, the legal obligation on FileTek to pay even if the end-user deal was not done?

A. No, 100% not. You know, there ’s more detail to that in that I would let them know that they were signing up to buy this software, that they were buying it non-refundably, no recourse to not pay, and Autonomy was not bound to do what it would intend to do if it went poorly, but that it would be our intent and that we would use every effort to backfill. Nobody wants to just burn a partner and leave them in the dust. The point of this was to create channels of revenue that were forward-looking, that they would be incentivised and go out and sell the software other times.

Q. And they always understood that although you were giving that statement of intention, that wasn ’t in any way legally binding?

A. I made it clear that it was intent but that if, you know, Mike or Sushovan decided that they didn ’t want to do that, they would absolutely be nothing I could do about it and I was telling them that if we were acquired or if there was any other change, I describe the fact that it was our intent if they think the relationship will work that way then that ’s something that they ’d consider.

Q. The whole idea behind this was that risk had to pass in order to get revenue recognition, wasn ’ t it?

A. I actually didn ’t understand that, that that was required for revenue recognition, but I understood that it was an absolute requirement of Sushovan ’s that I impart the risk and that it be absolute.”

I understood that this was intended as an exemplar and an explanation of the position across the impugned VAR transactions as Mr Egan perceived it: and I consider the challenge to the exemplar to have been a challenge to all. So I do not consider that the challenge failed for want of being put in respect of every transaction; it failed because Mr Egan ’s answers made clear the formulaic nature of the emphasis on the legal position.

Mr Miles also relied especially on two passages of evidence given in the course of Mr Egan ’s re-examination by Mr Rabinowitz as being “really crucial evidence” :

(1) The first was this:

“Q. Mr Egan, can I ask you to explain further what you meant by “backfilling” and ensuring that the VAR was “not left holding the bag” ?

A. When I used the term “backfilling” what I was referring to was if the deal that the reseller had taken at risk somehow were not to be able to be closed, and Autonomy needed to make sure that the reseller wasn ’t - not receiving any revenue and making those payments to Autonomy, I was going to take other deals from Autonomy ’s forward-looking pipeline and then give them to that reseller to backfill that amount, in other words basically substitute another deal for it.”

(2) The second was this: when asked whether Mr Egan had discussed with the VAR whether, if the specified end-user dealer failed to eventuate, the VAR would be expected to make payment to Autonomy during the period before Autonomy had found a way to backfill the deal, Mr Egan answered:

“ Effectively, yes, they had to make their payments in line with payment terms they ’d committed to, independent of the time it took to decide whether to backfill, the time it took to partially backfill, independently”.

Mr Miles relied on this evidence for two points. First, he submitted that what Mr Egan had said amounted simply to an assurance that he would see what Autonomy could do to find another end-user in the event that the first end-user dropped out. Secondly, and a point which Mr Miles submitted was critical, Mr Egan made clear that in the meantime the VAR would still be “on the hook” to pay, or in another expression Mr Egan often used, left “holding the bag” pending identification and substitution of a ‘backfill” deal. Mr Miles submitted that this evidence:

“undermines their contention that there was any arrangement that the resellers would not have to pay or would not have to pay until they were paid. On the contrary, this is Egan saying that it was expressly discussed that they would have to pay even while they were looking for another end-user” .

Mr Miles also made the more general point that although this passage referred to the VAR deal between Autonomy and FileTek, Mr Egan had earlier confirmed that both the assurances he gave and his perception of their intent and consequences were the same in each of the other deals also. As Mr Miles summarised it (after a forensic swipe that “the Claimants themselves are guilty here of ignoring the substance” ) this evidence was that:

“what Mr Egan was emphasising to them was that the reseller was indebted to Autonomy whatever happened with the end-user.”

Mr Miles went on to suggest that this not merely challenged, but disposed of, any notion of some unilateral intention on the part of Autonomy that neither risk nor managerial control should effectively pass, since Mr Egan was the only person supposed to have generated that intention.

The Claimants sought to reply to these points by presenting Mr Egan ’s answers as simply reiterating the distinction, on which they relied, between the legal position (in respect of which the Claimants accept legal enforceability) and the practical reality, which they alleged to be what they described as his:

“unchallenged evidence that he assured the VARs that Autonomy ’s intention was not to leave the VAR holding the bag.”

I am entirely persuaded that Mr Egan gave these assurances, and that he did so with the sanction, and indeed on the instructions, of Mr Hussain. I cannot accept Mr Miles’ submission that Mr Egan ’s oral evidence critically affected his presentation of this in his witness statement. In my view, nothing that Mr Egan said in cross-examination or re-examination substantially qualified or undid the evidence in his witness statement that I have quoted at some length in paragraphs 1983 to 1985 above. Taken in the round, the gist of his oral evidence was to reiterate that the relationship with the VAR operated at two levels: what the contract provided and how the parties in fact intended to and did in fact conduct their relationship. For as much as Mr Egan stressed to the VAR that it was on the legal hook, as Mr Hussain had insisted he should, he also made clear that Autonomy would do “everything in its power to help ensure that the VAR would not be left holding the bag.” Nothing that Mr Egan said contradicted the reality that neither party expected the VAR to have any role, or to carry any real risk, or to gain any reward from any resale (except a MAF which the contract made no provision for).

Overview of Mr David Truitt ’s evidence as to MicroLink ’s role in impugned VAR deals

As explained in paragraph 1963(1) above, none of the many VAR deals between Autonomy and MicroLink entered into prior to the Relevant Period has been impugned in these proceedings. At one time (in October 2018), the Claimants confirmed that they were abandoning their allegation that there was a side-agreement in relation to the MicroLink transactions in the Relevant Period. That concession was then withdrawn a month later; yet Mr Welham ’s witness statement records that he was asked to assume that in only one of the six MicroLink VAR deals with Autonomy in Q2 2009 was there a side-agreement; and of the three impugned MicroLink deals in Q3 2009, Mr Welham was again asked to assume a side-agreement in relation to only one of them. That is the basis of the Defendants ’ submission that the Claimants ’ case in relation to MicroLink has been in flux as regards the basis on which they seek to impugn the MicroLink VAR transactions.

Mr David Truitt, a practiced prosecution witness who gave evidence at the US criminal trial, as well as in the MicroTech litigation, accepted in his deposition in the MicroTech proceedings that Autonomy ’s objective in its VAR sales to MicroLink was to “take on these end-of-quarter deals so that Autonomy could recognize revenue on those deals in whatever was the then current quarter, the quarter that was about to end.” However, he consistently made clear in both sets of proceedings that MicroLink acquired full ownership and full control of the software under the VAR transactions and was unequivocally “ at risk” . He denied that there was any side-agreement or understanding such as to undermine revenue recognition.

Though cautioning as to his reliability, the Defendants cited the following three passages in support of their contention that Mr David Truitt ’s evidence in earlier US proceedings undermined rather than supported the Claimants ’ case:

(1) In his examination-in-chief in the US criminal proceedings, Mr David Truitt confirmed that MicroLink ’s relationship with Autonomy was an “at risk scenario” such that he wanted to make sure that the deal with the end-user would close quickly:

“Q. What did you mean by "end-of-quarter scenarios"?

A. "End of quarter" for me would mean deals that Autonomy had been working on their own, that they would -- they'd be similar, you know, to the one I just described in the sense that they were supposed to be far down the line in the sales process, and they would ask whether we would be interested in issuing orders for those -- for those scenarios.

So they would present a -- you know, a particular opportunity, they would talk to us about what it was, where it was in the sales cycle, and then we would decide whether to issue an order or not.

Q. The $200,000 order that you described that started this end-of-quarter scenario, why did you call the customer about that?

A. I wanted to make sure that it was going to indeed close quickly and that, you know, the agreement with Autonomy on those deals was -- you know, once you issue the order, it's your order. So there was, you know, an at-risk scenario. So, you know, I didn ’t know Mr. Cronin very long at that point and wanted to make sure that what he was telling me was correct. I didn ’t want to be out 200,000.”

(2) Mr David Truitt gave similar evidence in relation to MicroLink in the US MicroTech litigation. This was marked up as hearsay by the Claimants and (as the Defendants pointed out) must be taken to represent their evidence:

“Q. And were there occasions when you did not collect from the customer?

A. Yes.

Q. Okay. What happened on those occasions?

A. On the occasions where the customer did not buy the software, we would try to take that software and sell it to another customer, and sometimes we were effective at doing that. You know, we had -- we had to figure out a way effectively to cover -- we would call these transactions at-risk transactions, which meant to us that we took them and we had to figure out if it went badly, you know, what to do.”

(3) In another passage from his evidence in the US criminal proceedings, this time taken from his cross-examination, Mr David Truitt gave similar evidence against Mr Hussain, explaining what he meant by an “ at risk” transaction (albeit in relation to DiscoverTech). He confirmed that DiscoverTech was at risk if the end-user deal did not close.

The difficulty with this evidence is that (a) according to Mr Egan, whose evidence on this point was not challenged, it was Autonomy, and not MicroLink, which in the transactions in the Relevant Period to which it was a party invariably and exclusively dealt with the end-user to seek to conclude the indicated end-user transaction; (b) no instance was provided of MicroLink seeking to sell on the software it had acquired under license; and (c) no evidence was provided of MicroLink ever suffering a loss on any of these deals: on the contrary, the evidence adduced showed that its indebtedness (which it had no substantial resources to enable it to meet) was always resolved by some form of transaction with Autonomy ^{^[266]} , and ultimately by Autonomy acquiring MicroLink itself and its debts being ‘forgiven ’ (with the result that MicroLink ’s principal relevance is in the context of the Claimants ’ claims in respect of the alleged ‘ reciprocal ’ VAR transactions)^{^[267]} .

Mr David Truitt did not give evidence before me. I have only the transcripts of his evidence in the US in this regard. However, Mr David Truitt ’s depiction of MicroLink being substantially at risk in respect of its legal indebtedness, and of taking control and active steps in respect of the licensed software it had acquired, does not fit with the evidence before me. He spoke of MicroLink ’s risk and managerial control: but there is no evidence that he thought of this as defining the parties ’ understanding of their true commercial relationship.

Furthermore, Mr Matt Stephan testified at the US criminal trial that, following Autonomy ’s acquisition of MicroLink, he was asked to look at MicroLink ’s financial records and found that the VAR transactions between Autonomy and MicroLink were not recorded in MicroLink ’s own books: they were “ just not in the books at all ”:

“Q. When you looked at the financial records at MicroLink, was there anything that caused you concern?

A. Yeah. They did not have on their books an equal and opposite amount of accounts payable due to Autonomy that Autonomy had in their books as owed to them, if that makes sense.

Q. All right. Can you describe that for us? I mean, what were you seeing and what were your concerns?

A. I think —— I can ’t recall the figures, but Autonomy was owed over $10 million by MicroLink at that point and MicroLink had a much, much smaller balance on their books as owing to Autonomy.

Q. Was that because they were listing smaller figures for the deals?

A. No. The deals —— the deals were just not on the books.

Q. So debts ——

A. A number of the deals — sorry. A number of the big deals that we ’d signed in recent quarters were just not in the books at all.

Q. So debts to Autonomy by MicroLink were not reflected in MicroLink ’s books?

A. That ’ s correct.

Q. So why did this cause you concern?

A. It indicated to me that despite signing audit confirmation letters and different things about not having any side agreements in place, it suggested that from MicroLink ’s perspective, they were never going to have to pay that money if they didn ’t get paid or if they didn ’t close the deal with the end-user.”

Overview of Mr Baiocco ’s evidence as to Capax Discovery ’s role in impugned VAR deals

Mr Egan ’s opposite number in the 10 impugned VAR transactions between Autonomy and Capax Discovery (see paragraph 1963(2) above) was Mr Baiocco. Mr Baiocco was managing partner of both Capax Global and Capax Discovery and the central figure in the impugned VAR transactions in which the latter was a party.

I have made an assessment above (see paragraphs 426(7) and footnote 55 above) of Mr Baiocco ’s reliability. He was another practised witness, having met with HP ’s lawyers three or four times in 2013 and with US prosecutors six or so times before the Grand Jury proceedings prior to the US criminal trial. He also gave evidence to the Grand Jury and in the US criminal trial. He has had throughout the benefit of his own lawyer, paid for by HP. Parts of his evidence relating to the impugned VAR transactions struck me as the product of careful forensic lawyering: when cross-examined, some parts of his witness statement appeared unfamiliar to him; and his voice in cross-examination was a contrast with the words on the pages of his witness statement.

Like Mr Egan ’s evidence, the format of Mr Baiocco ’s witness statement was to set out matters of general application and then deal more particularly with the various transactions individually. I address general matters at this stage.

The Claimants portrayed Mr Baiocco ’s evidence as being to similar effect as, and thus complementary to, Mr Egan ’s, but looked at from the other side of the bargain between Autonomy and the VAR.

Again, like Mr Egan, Mr Baiocco paid lip service to the legal content of the VAR agreements made with Autonomy on behalf of Capax Discovery, and accepted that the VAR agreements were legally enforceable. But he sought to portray the risk of actual enforcement as attenuated, if not extinguished, by assurances he stated he had received from Mr Egan, and the ‘ pattern ’ which Mr Baiocco expected or came to expect the transactions to follow. Mr Baiocco sought to capture this outlook in stating that the real risk, as he thought of it, was that “Autonomy might not follow through on my handshake agreement with Mr Egan.”

There was a marked contrast between the honed fluency and consistency of Mr Baiocco ’s witness statement and the contradictory and sometimes evasive evidence he gave when cross-examined.

The gist and (to my mind) manufactured message and contrived tone of Mr Baiocco ’s witness statement is apparent from the following passages in it, which the Claimants, somewhat unrealistically, submitted were unchallenged:

“29. In around May or June 2009, Mr Egan asked me whether Capax would become a reseller for Autonomy. A value-added reseller, as that term is usually used, is a company that purchases a product from a manufacturer or supplier to which it adds features or services and then resells the package (usually to an end-user)as an integrated or completed solution. However, that was not the nature of the relationship we had with Autonomy. Instead, Mr Egan told me that Autonomy often faced the situation where it was very close to completing a sale to an end-user, which it was not able to conclude by the end of the quarter. Rather than Autonomy lowering the price to get the end-user to sign a contract for quarter end, Autonomy wished instead (a) to enter into an agreement with us at quarter end supposedly for on sale by us of the software in question to the end-user, and then (b) to continue to negotiate with the end-user and to close the deal with the end-user in the following quarter. Capax Discovery would receive a 10% fee.

30. There were a number of significant features to these transactions:

(a) First, I was usually approached by Autonomy (almost always Mr Egan) to enter into these VAR transactions right at the end of a quarter. As I understood it, Autonomy ’s primary objective was to close its deal with the end-user before the end of the quarter, thereby avoiding the 10% profit commission that it would pay us if we were involved. Thus, it appears that our participation was only requested where Autonomy failed to conclude a deal with the end-user. As a result, we were often approached by Autonomy at the very end of the quarter and the paperwork had to be rushed through before quarter end. All of the transactions that I describe in this statement, without exception, were entered into on the very last day of a quarter. I would only agree to enter into the transaction if I was assured that Autonomy was very close to concluding the deal with the end-user, as reflected in my email dated June 24, 2009 to Mr Sass, which made it plain that we would only consider taking transactions which Autonomy was “fairly confident” would close.

(b) Second, although on paper Capax Discovery licensed the software from Autonomy, we did not pay for it upfront. The agreement with Mr Egan - again by way of a “ handshake”- was that Capax Discovery would not be required to pay Autonomy until Autonomy closed the deal with the end-user and the end-user (i) paid Capax Discovery or (ii) Autonomy (in which case our debt under the VAR agreement would be forgiven), or (iii) Autonomy had otherwise put Capax Discovery in funds to make the payment. In actuality, in most cases, Autonomy either caused the end-user to pay Capax Discovery and we in turn then paid Autonomy, or if Autonomy failed to conclude a deal with the end-user or the end-user made its payment to Autonomy directly, Autonomy forgave Capax Discovery ’ s payment obligations. As with the EDD arrangement, there was always a risk that Autonomy might not follow through on my handshake agreement with Mr Egan. Again, our use of Capax Discovery as the vehicle for these VAR transactions was designed to protect the rest of the Capax group from this risk.

(c) Third, there was never any price negotiation between Capax Discovery and Autonomy: instead, in each case, Autonomy prepared a purchase order for submission by Capax Discovery in a specified amount, for specific software that was, on paper, to be resold by Capax Discovery to an identified end-user. We then executed the purchase order as requested and returned it to Autonomy.

(d) Fourth, while Autonomy provided us with a key which could be used to access the licensed software from an Autonomy website, as far as I am aware we never actually retrieved the software because we had no need for it. We never were asked to deliver the software to the end-user.

(e) Fifth, the end-users were presented to us by Autonomy and we played no part in selecting or identifying them. The negotiation of the sale to the end-user (before we submitted our purchase order and after we submitted our purchase order) was handled exclusively by Autonomy, without any assistance or participation by us. In fact, Autonomy did not allow Capax Discovery to get involved in the discussions with the end-user. When an agreement was ultimately reached between Autonomy and the end-user, Capax Discovery played no part in bringing it about, or in negotiating or approving its terms. In most (but not all) cases, payment was made directly by the end-user to Autonomy rather than to Capax Discovery. Capax Discovery did not actually “ resell” the product to the end-user.”

Like Mr Egan’s, his witness statement presented an overall picture of the VAR agreements to which Capax Discovery was a party as being a paper record to show Autonomy’s auditors “which did not reflect our actual relationship with Autonomy”, which was in reality defined by a “handshake” and characterised as a matter of substance by the “significant features” identified.

It was unclear from Mr Baiocco’s witness statement how many of these “significant features” it was being suggested Mr Baiocco and Mr Egan had agreed by such a “handshake” in advance. The impression sought to be created, I think, was that the features identified did reflect what had been agreed between them; or at least, that the substance of what the parties had in mind was reflected in the pattern of the transactions concerned. However, the carefully assembled “significant features” seemed to me to be far too precisely engineered and detailed to be comprehended in an informal exchange; and I am in little doubt that they reflected lawyers’ analysis, honed over many years, after the event and after many interviews and two substantial trials, and not Mr Baiocco’s recollection of anything agreed at the time. Indeed, in cross-examination Mr Baiocco accepted that the scope of any handshake agreement or assurances was far more limited, as I elaborate later.

Another carefully constructed passage in Mr Baiocco ’s witness statement addressed the fact that Mr Baiocco signed a series of audit confirmation letters which (he confirmed) he knew were sent to Deloitte. Both in his witness statement and under cross-examination, Mr Baiocco sought to portray these, at least in his understanding, as simply confirming “Capax Discovery ’s existing payment obligations to Autonomy.” He said that he had been told by Mr Egan that Capax Discovery had to sign them to “let the auditors know that we owed them, that the bill - the receivables on there were proper” (as he elaborated when cross-examined). This was an obvious effort to confine the effect of the signed confirmation letters to an acknowledgement of legal indebtedness under the VAR agreements, which Mr Baiocco always accepted. However, the audit confirmation letters were not so confined. In addition, as previously noted, each such letter formally represented that there were “no side letters or other agreements in respect of the subject matter of this request” and later versions included a further representation that Autonomy “retains no continuing managerial involvement in the delivery of this product or service, other than stipulated in the licence agreement.” As I also elaborate later, even Mr Baiocco, who was not a man to appear abashed, looked uncomfortable in resorting to saying that he had not spotted or understood this at the time.

Mr Baiocco did not mention in his witness statement the “entire agreement” provisions which were contained in every version of the Capax Discovery VAR agreements. When cross examined, he initially accepted both that such a provision was included in the VAR agreements and, more generally, that such a provision was standard in software agreements. Indeed, he was taken to Capax Discovery ’s own standard terms in its own contracts which contained a similar provision, which did not surprise him. However, when later in his cross-examination, Mr Baiocco was asked whether he accepted that the reason for such terms was that there was “always a risk that a salesman might try to say something to get a deal”, he said that he could not really answer that because he did not “understand this language enough to speak anything to it”.

His evidence under cross-examination when shown the numerous audit confirmation letters he had signed and pressed on their apparent inconsistency with the side agreements he asserted is exemplified by the following passage:

“Q. What you were doing was confirming that these amounts were due and owing, that there were no side agreements and that Autonomy retains no continued managerial involvement in the delivery of the product?

A. What I was confirming when I signed those was that I owed the money, because they told me that these were about owing the money, so I signed that knowing I owed the money. I never looked or saw the side letter agreement and as far as the continuing managerial involvement, that was not on the first couple of them, I believe that was something that was added down the road. So I did not knowingly sign those knowing that there was a –

Q. How do you know those words were added down the road?

A. Pardon?

Q. How do you know those words were added down the road?

A. Nine years of lawyers.

Q. Right, sitting with too many lawyers?

A. Yes.

Q. So they pointed it out to you.

You were signing this, it ’s on the very page that you were signing and it said that there were no side letters or other agreements in respect of the subject matter, and that was true as you understood it, wasn ’ t it?

A. It was true to me that we owed the money. I did not recognise that sentence when I signed these.

Q. Right. Did you think there were side letters or other agreements?

A. I mean, it depends how you define “side agreement”, but –

Q. All right, how do you - I ’m just asking you your understanding –

A. I personally do not

Q. You do not and you did not at the time?

A. I did not at the time but it ’s irrelevant because I did not see that at the time.

Q. Well, that ’s your evidence. It ’s on the same page, I suggest you did see it, but that ’s what you say.

A. You don ’t know that.

Q. But I ’m asking you now about what you believed at the time. You didn ’t think there were any side agreements or other agreements that affected the debt, did you?

A. If you consider they were going to swap out the deals, a side agreement, then there was a side agreement, I didn ’t sign that with that in mind because I didn ’t see it.

Q. When you say, if you consider they were going to swap out the deals, that ’s going back to what you said before, that Mr Egan said to you, “ We ’ll do what we can to try and get you another deal” ?

A. Correct.

Q. Nothing more than that? That ’s all he said?

A. If it fell through, we ’ll get another deal, yes, we ’ll swap out the deal, correct?

Q. But you never thought that what he was saying was legally binding, did you? You never thought that what he was saying about swapping out the deal was legally binding?

A. Correct.

I mean, I guess unless oral promises are legally binding. I ’m not a lawyer.

Q. Well, let me ask you about that. Did you or did you not think that whatever he said about swapping out the deal was legally binding?

A. I can ’t really answer that with knowledge. I don ’t know if –

Q, What did you believe?

A. I believed probably not.”

The contrast between this evidence under cross-examination and the bald assertion in his witness statement, crafted after “nine years of lawyers” , that the “confirmations were false” is stark. Mr Rabinowitz interceded in the course of cross-examination to suggest that the assertion as to the confirmations ’ falsity related only to the part relating to management involvement in the delivery of the goods; but had quickly to withdraw the suggestion and accept that the reference was to both aspects of the confirmations provided.

As to Capax Discovery ’s obligation to pay, Mr Baiocco ’s evidence as I understood him was that:

(1) The deals were usually made right at the end of a quarter; this was not unusual in the sector but amongst its consequences was that there was never any price negotiation between Capax Discovery and Autonomy: the price was that which Autonomy had established for the sale to its proposed end-user;

(2) Capax Discovery was, in terms of the law, ‘on the hook ’ to make payment in full; but neither party ever intended or expected payment by Capax Discovery unless and until the end-user deal was closed, whereupon the amounts paid would be applied in discharge of the VAR.

(3) If no end-user deal eventuated Autonomy would devise some way of extracting Capax Discovery from the ‘ hook ’: at one and the same time as emphasising Capax Discovery ’s legal obligation, Mr Egan always assured Mr Baiocco that Autonomy would do everything it could to ensure that Capax Discovery was not left exposed.

(4) A particular way of negating this exposure was specifically discussed: this was that Autonomy would do its best to “swap out” any deal which failed to result in an end-user transaction which would provide the funds to satisfy the obligation. But Mr Baiocco ’s evidence left me in no doubt that the assurance was general and the intention clear.

As to the other legal and economic effects of the VAR transactions as between Autonomy and Capax Discovery, and in particular, as to how it was intended the parties to the VAR sales should conduct themselves after the date of sale, Mr Baiocco ’s evidence was:

(1) He believed that delivery of the software licensed was always made electronically on execution of the VAR sale: this was effective delivery;

(2) However, he was never involved in that process, and (so far as he was aware) no-one at Capax Discovery ever actually retrieved the software, “because we had no need for it. We never were asked to deliver the software to the end-user.”

(3) Although the contract did not reserve to Autonomy any contractual right to do so, in practice the parties intended that Autonomy should be in actual and exclusive control of the onward sales process. The following passage from Mr Baiocco ’s cross-examination demonstrates this:

“Q. Then the next point is about the identification of the end-users. Now, you say that Autonomy didn ’t allow Capax to get involved in discussions with the end-user. Who do you say said that? Mr Egan?

A. Yes.

Q. Now, can we just consider that for a minute. He didn ’t actually say you couldn ’t, did he? He wasn ’t telling you, you couldn ’ t?

A. I think on occasions I asked -- because we may have had some people we knew at the end-user clients, maybe not in this specific area, and they were like, basically, “No, stay out of it”.

Q. Okay, so that ’s Mr Egan saying that?

A. Yes.”

(4) If the end-user sale proceeded, Autonomy would pay Capax Discovery a 10% fee (a MAF). Mr Egan regarded this as a fee for taking on risk, analogous to a guarantee fee. The letters relating to the MAFs stated that the products would be sold to Capax Discovery at a 10% discount and that Capax Discovery would get its ‘fee ’ from the full price on onward sale. But in fact Capax Discovery never did actually “ resell” to the end-user. Autonomy simply paid Capax Discovery the 10%. Mr Baiocco was not interested in and never read the letters relating to the MAFs: he was only interested in getting the fee.

(5) Capax Discovery never had dealings with the end-user prior to the sale. Part of the commercial rationale from its point of view was that it would have the opportunity to offer the end-user professional services thereafter: but this happened in only two sales (the TXU and FSA deals).

The Goldberg Segalla letter

In addressing Mr Baiocco ’s evidence, and in seeking to contradict the Claimants ’ contention that it demonstrated that in economic substance, Capax Discovery took on neither any real risk nor any managerial responsibility as regards the software it had contractually acquired, Mr Miles drew repeatedly from a letter dated 1 ^st November 2013 ( “the Goldberg Segalla letter”). This letter was written on behalf of Capax Discovery ’s holding company, Capax Global LLC by its lawyers, Goldberg Segalla LLP to the US Department of the Air Force ( “USDAF”).

The Goldberg Segalla letter was signed by Mr Baiocco ’s present lawyer, Mr Christopher Belter ( “ Mr Belter”), who was paid by HP and accompanied him at the hearing. The letter was stated to be “in response to the show cause letter dated September 6, 2013 inviting Capax Global LLC ( “ Capax”) to respond to allegations concerning accounting improprieties involving Autonomy” in respect (in particular) of a “government end-user business opportunity…brought to Capax in late December 2010”.

The avowed purpose of the Goldberg Segalla letter was stated to be “to show that at all relevant times Capax and its subsidiaries have been responsible contractors and should not be proposed for debarment from Government contracting.” To that end, the letter described first “the nature of Capax ’s business and the relationship between it and Autonomy, including the current relationship with [HP]” and then set out reasons why Capax Discovery should be regarded as a “responsible contractor with particular emphasis on its ethics, integrity, compliance with laws, and ability to perform adequately.”

Mr Baiocco told me that he had read and approved the Goldberg Segalla letter at the time, and had also had an opportunity to look at the letter recently. He confirmed his belief that the statements made in the letter were true: he told me he believed it to be “ an honest letter” , and he acknowledged that he was well aware at the time that there were serious potential consequences (including criminal consequences) of not giving an accurate and comprehensive account in reply to the USDAF. Of course, it was not realistic to expect him to say or suggest otherwise.

Dr Lynch attached very considerable importance to the Goldberg Segalla letter. Key points in it relied on and emphasised by the Defendants as contemporaneous support for its description of the nature and effect of the VAR sales were as follows:

(1) Capax Global and Capax Discovery were treated as one company ( “ Capax”), and Capax was presented as a sizeable entity which had been in operation since 1999 and as an

“information technology company that provided technology services, and solutions concerning cloud hosting, software application and development, systems integration and technical support.”

(2) The letter explained Capax Discovery ’s position in the market, the breadth and depth of its customer base (including 250 Fortune 500 companies) and its role as a professional services provider and reseller for Autonomy ’s products and especially its core software product, IDOL, from May 2009. That presentation (which reflected the position of Capax Global but not its subsidiary Capax Discovery, which was not established until early 2009) went on to state (again in a passage which did not apply to Capax Discovery) that:

“A mainstay of Capax ’s solution offering is the deployment and configuration of Enterprise Search and the incorporation of Enterprise Search into a customer ’s applications to achieve business value.”

(3) The letter addressed two impugned VAR sales, one (referred to in the letter as “The Government End-User Transaction” ) in Q4 2010 (VT20) for which the VAR was Capax Discovery and the end-user was Defence Knowledge Online ( “ DKO”, part of the US military) and the other (referred to in the letter as “The Entity B Transaction” ) in Q1 and Q2 2011 was a two-stage impugned VAR sale (VT28 and VT34) for which the VAR was Capax Discovery and the end-user was UBS.

(4) The letter gave an explanation of the way the reseller and supplier relationship worked in the market, explaining inter alia how reseller risk included revenue loss caused by the inability to collect payables.

(5) The letter also explained the role of a VAR in general and Capax Discovery ’s specific role as a VAR for Autonomy:

“In the computer industry, a VAR may add value to a product by customizing or implementing software and/or being a guarantor. As a VAR for Autonomy, Capax was a guarantor as well as a provider of services and support that allowed Autonomy ’s software to function effectively in its clients ' environments.” [Emphasis supplied]

(6) It stated that Capax Discovery viewed being a reseller for Autonomy as a way to “potentially create direct and valuable relationships with new clients in order to sell them additional products and services” .

(7) It emphasised that Capax Discovery took substantial risk in every reseller transaction with Autonomy. It stated unequivocally that:

“At no time did Capax ever think that it was not at substantial risk. Indeed, in all VAR transactions with Autonomy, if the deal with the end-user did not materialise, Capax would ultimately still be responsible to Autonomy for the payment of the software licences.

Although Autonomy would possibly assist Capax in finding a new client in a joint selling arrangement or Capax would itself be able to sell the licences to another client, there was no guarantee that Capax would ultimately find a different end-user. Ultimately, weighing the risks and reward potential, Capax decided to proceed with these VAR deals. Capax had additional confidence with respect to the risks involved in these deals because Capax was, and remains, a major service and support provider to Autonomy ’s clients. Over the years, Capax has provided, and continues to supply, substantial and effective services and/or support for over 1,000 Autonomy clients.”

(8) It stated that Mr Baiocco and Capax Discovery considered that they had no reason to believe that any Autonomy revenue was being improperly recognised.

(9) It confirmed that Capax Discovery had reviewed the audit confirmation letters which confirmed that Capax Discovery was on risk and that there were no side-agreements and believed them to be accurate; and in cross-examination, Mr Baiocco told me that when he signed the audit confirmation letters he “ absolutely believed what I was saying was true ”.

(10) Capax Discovery also confirmed that it was on risk in relation to the specific reseller deals addressed in the letter.

The Defendants placed considerable reliance on the Goldberg Segalla letter: for Dr Lynch it was submitted that:

“the Claimants ’ case on the Capax transactions is unsustainable in the light of this letter and Mr Baiocco ’s confirmation of its accuracy”.

In essence, the Defendants ’ case was that the Goldberg Segalla letter represented a reliable and more contemporaneous record of Capax Discovery ’s role and acceptance of risk in the impugned VAR transactions which was consistent with Mr Baiocco ’s statement that he never believed at the time that there was any side-agreement; and that the description given in the letter of the VAR arrangements should be preferred to the evidence in his witness statement in these proceedings. The letter was presented as reflecting the true relationship between the parties; his witness statement, and in particular his evidence about “side agreements”, as reflecting Mr Baiocco ’s business imperative to remain on good terms with HP ^{^[268]}, his many meetings with HP ’s lawyers and prosecution lawyers in the US, and lawyers ’ input: it was the result of “what had been suggested to him repeatedly over the years.”

However, the Goldberg Segalla letter must be set in context, and read with an appreciation of its objectives. It is true that the letter made no mention of any side-agreement, and focused only on legal commitments and their presented rationale. But the Goldberg Segalla letter was quite obviously heavily lawyered and carefully crafted to seek to maintain Capax Discovery ’s valuable status as a Government Reseller, and I do not think it realistic to treat the letter as an authentic and complete depiction of the relationship between Capax Discovery and Autonomy. Indeed, the Goldberg Segalla letter has unsettling features, and in particular:

(1) The misleading elision of the two Capax Discovery companies to give the impression of far greater financial solidity and a much wider customer base than could be said of the VAR in each case, which was Capax Discovery, not Capax Global;

(2) The rolled-up presentation of the usual activities or contribution of a VAR in general, and “Capax Discovery ’s” contribution in particular, which gave an impression that “Capax Discovery” did far more to “add value” than was the truth;

(3) The exaggeration of the risk borne by Capax Discovery, given that there was no real doubt that Mr Egan had given assurances of assistance with a view to doing all that was possible to ensure that Capax Discovery would not be left exposed, the real issue and dispute being not whether any assurances had been given, but whether they amounted to a proscribed “side-agreement”;

(4) Clear examples of half-truths in the depiction of the nature of the risk and Capax Discovery ’s role in its VAR deals. The risk depicted was the legal risk: but not the reality; and as to Capax Discovery ’s role, the suggestion, for example, that Capax Discovery “viewed being a VAR as a way to potentially create direct and valuable relationships with new clients in order to sell them additional products and services” may not have been untrue at one level: but in the context of Capax Discovery ’s impugned VAR transactions with Autonomy it is, at another level, either inapposite or misleading, given that in none did Capax Discovery deal directly with the end-user (Autonomy having itself (alone) negotiated directly).

In short, whilst I would accept that Mr Baiocco ’s evidence was influenced by the input and (I suspect) insistence of HP ’s lawyers and the prosecuting authorities that he should (as it were) label the assurances he was given as, and confess to, a ‘side-agreement ’, that was really a matter of legal labelling: the Goldberg Segalla letter does not disprove or even tell against Mr Baiocco ’s consistent and, to my mind, entirely credible position that Mr Egan gave the assurances that both he and Mr Egan described. The truth is that in both the letter and Mr Baiocco ’s witness statements, his own voice is for the most part barely audible: his lawyers ’ voice is what prevailed in both contexts and too often it followed what they perceived to be, or was made clear to them to be, the required legal line.

In summary, I agree with the Claimants that the picture revealed by Mr Baiocco on the other side of the impugned VAR sales is in substance the same as that painted by Mr Egan; and (in my own words) that:

(1) Mr Egan emphasised to him, and Mr Baiocco understood, that Capax Discovery was legally obliged to pay Autonomy in accordance with the terms of the VAR sales contract. There was no doubt as to the VAR ’s legal indebtedness as a matter of law.

(2) Mr Baiocco understood the letters of confirmation simply to confirm that legal indebtedness, principally for the benefit of Autonomy ’s auditors. He did not understand those letters to have anything to do with his handshake deals with Mr Egan, which he did not understand to be legally enforceable, but which he did consider to be reliable as a matter of reality, both parties ’ interests being aligned.

(3) The residual risk was that contrary to the intentions and expectations of both parties, and notwithstanding the best efforts of Autonomy to hold Capax Discovery harmless or release it from liability, something would happen which caused Autonomy to resort to its legal right to enforce payment: absent some change or control in Autonomy, or some cataclysmic collapse in its fortunes of the marketability of IDOL, this was most unlikely; and Mr Baiocco considered that this risk (which in point of fact never eventuated) was covered by the upside to Capax Discovery of enabling it to sell services to the end-user (as Mr Baiocco told me in cross-examination did happen) and in payment to it of a MAF, which Mr Baiocco accepted was payment for the nominal or residual risk assumed (which Mr Baiocco described as acting as a ‘ guarantor ’).

(4) Neither party to the VAR transactions between Autonomy and Capax Discovery ever expected Capax Discovery to play any material role in negotiating the end-user sale of which the VAR sale was in contemplation: Autonomy carried on such negotiations after the ‘sale ’ in every case as if nothing had happened: even offers of assistance volunteered on the part of Capax Discovery were roundly rebuffed by Autonomy.

(5) The much-emphasised benefit to Capax Discovery, used to justify the nominal risk, of an introduction to end-users was greatly exaggerated. Any contact with end-users was after the conclusion of the end-user contract, and was not in consequence of any involvement in or risk taken as a VAR, but in consequence of an end-user’s need for service support which Autonomy usually outsourced anyway and for which Capax (usually Capax Global) was a preferred provider.

Evidence in relation to FileTek ’s single impugned VAR transaction

FileTek’s single VAR transaction with Autonomy in September 2010 (VT18, for end-user USDVA) was in a sense an outlier. FileTek was not in business as a VAR and had no real wish to be so. Its President and Chief Marketing Officer from 2009 to 2013, Mr Szukalski, did not wish to do the deal and was not involved in any direct negotiations with Mr Egan.

However, he reported to Mr Loomis, FileTek’s CEO and CFO during the Relevant Period. Mr Loomis confirmed that he had never done a VAR deal with Autonomy before: it was “something new” for FileTek; it was Mr Loomis who directly dealt on behalf of FileTek with Mr Egan in respect of VT18, and who also reported directly on the deal to FileTek’s ultimate owner (Mr Bill Thompson).

The “direct” evidence relating to FileTek ’s perception of its role comprised the oral testimony of Mr Szukalski (in person) and of Mr Loomis (given in the US criminal trial and admitted under a hearsay notice into these proceedings). Mr Szukalski previously worked for Verity Inc ( “Verity”), another software company which was acquired by Autonomy in 2005.^{^[269]}

Their evidence was also directed to two transactions with FileTek which the Claimants alleged were, and sought to impugn as, “ reciprocal” transactions in Q4 2009 and again in Q1 2010 (under which Autonomy purchased StorHouse software from FileTek), and to a further succession of three purchases of StorHouse licences by Autonomy from FileTek in March, June and August 2011. I deal with the two “ reciprocals” in detail in the section of this judgment in which I address the Claimants ’ claims relating to “ Reciprocals” and other purchases by Autonomy from its own customers. As the 2011 transactions were alleged to have been contrived to put FileTek in funds to pay down its debt under FileTek ’s single VAR deal with Autonomy, I address them when dealing with this VAR transaction (VT18) (end-user USDVA).

There was no mention in the evidence of Mr Loomis in the US criminal proceedings that there was any side agreement or understanding qualifying the terms of the VAR transaction. Indeed, as was pointed out in Dr Lynch ’s written closing submissions, the prosecution did not suggest there was any prior agreement, but instead focused on an allegation that, while FileTek ’s debt was real and substantive, the later StorHouse purchases in 2011 were made primarily to assist FileTek meet its obligations, rather than because Autonomy had any need for further StorHouse licences.

I address the effect of the evidence given by Mr Szukalski and Mr Loomis in respect of VT18 at greater length when addressing that transaction later in this judgment. For present purposes a broad summary should suffice. In my view, the evidence of Mr Loomis and Mr Szukalski is again largely consistent with Mr Egan ’s account to the effect (in brief summary) that the legal risk was clear, but FileTek, like Capax Discovery, well knew that there was no real economic risk of the legal obligation being enforced, since some means would be found by Autonomy to avoid that. However, neither Mr Loomis ’ nor Mr Szukalski ’s evidence provides any support for the suggestion made by Mr Egan that pending the end-user sale the VAR might be made to pay: Mr Loomis made quite clear that this was never on the cards, and in the event, he felt no difficulty in withholding any payment accordingly.

Overview of Mr Steve Truitt ’s evidence on the impugned MicroTech VAR transactions

MicroTech was a fast-growing, and by 2011, already substantial reseller. By then, it had some 350 employees, having had no more than 12 or 13 in 2006. It acted as a VAR for a number of large US companies. As well as being a reseller for Autonomy, MicroTech was a Cisco Gold Partner, an HP Elite Partner, a Symantec Managed Services Partner and an Oracle Silver Partner, all corporations of very substantial size. Its COO was Mr Steve Truitt. Its CEO was a Mr Jimenez, who had decision-making power, and whose approval Mr Steve Truitt needed for each deal.

The Claimants sought to impugn nine VAR transactions between Autonomy and MicroTech on the basis that in consequence of side agreements or understandings risk did not pass to the VAR, and nor did managerial control of the goods (save for VT5, which turns on its own peculiar facts). The Claimants submitted that the transactions were designed simply to enable Autonomy to recognise revenue at the point of the VAR agreement, but in substance none constituted any real economic sale.

The relevant impugned VAR transactions with MicroTech, which took place between Q4 2009 and Q2 2011, are VT5 (end-user DiscoverTech), VT6 (end-user Honeywell), VT7 (end-user ManuLife), VT8 (end-user Morgan Stanley), VT13 (end-user Vatican Library), VT25 (end-user US Department of the Interior) ( “DoI”), VT32 (end-user Bank of Montreal), VT33 (end-user Xerox), and VT37 (end-user HP).

Although Mr Steve Truitt ’s evidence was that the deals with Autonomy were originally introduced by his brother, Mr David Truitt, and a Mr John Cronin (a Consultant to MicroTech, who appears to have been paid a commission by MicroTech for each such deal) the VAR deals with Autonomy were operationally controlled by Mr Steve Truitt. On the Claimants ’ case, therefore, any discussion giving rise to a side agreement or understanding would have had to be between Mr Steve Truitt and Mr Egan.

In respect of their claims that the nine impugned VAR transactions between Autonomy and MicroTech had involved no real economic sale the Claimants relied principally on Mr Steve Truitt ’s evidence given in a deposition dated 7 September 2018, which was made in the course of proceedings in the US between MicroTech and Autonomy and admitted into these proceedings under a hearsay notice.

The Claimants placed particular reliance on the following passage:

“Q. Did you - was it your understanding that on these deals that Autonomy had the relationship with the customer and Autonomy could continue the sales effort with respect to the ultimate - the true customer?

A. Yes.

Q. Okay.

A. As it says here, these - these are deals that were ostensibly very close to closing. It certainly wouldn ’t make sense to - to hand over, you know, to bring in the second team offense when you ’re on the 5- yard line.

Q. Okay.

A. They ’re supposed to score the touchdown and we get the extra point.

Q. Okay. So - so the - and was there - withdrawn. Who - who controlled the negotiation of the price with the end customer?

A. I don ’t know.

Q. Not MicroTech, though?

A. Not MicroTech.

Q. Okay. Somebody at Autonomy?

A. Somebody at Autonomy.

Q. Okay. And so was the concept at the beginning that you would take the deal, Autonomy would continue to sell to the end-user and - and, hopefully, there would be a sale to the end-user?

A. Well, I would - I would say that ’s correct, except that at this point in time, it was more than hopefully. I was - what I was looking at was the - was an attempt to duplicate the experience that my brother [David Truitt], who ’s my brother, said that he had had.

Q. Okay.

A. So it was more than hopefully.

Q. Okay. And - and it ’s correct, is it not, that you understood that Autonomy would be not only attempting to sell to the customer, but negotiating the price with the customer?

A. Yes.”

Mr Steve Truitt gave similar evidence when asked about specific purchase orders. For example, as regards the purchase order that MicroTech submitted in Q4 2009 for end-user Morgan Stanley (VT8), Mr Steve Truitt testified in the same deposition as follows:

“Q. Who did you understand would be dealing with Morgan Stanley about the purchase of a license to use the software identified on page 3 of Exhibit 37?

A. A salesperson who worked for Autonomy.”

Another example relied on by the Claimants was the Vatican Library purchase order (VT13). When asked about this, Mr Steve Truitt testified in his oral evidence in chief in the US criminal trial (also admitted into these proceedings under a hearsay notice):

“Q. In terms of the Vatican, was MicroTech making any effort to sell to the Vatican?

A. No.

Q. Who was going to sell the software to the Vatican?

A. Autonomy.”

However, in a passage in his cross-examination in the US criminal proceedings marked by the Claimants as evidence on which they intended to rely, Mr Steve Truitt explained the rationale, from MicroTech ’s perspective, for entering into the reseller deals, and also confirmed that he understood that MicroTech were fully on risk:

“Q. And these -- these deals you saw as a way to get the company to grow; right?

A. Yes.

Q. Meet new customers, you've talked about?

A. Right.

Q. And you knew that your brother had been very successful in doing similar kind of work when he'd been at MicroLink?

A. That's right.

Q. And you wanted to try to repeat that for MicroTech?

A. Yes.

Q. You had, as I understand it, three objectives for MicroTech being a reseller. The first one was -- and the primary reason was to get service business from the people that would buy the Autonomy software?

A. Yes. That's correct.

Q. And then another reason was to meet people -- meet customers and grow your business, hire more technical people?

A. Yes. Particularly commercial customers.

Q. And then a third reason was to get this -- you buy the software at a discount so you would get a 10 percent premium; right?

A. Right.

Q. But you wouldn't have done these deals just for the 10 percent premium. It was these other reasons that predominated, didn't it?

A. Yes. Those were the reasons that mattered to me.

Q. And the deals that you did, you knew that MicroTech was at risk when it agreed to buy Autonomy software, didn't you?

A. The first set of deals in particular as I testified earlier, yes, I considered us to be at risk.

Q. You considered -- and I think you used the words that MicroTech was on the hook for the debt?

A. Yes.”

In his deposition evidence in the MicroTech US litigation, Mr Steve Truitt was at pains to emphasise that, contrary to what was suggested to him, he regarded the indebtedness to Autonomy as “100 percent real”: “we owed it not just on paper. We owed it for real” ; and that when the end-user deal did not close in a “timely fashion” he was “a hundred percent freaked out”. He added, however, that there came a time (which he could not specify) when he felt less concerned

“because Autonomy was going to work with us to make arrangements to pay these debts with money that we didn ’t have to go make elsewhere.”

His evidence suggested that this was in part something which he came to appreciate from experience: but he also stated (in his deposition evidence in the MicroTech litigation) that his approach was informed by his brother Mr David Truitt’s earlier experience at MicroLink in 2009 (see paragraph 2041 above); and Mr David Truitt was clear in his own deposition evidence in those proceedings that it was understood that MicroTech would not pay Autonomy for the VAR deals unless it had first received the money from someone with which to make such a payment.

Pulling these strands together, it seems to me that, though with perhaps a little more equivocation, Mr Steve Truitt ’s evidence also substantially endorsed Mr Egan ’s presentation and that in summary:

(1) Even if he was more concerned initially, it was not long before he well understood that the legal risk would be most unlikely to eventuate, that Autonomy had no intention of relying on its legal rights and on the contrary, that it fully intended to ensure that MicroTech was not left “on the hook”.

(2) He did not regard this as a side-letter or side deal at the time because, as he put it, there was “never a literal side letter”. But he was clear that, whether or not labelled a side deal (as he said in his deposition evidence he was only later persuaded by HP ’s lawyers and the US prosecution counsel to call it), it was made plain to him by Autonomy that they would see to it that they would, if necessary, arrange “to do business… separate from those deals so that we could somehow get the money to pay those debts off”. As to the label he was later persuaded to put on it, he said (to quote again from his deposition evidence):

“…look, I didn ’t think of it this way at the time. There was never a written side deal, but if you want to call that a side deal, call it a side deal. I will acknowledge that this situation existed.”

(3) At no stage was MicroTech involved in any negotiations for or the closing of end-user deals; and it should be noted that the customer contact which Mr Steve Truitt anticipated and placed value on was not to eventuate until after the end-user sale, for the simple reason that only Autonomy would be involved in the negotiations for it; and if none eventuated, that benefit would evaporate. As he explained in his direct evidence in the US criminal proceedings:

“…the way it was going to work was that when the [end-user] deal closed, we would be introduced to the key people and the customer and the technical contacts, and we would then go actually install and configure the software and start building relationships with the customer. And as they started learning what the software could and couldn ’t do, we would - we would extend and modify and reconfigure the software, sell them additional packages to complement what they already bought. I mean, we would put a relationship in place.”

(4) He accepted that in such circumstances, Autonomy retained managerial involvement. When asked directly in the MicroTech litigation whether the additional sentence in the audit confirmation letters confirming on behalf of MicroTech that Autonomy retained “no continuing managerial involvement in the delivery of this product or service, other than stipulated in the license agreement” was true, he accepted that it was not true. He offered only that “to my discredit, I didn ’t even notice that language…”.

(5) A point of detail not mentioned above remained unexplained but nevertheless supportive of the point that MicroTech did not regard the notional debt as an economic reality: MicroTech ’s accounts did not reflect its exposure to Autonomy at anything like its face amount^{^[270]} . Thus:

i. As at 31 December 2009, MicroTech owed more than $18 million to Autonomy on VAR transactions: $10 million on the DiscoverTech purchase order (VT5), almost $1.9 million on the Honeywell purchase order (VT6), about $1.1 million on the ManuLife purchase order (VT7) and over $4.8 million on the Morgan Stanley purchase order (VT8).

ii. However, MicroTech’s own audited financial statements showed current liabilities of only $17.2 million as at 31 December 2009, of which $3.5 million related to a bank line of credit. Plainly, therefore, MicroTech’s accounts cannot have accounted for the full amount of the $18 million debt owed to Autonomy.

iii. MicroTech’s 2009 financial statements also contained no indication that Autonomy software was being held on its balance sheet as inventory, as at 31 December 2009.

iv. Steve Truitt offered no explanation of these discrepancies during his deposition:

“Q. Is it correct that you do not see either the debt owed to Autonomy or the associated asset owned by MicroTech at year end on this balance sheet?

MR RINGER: Objection.

THE WITNESS: I think that ’s a reasonable statement, looking at these numbers.

BY MR FRANK: Q. Okay.

A. But it doesn ’t mean that some - some part of it isn ’t here. I don ’t know exactly what these numbers do represent.”

(6) The position a year later was no different. MicroTech ’s audited financial statements for 2010 do not appear to show, on the balance sheet as at 31 December 2010, the debt of over $10.7 million which MicroTech owed Autonomy at that time. Again, Steve Truitt was asked about this in his deposition:

“Q. Do you see on the balance sheet anything that you believe corresponds to that more than $10 million debt?

MR RINGER: Objection.

THE WITNESS: It ’s difficult to say exactly what these liabilities are, but there aren ’t enough of them to correspond to this amount.

BY MR FRANK: Q. Okay. Do you know why that debt does not appear on MicroTech ’s audited financial statements?

MR RINGER: Objection.

THE WITNESS: No, I don ’t.”

It is not difficult to see why he did not wish to hazard an explanation; but it seems to me to be a fair inference that there was no debt recognised because in commercial terms it was perceived that it would unbalance the accounts: the debt was to be covered by the understanding with Autonomy that any debt would be funded by another (contrived) deal if necessary (see above).

Mr David Truitt ’s evidence on the impugned DiscoverTech VAR transactions

As mentioned in paragraph 1963(4) above, DiscoverTech was a newly incorporated entity in 2009. It was formed by Mr David Truitt, who was DiscoverTech ’s CEO from its inception, and in effect was MicroLink ’s successor as a favoured VAR after MicroLink was acquired by Autonomy (in circumstances and for purposes which the Claimants also seek to impugn, as I explain later).

Mr David Truitt, who continued as CEO of MicroLink after its acquisition by Autonomy, accepted in the US criminal trial that DiscoverTech ’s financial position as a start-up was not such that it could fund VAR purchases out of its own resources, at least without “recapitalising the company”. Even though theoretically it could, of course, take into account receivables from any onward sale to an end-user it is of some note, therefore, that DiscoverTech took on so many high-value VAR transactions (BofA was a “big deal” as Dr Lynch put it, even for Autonomy), especially since it cannot have been confident of receipts from an end-user sale in which it was not to be involved.

Mr David Truitt also confirmed in his evidence in the US criminal proceedings that DiscoverTech never made any effort to resell the software it had acquired from Autonomy under licence. Like Messrs Egan, Baiocco and Szukalski, and like his brother Mr Steve Truitt, he accepted that in none of the impugned VAR sales was the VAR involved in any way after the VAR transaction, and that in reality Autonomy controlled what to sell, to whom to sell and at what price.

A corollary was that in none of the impugned VAR transactions to which DiscoverTech was a party was an end-user sale by it ever effected. DiscoverTech was reliant on Autonomy either procuring payment to it out of the proceeds of a direct deal, or cancelling the initial VAR transaction (which, in the same testimony, was something that he had never had to do in any MicroLink or other VAR deals prior to the impugned VAR deals in question).

Even so, the Defendants much relied on Mr David Truitt ’s evidence (in the US criminal trial and/or the MicroTech litigation) asserting that DiscoverTech was “ at risk” and could be “stuck with it” even if the end-user deal failed to materialise and further asserting that DiscoverTech “had control of the software and [was] fully the owner” . In the latter context, they also relied on his evidence in the US criminal trial on the issue of managerial control:

“Q. And then some questions about -- you keep asking about control, who controlled what. And your answers were, "The documents we signed were the terms of the agreement. The terms are stated in the agreement." Do you remember that testimony?

A. Yes.

Q. And is what you meant is that once you made the agreement, you, be it DiscoverTech, MicroTech, whoever made the agreement, owned the software? It was your software?

A. Yes.

Q. And was -- if it didn't close, it's still your software and you owed money; right?

A. Yes.

Q. If you sold it to the end-user, the terms were already set. It was there in the agreement; right?

A. Correct.

Q. So you had control of the software, you owed the money, and you were fully the owner; right?

A. Yes.”

This evidence mirrored his evidence in respect of MicroLink VAR sales in the MicroTech litigation (see paragraph 2001 above). The Defendants made the point that such evidence had been adduced by the Claimants themselves (pursuant to hearsay notices). Once again, however, it seems to me that the only way that this evidence can be treated as consistent with his other evidence (as relied on by the Claimants) and with what in fact transpired is to treat it as reciting (somewhat ritualistically) the legal position only.

Whatever the legal position, I take it and find that Mr David Truitt accepted that in reality, and like MicroLink before it, DiscoverTech never had the means to pay Autonomy out of its own resources, never did so, and never participated in any negotiations with end-users after the VAR sale to it: and that was entirely as both he and Mr Egan expected and intended.

Dispute as to the effect of any assurances given

Both Defendants submitted that the oral testimony of Mr Egan and Mr Baiocco, and the hearsay evidence from Messrs Steve Truitt, David Truitt and Loomis in the US criminal trial and in the MicroTech litigation, suffered from its genesis in numerous meetings with the US authorities playing on the self-interest of the actors. They submitted that the evidence had “been contaminated by the forensic process” , was unreliable and inconsistent, and could not provide any secure basis for a plea of fraud. Nevertheless, they added that, read as a whole, and contrary to its purpose, such evidence served on balance to support their case, rather than the Claimants, since it demonstrated that all the witnesses well understood that nothing that Mr Egan said could or would dissolve their legal obligations.

The Defendants ’ position was thus that whatever assurances were given, none could reasonably have been or were taken by the VAR to release, qualify or modify the VAR ’s right to the software which (it was common ground) was in each case electronically transferred to the VAR when the VAR agreement completed and was then available for use, nor the VAR ’s obligation to pay within the stipulated time. The Defendants emphasised the unanimity amongst the witnesses that they were acting legitimately in VAR transactions, which were not out of the usual, that they had confirmed their belief in the accuracy of the audit confirmation letters when they signed them, and their understanding throughout that the VAR they represented remained in substance and in law fully at risk to pay Autonomy on demand.

In his oral closing submissions, Mr Miles summarised the Defendants ’ position as follows:

“ Now, the next point is to turn to the witness evidence and make some observations on that. There are several factual witnesses. There were three live witnesses, Mr Egan, Mr Baiocco and Mr Szukalski, and then the worthier hearsay statements. Now, as regards both Mr Egan and Mr Baiocco, we do suggest that you have got to be quite careful about the evidential status of certainly much of their witness statements. But even if you do… exercise caution, as we suggest you should, what we suggest the evidence as a whole shows is that, even taking the position most favourable to the Claimants there ’s nothing in the discussions that Mr Egan had with the resellers which would impact on revenue recognition.

The evidence of Mr Egan and indeed the others was that the resellers understood that they were fully on risk. When they explained in their evidence that they thought they were on risk, what they meant, we suggest, is that they would have to pay from their own resources.

If you were to think the evidence came to anything at all… It was just that Mr Egan would do what he could to find another end-user or possibly that he would try and do what he could to help. But the evidence was that the resellers knew that any assurance or words of comfort he gave them was just that, non-binding words of comfort. But more than that, even if you accept their evidence, the resellers knew that their relationship was governed by the contract, that they were bound to pay in accordance with the contract, that if things did not go right Mr Egan would try and help them out, that they also knew that was just a hope and not something that was binding on Autonomy in any way.”

This summary correctly, in my view, identified the nub of the factual issue revealed by the direct evidence: whether the assurances given were “ warm words” or “soft soap”, or whether they were intended to be and were acted on as the basis of the parties ’ relationship.

Mr Rabinowitz, rejecting the suggestion that the assurances given were simply “ warm words” or “soft soap”, summarised the Claimants ’ position in his oral reply as follows:

“…the expression… also has pejorative connotations of the kind of thing an unscrupulous salesman might say to get the deal done. But here no one is suggesting that the expressions of intent communicated by Mr Egan were anything other than his honest intention backed by the express authority of Mr Hussain and, we say, in turn backed by Dr Lynch. Mr Egan was not duping the VAR into thinking that Autonomy would avoid leaving the VAR “holding the bag”. That actually was the intention.

My lord, the best evidence of that is what actually happened. Your lordship will recall that I described the pattern where Autonomy never did require the VAR to pay in full from its own resources, leaving aside a few modest part payments. That is evidence of Autonomy ’s intention at the outset. I would note that Mr Miles did not seek to answer my patterns point.

… The proof of the seriousness with which the VARs treated these assurances, despite their non-binding character, lies in the VAR ’s willingness to assume the enormous risk that would arise if they nevertheless ended up being held to their contractual obligations. From Autonomy ’s point of view, as Mr Egan explained, any failure to honour the assurances would have made it harder to get a VAR to take on a deal in the future. For your Lordship’s note, that is what he said at Egan 1, paragraph 29, last sentence {C/5/9}. From a repeat VAR ’s point of view, the fact that the intention was always honoured would have added more weight to the assurances as time went by.”

Assessment of the effect of the direct/witness evidence

Supported by the details of the way each VAR sale was implemented which I describe in the next section, it is to my mind clear, and I find, that in every one of the impugned VAR sales except the five Collectability VARs and the two VAR transactions (VT5 and VT14) which turned on their own facts, Mr Egan both (a) emphasised, as he had been instructed by Mr Hussain he must emphasise, that legally the VAR was unconditionally obliged to pay and became upon delivery the owner of and responsible for the licensed software with a view to its resale, but at the same time also (b) assured the VAR that it would not be expected to be involved in negotiating and closing an end-user sale, and that if no end-user sale eventuated, Autonomy would not leave it “on the hook” or require payment in the meantime, and would one way or another ensure, whether by “backfilling” or some arrangement which realised funds sufficient to discharge the legal indebtedness, that it would not be left “holding the bag” .

I think it clear also, and I find, that even if Mr Egan did not make any formal representation that he was able to bind Autonomy ^{^[271]} , he made it clear, as was the fact, that he was acting in conformity with the wishes and suggestions, and probably on the actual instructions, of Mr Hussain, both in emphasising the legal position and in giving the assurances as to the real intentions of Autonomy.

Reading between the lines of their incantations (with varying degrees of insistence) about their legal exposure, the direct evidence of all the VAR witnesses confirmed, and I find, that but for these assurances, and the shared intention (on the part of Autonomy) and expectation and belief (on behalf of the VARs) that they would be fulfilled, none of the VARs could rationally have agreed to take on the role in the circumstances in which it was almost invariably obtained; these being that (a) the VAR deal would be put forward at the end of a quarter, sometimes on the very last day, (b) for completion before the end of the quarter, with (c) little if anything being known by the VAR of the state of Autonomy ’s negotiations, nor of the real prospect of a deal eventuating, with the prospective end-user; and (d) the VAR would usually not be able to cover the legal obligation out of its own resources, and in any event the risk/reward balance being plainly and obviously commercially unacceptable.

In my view, it is also clear, and I find, that in every one of these cases the VARs expected, and were content, that Autonomy would carry on negotiations with the end-user as being the best or most likely way of concluding a sale to the end-user, and that the VAR would play no part. Some might have expected an introduction to the end-user after conclusion of an end-user sale, and commercial advantage after that from it; but the most reliable source of reward would be payment to them of a MAF, of which they were assured by Mr Egan, although the VAR sales agreements made no legal provision in that regard.

The question then is as to how IFRS accounting principles are to be applied in circumstances where the parties accept that the contract is binding, and provides (amongst other things) for their agreement to be entirely expressed within it and that there are no side agreements (or the like) to qualify their bargain, but their evidence casts an entirely different light as to their real intentions and expectations and the real substance of their relationship. In short, whilst it will depend on the circumstances it is almost invariably the case the real substance must prevail.

Before considering the application of IFRS accounting principles and answering that conundrum, I turn to consider in more detail the Claimants ’ case that the impugned VAR sales and the way they were implemented followed a “ pattern”, which the Claimants relied on as the best proof of the parties ' intentions and the true substance of their arrangements.

(5) The alleged “ pattern” and the Defendants ’ response that it reveals “neutral features”

In parallel with their case based on the direct evidence of the VARs as to their understanding of the real substance of their engagement with Autonomy, Mr Rabinowitz described as “striking beyond belief” the ‘ pattern’ evident in every one of the impugned VAR sales indicating behaviour and its inconsistency with the ostensible contractual arrangements. He submitted that this ‘ pattern ’ could only be explained by reference to shared understandings between the parties as to the true economic substance and nature of the arrangements which were very different from those expressed in or apparent from their contract.

In this way of putting their case, the Claimants sought to focus on what the parties actually did (and did not do) as the litmus test of the true economic substance of the transactions and the existence of the shared intentions which in truth explained their true nature. This was a facet or development of Mr Holgate ’s description of economic substance as “ the commercial effect in practice and things like that” and “a common sense point of view [of] what is really happening in a transaction”. In their written closing, the Claimants described this as “a further string to their bow” ; but in his oral reply Mr Rabinowitz appeared to give primacy to this aspect of the Claimants ’ case on VARs.

As in the context of the witness evidence, the Claimants submitted that the conclusion should be that the impugned VAR transactions, though effective in law on the terms of the VAR agreements made in respect of them, did not reveal their true substance because they did not reflect how in fact the parties proposed to act. Once the way in which the parties acted was examined, it was contended that it was quite clear that nothing really changed after the VAR transaction: Autonomy continued its negotiations with the end-user, and the VAR did nothing to signify either a change of ownership, nor make any payment. This demonstrated that (a) it was Autonomy, and not the VAR, which in fact had ‘ managerial control ’ of the goods in question; and (b) the VAR had neither any role nor any real interest in an end-user sale, because in reality it was just a “placeholder” and would never be required to pay since Autonomy ’s true source of payment was always intended to be the end-user.

The Claimants suggested that a further advantage of this way of putting their case, in addition to the fact that, although complementary to, it does not depend on the evidence of witnesses who may be considered self-interested or parti pris , is that (so they submitted) if the facts do indeed show that the parties’ actual intentions and conduct differed from the ostensible effect of the written sale contract, then neither the entire agreement clause in each of the VAR agreements, nor the auditor confirmation letters, could “insulate the transactions from an accounting treatment which reflects the true facts.”

The Claimants contended that the plain inference from the ‘ pattern ’ revealed was that:

(1) in none of the impugned VAR sales did or was the VAR ever intended or even permitted to negotiate or close a direct deal with an end-user; and

(2) in none of the impugned VAR sales was the VAR ever intended and / or ever expected to pay out of its own resources.

Was there demonstrated to be a pattern showing that VARs did not, and were not intended to, negotiate or effect an end-user sale?

The circumstantial facts the Claimants relied on as establishing a ‘ pattern’ demonstrating that at the time the VAR transaction was entered into, no-one intended that the VAR should be involved in any attempt to on-sell the software licensed to the end-user, and that the true intention was that Autonomy should contract directly with the end-user, can be summarised as follows.

First, in all of the impugned transactions it was Autonomy, rather than the VAR, which was negotiating with the end-user at the time of the VAR transaction; and it was Autonomy, and not the VAR, which had the relationship with the end-user, often one of long-standing and involving repeat business. The Claimants’ submission was that it is inherently unlikely that Autonomy would have wanted to cede the conduct of that commercial relationship to the VAR. Further, they suggested that it was equally unlikely that the end-user would want to deal with an unknown (and in some cases, newly incorporated) VAR in those circumstances either. Both considerations pointed to the conclusion that the VAR was not expected to do anything beyond acting as a “placeholder”, the true arrangement both existing and prospective was between Autonomy and the end-user, and the VAR agreement in each case lacked any real substance as a sale.

Secondly, and as already noted from the unchallenged witness evidence, in none of the impugned VAR sales was there any sign of any effort being made to give the VAR the introduction and information necessary to enable the VAR to intervene and take over the pre-existing negotiations. That again supports the conclusion that no such involvement was ever in reality intended of the VAR.

Thirdly, there is no evidence of Autonomy ever asking the VAR why it had failed to make any effort to on-sell to the end-user. The Claimants suggested in this context that if Autonomy had really expected the VAR to attempt to sell the deal to the end-user, it would surely have expressed its disappointment at the VAR ’s lack of effort and regarded the VAR as the author of its own misfortune in missing out on an onward sale.

Fourthly, there is no evidence that in any of the impugned transactions did the VAR actually make any effort, after the VAR transaction, to achieve an onward sale of the Autonomy software license to the end-user; which once more supported the same conclusion that the VAR was intended to be entirely passive.

Fifthly, and perhaps most strikingly, Mr Rabinowitz emphasised especially that in none of the impugned VAR transactions (except perhaps VT20, see footnote 273 below) did Autonomy ’s negotiations ever actually bring about a contract between the VAR and the end-user. The end-user contract, if one eventuated at all, was invariably on terms (including as to price) negotiated and determined, not by the VAR, but by Autonomy, and in every case where a contract eventuated, Autonomy and not the VAR was the party to it.^{^[272]}

The Claimants submitted that such direct deals shed light on the true intention at the date of the VAR transaction: they gave rise to the inference that the reason why the VAR never ended up concluding a deal with the end-user was because the VAR was never expected to do that, and/or that Autonomy in reality retained managerial control throughout, with the VAR being a purely passive placeholder and any purported purchase and sale to it being illusory. As Mr Rabinowitz put it in his oral reply:

“we say that that pattern makes it very unrealistic to suppose that Autonomy ’s intention was solely to negotiate a contract between the VAR and the end-user. At times Autonomy tried to achieve that, but always failed. At other times it did not even try. It simply negotiated a direct deal between Autonomy and the end-user.”

The Claimants submitted in conclusion on this aspect that this ‘ pattern’ demonstrated that Autonomy must from the outset have been intended to have, and did in every case in fact have, exclusive control of negotiations with the end-user, and that the VAR had no say or involvement in the process; and that this in turn demonstrated, to use the wording of IAS 18.14(b), the retention by Autonomy of “managerial involvement to the degree usually associated with ownership” and “effective control” in each case “over the goods sold . ”

As to the second aspect of the ‘ pattern ’, which the Claimants alleged also gave rise to the inference that there was no true ‘sale ’, was that the facts demonstrated that the VAR was also not intended to be required to pay Autonomy from its own resources if no end-user sale eventuated, and that in reality Autonomy ’s true source of revenue was never intended to be the VAR but the end-user. Failing an end-user sale, Autonomy would have to take some step, whether by way of “round-trip”/reciprocal transaction, debt forgiveness or some other expedient to protect the VAR whilst preserving the fiction of revenue generation. The Claimants relied on the following.

First, the Claimants submitted that if it is found that it was intended at the time of the VAR transaction that the VAR would do nothing, that fact, of itself, was also a reason to infer that the VAR was also not intended to be required to pay Autonomy from its own resources. The rationale was this: if the VAR was genuinely expected to pay Autonomy from its own resources, it would be vitally interested in the end-user sale: the most obvious way for the VAR to put itself in funds would be by working hard to secure an end-user sale. To step back altogether, and leave further negotiation and completion of the end-user sale entirely to the seller, would be most unnatural, to the point of being commercially inexplicable, if the VAR was really on the hook.

Secondly, the Claimants relied on the fact that the VAR was first approached by Autonomy regarding a transaction involving an identified end-user at the very end of the quarter, and usually concluded the VAR transaction within a matter of hours thereafter, with no chance being given to the VAR to scope and assess the end-user deal. The Claimants rejected Dr Lynch ’s suggestion (which I examine further below) that this was entirely usual, on the basis that it was exceptional for such deals to be struck at the very end of the quarter without the VAR negotiating any of its terms. Further, having had no prior contact with the end-user, the VAR had no insight of its own as to the prospects of an end-user deal being concluded, whether at all or within a reasonable time, beyond whatever Autonomy may have told it. The Claimants referred in this regard to the evidence of Mr Welham when asked about this in re-examination:

“Q. In your experience as an auditor of software companies, would you consider it usual or unusual for a VAR to be first approached about doing a deal right at the end of or on the last day of the quarter?

A. Probably unusual.”

The Claimants suggested further that if it was genuinely intended that the VAR would pay from its own resources, the VAR would have had, and been keen, to carry out elementary due diligence: it was not realistic and not to be assumed that astute businessmen would ever act so recklessly.

Thirdly, the absence of any effort at all, given that in every case the impugned VAR sale was right at the end of the quarter, to even make time for price negotiation with Autonomy before executing the VAR transaction, made no commercial sense if it was intended that the VAR would have to pay Autonomy from its own resources; and all the more so given that the amounts involved would have presented an existential threat to the VAR in many instances had it truly been at risk.

In short, the continuing responsibility of Autonomy and the abdication of responsibility by the VAR demonstrated that Autonomy regarded its source of recourse and reward as being from the end-user, and the VAR, relying on Mr Egan ’s assurances, did not regard itself as being at substantive risk in the meantime.

Fourthly, the Claimants maintained that in many cases the VAR did not appear to have the means to pay Autonomy, in the absence of payment further to an end-user sale. They submitted that the fact that Autonomy must have understood at the time of the VAR transaction that the VAR could not pay supports an inference that it was not intended to do so, at least unless and until the end-user sale had been completed.

Fifthly, the Claimants relied on the fact that, leaving aside any modest deposit, Autonomy never did require any ‘friendly’ VAR, in any of the impugned transactions, to pay Autonomy out of its own resources prior to receipt of funds from the end-user or (in effect) from Autonomy itself in a ‘reciprocal transaction. ’ Instead, and as I describe and discuss at greater length later, Autonomy either caused the end-user to pay the VAR, so that the VAR in turn could pay Autonomy, or Autonomy found another means to excuse the VAR, whether by issuing a credit note, and cancelling the VAR ’s legal obligation; or by buying something from the VAR in order to put the VAR in funds; or by simply writing off the VAR ’ s debt.

The Defendants ’ position as regards the alleged ‘ pattern ’

The Defendants resisted all of this, and both limbs of the Claimants ’ case as to the VAR agreements ’ lack of economic substance, primarily on the basis that the Claimants ’ reasoning was flawed because it was based on false a priori assumptions. They contended that the Claimants placed undue reliance on features of some of the impugned transactions on the assumption that they revealed impropriety, whereas in reality they were entirely usual features of VAR transactions in the software field and were “ neutral” in terms of the criteria for revenue recognition.

In Dr Lynch ’s written closing submissions, the Defendants identified and sought to justify as entirely “ neutral” the following features of the impugned VAR transactions on which the Claimants had placed undue reliance:

(1) Deals being reached at the quarter end;

(2) Bringing in a reseller to enable revenue recognition in that quarter when an end-user deal could not be closed at a quarter-end;

(3) The status or prospects of the deal between the reseller and the end-user;

(4) The extent of price negotiations between Autonomy and the reseller;

(5) No “added value” by the reseller;

(6) Autonomy continuing to be involved with the end-user after the VAR sale;

(7) Sales to VARs of hosted products;

(8) Closing a sale directly between Autonomy and the end-user;

(9) Arrangements for the VAR to be paid by the end-user notwithstanding a direct contract between the end-user and Autonomy (which Dr Lynch referred to as “Designated payee situations”);

(10) Cancellations of a VAR ’s obligations after a direct deal had been closed between the end-user and Autonomy;

(11) Occasional write-offs of debts owed to Autonomy by resellers;

(12) Absence of any occasions on which Autonomy demanded payment from the VAR and sued when none was received;

(13) A VAR ’s unwillingness to pay; and

(14) The payment of MAFs notwithstanding that the VAR had added no value nor taken any part in the negotiations for and conclusion of the end-user sale.

Put broadly, the Defendants ’ case was that there is nothing unusual, let alone improper, in a VAR taking software on its books and committing in law to pay the purchase price, with a view to assisting Autonomy (for a fee and other commercial advantages) both (a) to recognise revenue on the immediate sale to the VAR and (b) to achieve a subsequent sale of that software to an end-user. In other words, a VAR may be, and in ordinary commercial dealings often is, nothing more than a guarantor (as, for example, Capax Discovery described itself in the Goldberg Segalla letter), or in other cases, a “ fulfilment partner”; and a sale to a VAR for that purpose is nonetheless a sale in economic substance as well as law.

The “neutral features” as they were identified and described in Dr Lynch ’s written closing submissions are differently organised and differently expressed ways of putting the converse to the Claimants ’ case that the ‘ pattern ’ reveals lack of substance and impropriety. The easiest, and for the present the most convenient, course is to explain how Dr Lynch elaborated his case in this regard, and then to assess together the two opposing perspectives in reaching my determination on this aspect of the case.

Quarter-end timing

First, the Defendants sought to dispel the suspicion that the Claimants tried to create with respect to the timing of the VAR deals. They contended that, contrary to the Claimants ’ case, quarter-end deals with VARs were and are common-place in the software industry. Their timing is not of itself indicative of impropriety; and that is so even if, as a corollary, there may well have been no discussions with the reseller previously, nor any time for the reseller to assess the prospects of an end-user deal, or the reliability of the end-user if the deal was done.

Recourse to a VAR simply to recognise revenue if the end-user deal was delayed

Secondly, the Defendants submitted that there was nothing improper or indicative of impropriety in Autonomy bringing in a reseller at a quarter end simply because a deal with an end-user could not be closed in time to enable recognition of the income to be derived from a sale of that software in the preferred quarter. Deloitte knew that Autonomy closed deals with resellers in these circumstances and there was no suggestion that this affected revenue recognition. Dr Lynch under cross-examination described this as “de-risking”: Autonomy was getting the commitment to pay from the VAR, constituting a “bird in the hand”. As Mr Miles put it in his oral closing:

“ it meant that Autonomy could actually get a commitment to the revenue on the deal without further delay or pressure on prices.”

Mr Welham confirmed in his oral evidence that Deloitte had no problem with Autonomy selling to a reseller for the purposes of recognising revenue. He accepted that the fact that Autonomy was unable to close the deal for the quarter might be a good reason for selling to the reseller.

No assessment of status or prospects of end-user deal by VAR

Thirdly, the Defendants submitted that the Claimants ’ focus on the status and prospects of an end-user deal was misplaced. Dr Lynch relied on the following passage from Taylor Wessing ’s letter of January 2015 on behalf of Deloitte in order to rebut HP ’s letter before action under the Professional Negligence Pre-Action Protocol:

“The relevant agreement for revenue recognition purposes is that between Autonomy and the VAR. It is not relevant for the purposes of revenue recognition whether the party with whom Autonomy has contracted is a VAR or an end-user. It is not necessary to identify an end-user (other than for licence control purposes.”

The Defendants pointed out that Mr Welham had agreed also that it was not necessary for revenue recognition purposes for Autonomy even to identify the end-user, still less for the VAR to have considered the prospect or commercial terms and rationale of any onward sale by the reseller to the end-user. The Defendants surmised that the Claimants, in suggesting otherwise, may have been influenced, and indeed misled, by the more familiar (at least to them, as for the most part, US entities) but different, rules of US GAAP which (putting the point broadly) preclude revenue recognition until the end-user sale. The Defendants repeated once more that under IFRS accounting, the test is a “sell in” test: provided the IAS 18.14 criteria are met in the case of the immediate sale to the VAR, revenue is required to be recognised. Under US GAAP, the test is a “sell through” test: revenue may only be recognised once the end-user sale is closed. Influenced by the US approach, the Claimants tended not to recognise or accept the reality and the consequence of the fact that even though the reseller would not be expected to do anything, it was for the purposes of both the law and the IAS (unlike US GAAP) Autonomy ’ s customer.

No price negotiation for VAR sale

The Defendants argued that a lack of negotiation on price between Autonomy and the reseller may be a neutral factor, and is so on the particular facts of this case.

The Defendants submitted that it is clear on the evidence that, as a matter of law at least, all the resellers were genuinely on risk in respect of their purchases, and took the risk in anticipation of advantage to themselves. Autonomy was coming to them with a purchase opportunity which they were happy to accept on the terms that were offered, as Mr Baiocco made clear in his evidence.

VAR sale but no added value

Dr Lynch ’s fifth point was that although the appellation “Value Added Reseller” may encourage the contrary supposition, a VAR sale is not improper or lacking in substance so as not to merit recognition of its revenue just because the reseller was not to add value to the sale in the sense of enhancing the goods sold or providing services to the end-user.

According to Dr Lynch a VAR transaction would nonetheless qualify as a proper sale even if it was not intended that the VAR should fulfil any role at all beyond purchasing the goods on the terms of the VAR agreement, going on risk for that purchase by accepting thereby the unqualified commitment to pay, and thereafter fulfilling the end-user's purchase order.

Deloitte accepted that it would be a legitimate and normal reseller transaction if the only role of the VAR was to purchase from Autonomy, going on risk in the process and thereafter to sell to the end-user. Thus, for example, Deloitte ’s Defence in the FRC proceedings states as follows:

“Contrary to the impression given in the Formal Complaint, this is not to say that a VAR has to perform additional services or “add value” in every sale insofar as that is intended to mean “provide additional services”; it can and often does simply re-sell software licences to customers, acting as a reseller.”

In his evidence under cross-examination, Mr Welham accepted that this was Deloitte ’s actual perception at the time:

“Q. …The VAR could simply act as a reseller, it didn ’t have to provide any additional services?

A. Yes.

Q. And Deloitte understood that at the time, that was part of your understanding at the time of the audit?

A. That could happen, yes.”

Thus, the Defendants submitted that although the Claimants spent some time in argument and cross-examination questioning what services or added value a VAR would be likely to bring to the relevant transaction, that was based on a false premise: it is neither unusual nor improper in the business for a VAR to be what Mr Szukalski called a “ fulfilment partner” with its role being simply to enter the contract for the purchase of the goods, go on risk for the purchase, leave Autonomy alone to negotiate with the end-user and thereafter fulfil the end-user ’s purchase order. In effect, only the “paperwork” would go through the VAR.

In the same context, the Defendants also relied on Mr Szukalski ’s evidence that all this was “very common practice”. However, as I elaborate later, FileTek ’s only deal with Autonomy was not really a true VAR deal; FileTek was not really a VAR in any usual sense; and Mr Szukalski did not really operate in that business at all.

Autonomy’s involvement with the end-user after the VAR sale

Dr Lynch ’s sixth point was his contention that it did not undermine revenue recognition for Autonomy to continue to negotiate or be involved with the end-user after the reseller sale.

Issues with regard to managerial control could potentially arise if Autonomy had been contractually obliged to exercise managerial duties with respect to the goods, since that obligation would connote that managerial responsibility and effective control of the goods had not been transferred under the contractual terms; but this was not in fact ever the case, and would have been inconsistent with the provisions of the contract, which made it clear the risk passed in all respects to the purchaser, and did not impose any relevant continuing obligations on Autonomy.

Further, in circumstances where the VAR was simply fulfilling that role, it is not surprising or undermining of the VAR ’s purchase that, with the VAR ’s consent, and without obligation on the part of Autonomy, Autonomy should be conducting the subsequent discussions with the end-user.

Dr Lynch, supported by Mr Hussain, presented this as the unobjectionable provision of voluntary assistance by Autonomy, with a plain and obvious and entirely proper commercial rationale; as he put it in his first witness statement:

“Many of the end-users were repeat customers with whom Autonomy had other business, customers often buying the same product over and over again. It would make no commercial sense for Autonomy not to continue discussions with the end-user. If a customer bought millions of dollars ’ worth of products from the company every year, the company would not stop talking to them because a particular sale for a couple of hundred thousand dollars had been made to the reseller who was going to on-sell it to that customer. The reseller was on risk and owned the relevant product. In order for the sale to be recognised, Autonomy was not required to step out of the picture altogether. In fact, Autonomy ’s continued involvement in the deal could increase the collectability of the debt from the reseller, I understand, based on my discussions with Autonomy ’s Finance Department, and the auditors at the time, that there was no issue from an accounting perspective with Autonomy voluntarily providing assistance to the reseller in closing an end-user deal, or leaving the reseller to complete the sale alone. That made perfect sense to me.”

However, in all this the impression given by the Defendants was that the objective and expectation was that the ultimate sale to the end-user would still be expected to be made by the reseller fulfilling the purchase order from the end-user, and that Autonomy ’s involvement was to facilitate that result. The Claimants ’ point, to which I shall return, was that (apart from the two ‘exceptions ’ identified and discussed above which, on analysis, were not really exceptions at all) in not a single case was that the result: any end-user deal which eventuated was not only entirely negotiated by Autonomy but also a direct deal with Autonomy.

Sales to VARs of hosted products

Dr Lynch contended that the fact that some of the impugned VAR sales (such as the deal with Capax Discovery for end-user Kraft, “VT3”) involved sales of licences of products likely to be (and arguably only capable of being) used in the hosted environment, and in particular, Digital Safe, indicated no difficulty either: the licence remained a valuable asset and its sale was substantive and legitimate, so that revenue could be recognised from it accordingly whether or not it could be brought on site by the end-user.

Further, Deloitte were aware that some of the reseller deals were of licences in connection with hosting arrangements and do not appear to have objected to the recognition of revenue in respect of them.

Transactions going direct

The Defendants submitted that the Claimants had no basis for calling into question, or seeking to draw adverse inferences from, the recognition of revenue at the point of the VAR sale in certain cases where thereafter a direct deal was made between Autonomy and the end-user, and the revenue received on the end-user sale would either be remitted to the VAR or the VAR would be treated as released. They presented this as an occasional occurrence, and never Autonomy ’s objective in continuing to be involved with the end-user (see above). According to the Defendants, only

“very few transactions became direct deals; based on a historical average of 2009/2010, fewer than one transaction per quarter”.

The Defendants also made the point that the occasional replacement or reversal deal of that sort was not objectionable, either generally or in the view of Deloitte. That is illustrated by Deloitte ’s approval of the introduction into certain VAR purchase orders in Q4 2010 of a clause expressly contemplating, permitting and providing for direct end-user sales.

Dr Lynch additionally drew attention to Deloitte ’s own Defence document in the proceedings against it brought by the FRC, where as regards the new clause, it was stated as follows:

“The clause therefore envisaged that on occasion Autonomy might enter into a transaction with a VAR, but subsequently deal directly with the end-user. In those circumstances, once Autonomy had received payment from the end-user, it would pay a fee to the VAR.

The audit team concluded that this new clause did not affect revenue recognition. Upon the VAR entering into the agreement, it had accepted the risks and rewards of ownership. It had possession of the licence, and the ability to sell it to the end-user without requiring the involvement of Autonomy. It was legally required to pay Autonomy. The existence of the new clause did not change the analysis in respect of IAS 18 paragraphs 14(a) and (b) as the risk remained with the VAR, unless a direct deal was concluded with the end-user, a risk the VAR could not control.”

However, the Defendants acknowledged that, by reference to two VAR deals with MicroTech in Q4 2009 where there was a direct sale to end-user (ManuLife and Morgan Stanley), Deloitte highlighted a concern about “ revenue reversals” in its Audit Committee reports for Q1 2010 and Q2 2010. Deloitte had warned that “significant evidence of such further revenue reversals may jeopardise management ’s ability to recognise revenue at the point of sale to the reseller ”.

As again I return to later, the problem for the Defendants is that the truth is that direct deals and revenue reversal were or became characteristics of end of quarter deals transacted for revenue purposes, rather than occasional unexpected expedients. Their contention that “very few transactions became direct deals; based on a historical average of 2009/2010, fewer than one transaction per quarter” (which was based not, of course, on an analysis of impugned VAR transactions but on a comparison with the entire corpus of VAR deals) did not really address the point that 19 out of 30 of the relevant deals went direct, and in the rest, no end-user deal eventuated at all: see paragraph 2152 below.^{^[273]}

VARs as ‘designated payees’ in respect of direct deal between Autonomy and end-user

Another post-transaction event referred to by the Claimants with suspicion, but which the Defendants contended was unobjectionable and certainly not such as to upset earlier revenue recognition, is where there was a “designated payee”. That is to say, transactions where the end-user was either asked ad hoc for, or (from 2010 onwards) a contractual term in the end-user sale between it and Autonomy required, payment of the purchase price to the reseller as Autonomy ’s designated payee.

A variant was in the VAR transaction with DiscoverTech for end-user Citigroup, in the context of which the end-user wished to deal directly with Autonomy and Deloitte noted:

“Autonomy resolved the issue with the VAR by entering into a tri-party agreement, whereby Citigroup paid Autonomy in full and Autonomy then paid the VAR their margin based in the amounts due [i]n the original PO”.

I have considered whether Deloitte ’s approval of that sort of mechanism suggests acceptance by them, or apparent acceptance by them so as to falsely reassure the Defendants, of a more common-place practice of ‘direct deals ’, so as to invalidate or neutralise the professed concern about direct deals which became an issue in 2010. I have concluded that it does not.

It appears to me that Deloitte always regarded a direct deal as the exception and the designated payee provision as designed to cater for the exceptional case where notwithstanding negotiation and conclusion of an end-user deal by the VAR, the end-user nevertheless insisted on completing with the software producer/supplier (namely, Autonomy). That was never so in any of the direct deals following impugned VAR sales.

Autonomy occasionally cancelled a reseller ’s obligations by providing a credit note. Again this was a post-transaction event, which the Defendants maintained could not have affected the original recognition of revenue. In circumstances where the end-user had chosen to deal directly with Autonomy so that Autonomy received payment for the software from the end-user it was a sensible business decision to cancel the reseller deal. The alternative would have been to leave the reseller on the hook despite Autonomy having taken the customer opportunity. This would have been damaging to the relationship.

Of course, where an end-user contracted with Autonomy, the revenue was not recognised twice. Revenue was not recognised again on the end-user sale: Deloitte understood this and was satisfied. The Defendants adopted substantially the same approach and argument in the limited number of cases where the end-user was told to make payment to the reseller to permit the reseller to make payment to Autonomy.

Dr Lynch summarised the position as he saw it in his first witness statement:

“From my vantage point, Autonomy's reseller relationships operated without any major difficulties. I had a general understanding that, on occasion, reseller deals departed from the conventional pattern of "Autonomy sells to reseller/reseller sells to end-user", for example, where an end-user decided to purchase from Autonomy directly. HP has cherry picked 37 transactions to try to identify deals of this nature. In those circumstances, Autonomy had a couple of options. We could say to the end-user that we would not supply them and direct them to the reseller, but we would avoid doing this if it was likely to cause problems in our relationship with the end-user, particularly if the end-user was a repeat customer. I had a general awareness that Autonomy could take the commercial decision to supply to the end-user and unwind the reseller deal or designate the reseller as Autonomy's payee, although I would not have been the person at Autonomy to make that decision. In those circumstances, Autonomy would generally pay the reseller a MAF for their lost margin on the end-user deal. I now know that this happened in a handful of cases, a very small percentage of the overall total of sales to resellers. Autonomy could also have held the reseller's feet to the fire and demanded payment from the reseller, even though Autonomy had sold directly to the end-user and received payment from the end-user. The consequence would be that Autonomy would have double the cash and double the revenue on the sale of the same software, which would seem inappropriate. That rarely made commercial sense. Either way, it was a commercial decision for Autonomy at its discretion. Despite these complications, it was my understanding, then and now, that Autonomy's products were almost always sold to and used by the end-users, and that cash was received against the revenue recognised (whether from the reseller or the end-user).”

In the Taylor Wessing/Deloitte letter of January 2015 it was stated:

“Deloitte did in fact carefully consider and review each of the situations in which a contract with a VAR was cancelled and the deal was completed directly with the end-user, in order to understand the reasons for the cancellation and entry into a direct contract with the end-user. This included ensuring that the sale had not been double counted. With the exception of the Capax / Kraft transaction, which had no impact on the 2009 year end results, and which was the first time Deloitte had seen such a cancellation, these transactions were also all raised with the Audit Committee.

The fact that the earlier transaction with the VAR was ultimately cancelled does not however change the fact that in each case it had been appropriate to recognise revenue on the original contract between Autonomy and the VAR. In each case, as at the time of considering the original transaction with the VAR, there was a binding and non-cancellable agreement in place between Autonomy and the respective VAR, and under that contract the risks and rewards had passed to the VAR.

It is not appropriate in such circumstances to reverse the earlier recognition of revenue. The revenue was properly recognised. Circumstances then changed. As Deloitte was ultimately satisfied with the circumstances of the cancellation of the original contract with the VAR, there was no need to consider the need for Autonomy to adjust and re-issue the previous quarter results.”

[Underlining was in the original]

The Defendants also relied on Mr Welham ’s confirmation in cross-examination that the fact that Autonomy cancelled a deal with a reseller did not impact on the original revenue recognition, because at the time of the reseller deal there was a binding and non-cancellable obligation on the reseller to pay. It is of some note, however, (and I shall return to this) that Mr Welham added the following caveat, which though a reference to it was provided in a footnote, was given less prominence in Dr Lynch ’s closing submissions, and sought to be minimised by a statement which I do not accept (and see paragraph 2153 below) that “in the event the instances of direct deals were rare” :

“…it is important to highlight though what we did highlight to the audit committee at the time, which is that clearly if you have a continuing trend of that, then you build up a bank of evidence that would suggest that you wouldn ’t recognise revenue on the full 100%....”

Write-offs

The Defendants eleventh example of a “neutral feature” which did not call into question revenue recognition was what they described as the occasional need to write off debts owed by resellers. They ascribed this need to the occasional vicissitudes of business ( “Bad debts arise in business” ), and such an event long after the VAR deal in question could not sensibly affect the accounting analysis.

This is another matter to which I shall have to return, not least because (a) Mr Welham made clear that credit notes and write offs would sound a warning note if they were other than occasional; (b) most of the write-offs were not “ occasional” but done in the “Dark Period”; and (c) bad debt write-offs and credit notes resulted in $107.9 million of sums accounted for as owed to Autonomy and counted as recognised revenue being removed from its balance sheet.

Not suing debtors

The Claimants alleged that the reseller was not “pursued” for outstanding debts. The Defendants maintained that the documentary record in fact shows that resellers were chased up frequently; and that in any event, it is not commercially surprising if Autonomy did not race to court when there was an outstanding debt from its partners. Deloitte monitored debtors and recognised that resellers ’ repayments were being delayed in circumstances where the end-user deal had not closed, or amounts had not been received (see below).

Reseller unwillingness to pay

Nor, the Defendants submitted, was there anything in the Claimants ’ complaint that there were instances when Autonomy did not pursue the VAR for the outstanding debts in the meantime (and there were cases, which the Defendants depicted as occasional, in which it had made clear that it would not pursue the debt).

Whereas the Claimants relied on this as referable, not to unexpected commercial exigencies, but to the understanding between Mr Egan and the relevant VAR that there was no intention or expectation of the VAR ever being required to pay out of its own resources, the Defendants characterised this as ordinary commercial positioning and behaviour, which did not signify that the VARs were not on risk. They also relied on the fact that Deloitte monitored debtors and aged debt: and did not make any objection on that score.

Almost as a throw-away additional point, the Defendants suggested that “in point of fact the documentary record show that VARS were chased up frequently” : but although there were a few examples showing this, chasing letters were sent by members of the Finance Department who were unaware of the arrangement in place, and who were told to desist by those in the know (and see paragraph 2153(4) below).

MAFs

I have referred above to the payment of MAFs but it is convenient to bring together the strands of the Claimants ’ case in this regard before addressing the Defendants ’ case that they were a yet further example of “neutral features” distorted and incorrectly criticised by the Claimants.

The Claimants ’ case on the payment of MAFs was not that they were intrinsically abnormal or inherently objectionable, but that none of the VARs did anything substantial to merit their payment. The Claimants focused especially on the following aspects:

(1) The absence of any contractual provision for, and yet the expectation on the part of VARs of, payment of a MAF, which the Claimants submitted illustrated how such arrangements between Autonomy and its VARs were established informally, outside the “four walls of the contract”, and despite express contractual provisions precluding side arrangements. This, they submitted, demonstrated a readiness on the part of both Autonomy and the VARs to keep important parts of their agreement out of the written contract. They also made the more general submission that the fact that this important element of the overall arrangements was not expressed in the VAR agreements demonstrated that it would be wrong to assume that there is anything improbable about the notion that the parties might have agreed upon important matters that were not recorded in the written contract.

(2) Secondly, the Claimants also criticised some of the payments of MAFs to resellers when in the event a direct deal was done and the VAR deal was, in effect, replaced. They contended that in such circumstances, the VAR had in reality done nothing to warrant reward (still less provided any marketing assistance), had been released from any risk or obligation, and deserved no reward. They submitted that this was a concrete illustration of “Autonomy…buying revenue again” .

(3) Thirdly, the Claimants relied on instances where (as was for example the case in the Capax Discovery/Eli Lilly transaction) the VAR ’s entitlement to a MAF was recorded in writing in terms which (they submitted) did not fairly and honestly represent the true reason for the payment, thus further demonstrating Autonomy ’s propensity dishonestly to misrepresent the true nature of its payments and inducements. The Claimants provided as an example a letter drafted by Mr Kanter in the course of the Capax Discovery/Eli Lilly transaction which they contended was typical and stated as follows:

“To formalize our prior discussion, Autonomy Systems Limited ( “Autonomy”) and Capax LLC ( “ Referral Partner”) agree to the terms and conditions of this letter agreement ( “Agreement”) as follows:

Referral Partner will: (1) introduce Autonomy into the deals with Eli Lily ( “ End-user”); (2) obtain quotes from Autonomy on behalf of the End-user; and (3) work with the End-user to assist in executing purchase orders and contracts with Autonomy.

Autonomy will: (1) pay Referral Partner commissions in the amount of US$629,000, as a result of Referral Partner ’s direct and proximate participation in the account; (2) deliver products directly to End-user; and (3) use reasonable efforts to provide mutually agreed upon sales assistance. …”

The Claimants submitted that the letter was plainly designed to give the impression that it was formalising prior discussions as to what role the VAR would perform - and that, in return for the performance of that role, Autonomy would pay the commission; and that was so although by the date of the MAF letter quoted above (30 September 2010), it was perfectly clear that Capax Discovery had not done any of the things the letter said it had agreed to do. The Claimants gave another example of an email from Mr Kanter approving a MAF on the Capax Discovery/FSA transaction, in which Mr Kanter stated that he had been impressed by Capax’s contribution whilst in fact knowing there had been none.

The latter point posed two rhetorical questions, one bearing on impropriety and the second bearing on the Defendants ’ knowledge of it:

(1) First, why, if Autonomy had nothing to hide in relation to the reasons for this payment, Mr Kanter chose to resort to relying on a fictitious basis for the payment and the documents under which they were made. In connection with the payment of MAFs the Claimants also contended that the indications of what they characterised as pre-textual emails seeking to describe why a MAF was being paid indicated the substantive impropriety of the payment.

(2) Secondly, whether it is likely that they were intending to mislead Dr Lynch, or whether it is more likely that there was a discussion in which Dr Lynch knew and understood the real reason why the VARs were being paid and that it was better dressed up.

The Defendants submitted that the Claimants ’ premise, and the assumptions they had made and required their experts to make, was that (a) any impugned VAR transaction was improper and thus (b) any MAF that was paid, being (on that assumption) a fee paid in return for some improper revenue recognition, was another facet or demonstration of impropriety; but there was no basis for these assumptions, and nor was the inference justified.

The Defendants submitted that, contrary to the Claimants “ circular” approach, the starting point should be, and the evidence was, that the VAR did take on risk, for which it expected reward; and that in those circumstances, the payment of a MAF was a sensible and appropriate course where for one reason or another the reseller which had gone on risk could not otherwise make its margin (for example, where Autonomy and the reseller eventually made a direct deal, and the VAR lost its customer).

It was not disputed that MAFs were common in the software industry, usually in the range of 10% to as much as 30% of the deal price and something HP itself employed. The Defendants reminded me especially that Deloitte were aware that Autonomy paid MAFs, including in situations (such as the Kraft deal addressed in more detail below) where the deal went direct and the reseller was being compensated for lost margin.

Mr Welham confirmed that he did not think that the fact that fees were paid to VARs for their agreeing to become fully on risk for VAR deals was improper ; and he confirmed in his evidence that if a deal had gone direct and Autonomy thereafter agreed to pay the reseller a commission or payment fee notwithstanding that the risk they had undertaken would no longer eventuate since they had been released, this was essentially a commercial decision for Autonomy and did not affect revenue recognition.

The Defendants also drew my attention particularly to the following passage from the Deloitte memo on changes to Autonomy’s VAR agreements:

“In 2009, four deals with Morgan Stanley ($4.6m), Eli Lilly for ($6.1m), Kraft ($4.3m) and Manufacturers Life Insurance ($1.1m) that were initially sold through VARs Microtech (Morgan Stanley and MLI) and Capax (Kraft and Eli Lilly) were eventually done through Autonomy, Kraft was signed with Autonomy in Q4 2009 and the remainder were signed in 2010. In 2010 two deals initially sold through Discover Technologies (Phillip Morris International $2.9m, Citigroup $5.5m) were eventually done direct with Autonomy.

In a normal deal, the VAR signs with Autonomy, then signs a (slightly higher value) deal with the end-user, making a margin on the way. However, when the end-user decides to sign direct with Autonomy, then there is a contract between the VAR and Autonomy, a contract between the end-user and Autonomy, but no contract between the VAR and the end-user. In this situation, per the explicit terms of the VAR agreement, the VAR would owe Autonomy for the deal signed; without having a customer.

The issues were resolved in a way that ensured the VAR retained the margin they would have got, had the end-user actually signed with them. In the example of Citigroup, Autonomy resolved the issue with the VAR by entering into a tri-party agreement, whereby Citigroup paid Autonomy in full and Autonomy then paid the VAR their margin based on the amounts due on the original PO.

These events highlighted the issue of risk for the VAR ’s - relying only on the goodwill of Autonomy towards its resellers rather than formal legal recourse was deemed to be too high, and hence the clause was agreed.”

Mr Egan ’s evidence was that MAFs were justified in either context, and that he saw nothing wrong with paying them. In cross-examination he said this:

“Q. Now, from time to time Autonomy paid what was called a marketing assistance fee to a VAR, do you remember that?

A. I do

Q. And that was a way of compensating the VAR for its agreement to take on the liability to pay?

A. Correct.”

From Mr Baiocco ’s and other VARs stated perspective, there was nothing untoward, unusual, or still less improper, in being paid for taking on the risk which the VAR agreements unequivocally imposed on the VAR. Mr Baiocco saw it as a fee for the risk undertaken, sometimes in effect as a guarantor (as it was put in the Goldberg Segalla letter) (and see paragraph 2024 above). If he was released because a direct deal was substituted that position was historic, but nonetheless real; and the VAR should properly be compensated for lost margin. He never read or cared about the various letters that referred to MAFs.

The Defendants ’ answer to the second point in paragraph 2133(2) above was thus that the MAF was in each case in effect a fee for standing as guarantor, and thereby enabling Autonomy to recognise revenue at a time that was commercially advantageous to it; and that there was nothing wrong with that. But the converse of the Claimants ’ argument against a MAF was circular insofar as it depended on establishing impropriety but likewise the Defendants ’ argument was premised on the VAR actually being at real as distinct from nominal risk (which, of course, the Claimants submitted was never the case).

The Defendants did not substantively address the Claimants ’ first point (see paragraph 2133(1) above) as to the absence of any contractual provision for payment of a MAF and what that demonstrated as to the parties’ real attitude to the “entire agreement” clause and the prohibition of any side agreement or understanding. I recognise that neither such a clause nor such a prohibition precludes uncovenanted payment; but the convention of payment of a MAF does resonate with the alleged convention that the VAR would never be called on to fulfil its obligations.

As to the two questions raised in paragraph 2135 above, the Defendants sought to dismiss as “an overblown forensic point” the Claimants ’ criticism of the various ‘MAF letters ’ and the suggestion that, by referring to the VAR as a “ Referral Partner” and stating that the payments were for “direct and proximate participation in the account”, they were apt and designed to mislead. Dr Lynch contended that the letters were in standard form for a MAF and not bespoke; each contained language which was properly usable, even if not entirely suited to the particular case, and also contained language imposing a restriction on competition from the reseller, which was of value to Autonomy, supporting their substance.

In other words, the Defendants ’ answer was that there was no intent to disguise: the explanation (use of a standard form) was more prosaic. But that does not answer the email referred to in paragraph 2134 above; nor to my mind, what happened in various of the transactions, including the impugned VAR sale for end-user FSA. Thus:

(1) In his email (sent to Mr Hussain and Mr Egan, dated 6 October 2010 subject “ Capax” ) Mr Kanter stated:

“Having been impressed with Capax ’s contribution to the FSA transaction, I am comfortable that they have earned a marketing assistance fee in line with our standard terms. I have prepared the attached to document properly the transaction. Please can I have your views.”

Mr Egan responded “Agreed” and Mr Hussain responded “ok”.

(2) The impression given was of a specially produced draft recording the particular transaction and of a payment in exchange for a contribution to the end-user transaction with the FSA.

(3) Dr Lynch told me in cross-examination that although he was not involved in the transaction, he did know that Capax Discovery “were doing a lot of the work to actually show the technology and use it for FSA and that would be a very helpful contribution”. However, there is no record in the evidence of Capax Discovery having done such work in 2010 and Dr Lynch accepted that if there was none, his assertion would fall.

(4) The letter sent to Mr Baiocco by Mr Kanter dated 6 October 2010 which was in similar terms to one sent in respect of the Eli Lilly sale (see paragraph 2133(3)), described Capax Discovery as Autonomy ’s “ Referral Partner” and recorded that Capax Discovery was to introduce Autonomy into deals with the FSA; obtain quotes from Autonomy on behalf of the FSA and assist in executing purchase orders and contracts.

(5) Mr Baiocco ’s evidence (which was not challenged, and which I accept) was that Capax Discovery “did none of these things; and I never discussed any of this with Autonomy”. In further unchallenged evidence, which I also accept, he stated:

“In actual fact, we did not make any attempt to license Autonomy software to the FSA or participate in setting the terms of the license that I understand Autonomy sold to the FSA”.

The Claimants made the further point that in opening they had especially sign-posted that they wished to cross-examine Mr Kanter about the MAF letters (amongst other things), and were deprived of the opportunity to do so because Dr Lynch then decided not to call him. They invited the inference that Mr Kanter was not called because (a) he could give no honest explanation of the misleading way in which the MAF letter agreements were drafted and (b) he would have implicated Dr Lynch. As to (a), I prefer to approach the matter on the simple basis that I have been left to determine the matter on the basis of the balance of probabilities without the benefit of any contrary evidence from Mr Kanter, or any verification of an innocent explanation. On the evidence available, and in all the circumstances, I consider that it is more likely than not that the letter was in crafted (and not template) terms, and was intentionally misleading. I return to (b) later.

As will appear from my assessment of individual impugned VAR transactions later in this judgment, the FSA example was typical of the payments of MAFs made to “friendly” VARs, with variations as to the manner of payment according to whether or not any end-user eventuated. As will be seen, their payment was a pattern; the justification set out in the letters sent to VARs to document their payment was invariably included as a pretence to the reasons for their payment; and that pretence was kept up to both the Audit Committee and Deloitte.

Summary of Defendants ’ position

In the round, the Defendants sought to present a position that even if in many instances of impugned VAR sales Autonomy was involved in negotiations with the prospective end-user, Autonomy ’s objective was always voluntarily to assist the VAR to obtain and close a contract between the VAR and that end-user; and that even though there were occasions on which the culmination of the negotiations was that the end-user preferred to contract directly with Autonomy, none of this revealed anything such as to indicate that for the purposes of IAS 18.14, “the risks and rewards of ownership of the goods” had not transferred to the VAR, nor signified that Autonomy retained “managerial involvement or effective control over the goods sold” to “the degree usually associated with ownership.”

They ascribed the various ways in which VARs were released from liability as part of the ordinary course of commercial dealings necessitated by unexpected commercial exigencies.

Dr Lynch also noted that for the vast majority of the impugned VAR transactions Autonomy did receive the cash. As stated in his written closing submissions, of the $176,400,000 of revenue which the Claimants allege in Schedule 3 of their Statement of Case was reported by Autonomy on those deals, Autonomy received approximately three-quarters in cash. The dispute, of course, was as to how such payments had been funded, though the Defendants rejected the Claimants ’ allegation that in no case had the VAR provided funds which did not originate from some transaction with Autonomy such as a ‘ reciprocal ’ or ‘circular ’ transaction.

As part of this presentation, the Defendants also sought to depict the direct contracts between Autonomy and an end-user, and the “neutral features” identified above, as occasional, and not part of any invariable or even standard ‘ pattern ’. The Defendants did not really offer an answer if those features presented in a ‘ pattern ’, as was the Claimants ’ case that they did; and they offered nothing except assertion to the contrary, and reference without any evidence to it being atypical of an unidentified mass of VAR transactions which were not impugned, to contradict the evidence which emerged that in not a single one of the impugned VAR sales was the VAR actively involved in negotiating with the end-user, and in not a single one did a contract between the VAR and the end-user result. In every instance, any end-user contract was between Autonomy and the end-user, and in 11 out of 30 no contract eventuated at all.^{^[274]}

Thus, and as in some instances already observed in the course of examining what the Defendants averred to be “neutral features” ,

(1) Although the Defendants submitted that it would not prevent revenue recognition if materially all the interaction post-sale was between the end-user and Autonomy, they steered away from the reality that time and again, Autonomy was not assisting the VAR: it was excluding the VAR, and dealing exclusively with the end-user as if no VAR transaction had occurred.

(2) Linked to that, whilst the Defendants submitted, and Mr Welham agreed, that the prospect (or in the case of deals subject to the revised contract wording, express contemplation) that there might be a direct deal between the end-user and Autonomy did not prevent the risk being transferred to the reseller at the time of the deal, and nor did the occasional decision to deal directly at the insistence of the end-user, Deloitte warned the Audit Committee, and the Defendants appear to have appreciated, that the opposite conclusion would follow if in truth a direct deal was the intention and a routine or repeated result.

(3) The Defendants ’ assertion that the fact that in most cases a transaction between Autonomy and the end-user resulted made clear their awareness of the importance of the result in terms of the propriety of revenue recognition: the actual position, which was that no such transaction ever eventuated in the case of any impugned VAR transaction tells its own story against them.

(4) Dr Lynch ’s suggestion that “in point of fact the documentary record shows that VARS were chased up frequently” was obviously directed to buttressing the notion that Autonomy ’s behaviour did not depart from an ordinary commercial norm, with Autonomy doing what might be expected in the case of a reluctant payer. However, in the limited number of cases in which Autonomy did chase payment, the efforts were unsuccessful and not pursued; and, for example, when, in respect of the Eli Lilly transaction (VT4), Mr Baiocco complained on behalf of Capax Discovery that, contrary to his understanding of the basis of the VAR transaction, he was being pressed for payment before receiving any payment from Eli Lilly, Mr Egan interceded and the calls pursuing payment were stopped.^{^[275]}

(5) Whilst the Defendants submitted that the occasional cancellation of a VAR deal, and the issue of a credit-note and payment of a MAF to a VAR, did not necessarily impact on the original revenue recognition, Deloitte warned, and the Defendants appeared to accept, that if that became the norm or routine, again the assessment could well change. Their attempt to depict the issue of credit notes and very substantial write-offs on the basis that “bad debts arise in business” and to pass off the write-off and credit notes issued in the short “Dark Period” as a “normal housekeeping exercise” , reveals an acute consciousness of this; and the like applies to the cancellations.

(6) With particular reference to the write-offs made and the credit notes issued in the “Dark Period”, debts written off (as shown in Autonomy’s main general ledger (“DDS”) (included $2.3 million owed by MicroTech (VT13) and some $2 million owed by Realise (VT15) as well as £1.6 million owed by Sales Consulting (VT9) and £1.8 million owed by Red Ventures (VT19)) – the first two being transactions impugned by reference to a side agreement, and the latter two being “Collectability VARs). Ms Harris originally tried to claim that all of these debts had already been fully provisioned, so that the write-offs made “no difference to profit and loss, it’s just a balance sheet tidy-up exercise.” However, when she was shown an analysis that made clear there had been no provision for Red Ventures or MicroTech, and only a partial provision for Sales Consulting, she had to accept that these were substantive write-offs, and no mere “tidy-up”.^{^[276]} Other impugned VAR transactions that led to write-offs in the Dark Period, as shown in the same analysis included VT14 (£2.7 million net write off shown), VT17 ($1.7 million net write off shown), and VT22 ($1.1 million net write off shown). The list of credit notes included several credits to VARs, including Capax Discovery (through its parent company Capax Global), DiscoverTech and MicroTech, including (for example) $4 million in respect of VT25.^{^[277]}

My conclusion as to whether a pattern emerges and what is revealed

The Defendants accepted that the transactions did, for the most part, consistently evince the features identified by the Claimants. The real dispute was whether those features were “neutral”.

Perhaps unsurprisingly, the Defendants could point to features which were not, in themselves, unusual. I accept that it is not unusual or objectionable, of itself:

(1) For a software seller to effect a genuine sale to a VAR at the very end of a quarter in order, for example, to avoid being squeezed on price by a prospective end-user hopeful of exploiting (entirely legitimately) quarter, half or year-end pressures on the software seller to get a better deal;

(2) For a software seller to effect a genuine sale to a VAR after failing to conclude a transaction with the end-user before the end of the relevant accounting period (as the Claimants conceded in their Reply and Mr Welham accepted);

(3) For the VAR, in the context of such a deal late in the quarter, to have no time to assess the status or prospects of an end-user sale, or even the identity of the end-user, otherwise;

(4) For there to be no negotiation on the VAR sale price where there was no real issue as to the prospective onward sale price (typically because of a standard price or price range for the software sold);

(5) For the VAR in certain instances to act merely as a fulfilment partner (usually when acting, as FileTek did, as an approved intermediary for end-sales to US government institutions such as USDVA) and not to add any real value beyond processing the paperwork;

(6) For the software seller to be involved with the VAR in negotiating an end-user sale (for example in the case of an established customer of the software supplier);

(7) For there to be occasions, though they would be rare, where the end-user decided it would only deal directly with the software supplier and some arrangement would have to be made with the VAR to ensure that it was not out of pocket and that Autonomy was not paid twice;

(8) For there to be occasions (again rare) where the end-user might be instructed by Autonomy to pay the VAR instead of Autonomy again to avoid double exposure or double recovery;

(9) For the VAR to be issued a credit note for similar reasons (though again this would be rare);

(10) For the VAR to be paid a MAF either as reward for its assistance or compensation (if a deal went direct) for lost margin.

However, it is unusual, and both revealing and potentially objectionable:

(1) For the VAR to have no expectation at any time of any involvement at all in any discussions or dealings with the prospective end-user or any other potential end-user;

(2) For the VAR in fact to be entirely passive and cede entirely to the software seller all negotiations with the prospective or other end-user and to take no part in them;

(3) For the VAR not to expect to be consulted about the terms or even told of them (until after the event), or be party to any contract concluded by Autonomy with an end-user, or be involved in the delivery of software to such end-user;

(4) For the VAR routinely to agree to payment instalment dates for which it routinely made no apparent provision and failed to pay in time, as if they were never intended to have any actual effect^{^[278]} ;

(5) For the VAR not to be pressed or even reminded by the software seller about failures to pay on instalment payment dates;

(6) For the VAR to take on legal risk in an amount well beyond its ability as a practical matter to pay on the instalment dates;

(7) For the VAR to be dependent on the software seller to remit to it or credit it with the proceeds of the end-user sale, and, if no end-user sale eventuated, simply to await some other form of rescue by Autonomy;

(8) For a combination of the above to be typical of its largest VAR purchases from the software seller time and time again.

Further, the Defendants efforts to validate the impugned transactions require careful scrutiny. In many instances, those submissions were premised on fact patterns which at first blush appeared similar to what in fact occurred but which were on analysis subtly but materially different. In other instances, the factual matters relied on by the Defendants were not established, or they were revealed to be either incorrect or confined to special situations.

For example, it was submitted in Dr Lynch’s written closing that the Claimants were mistaken in repeatedly questioning what services or value a reseller was likely to bring to the transaction, and that “it would not matter if the reseller was not going to fulfil any role at all, other than purchasing the goods, going on risk for that purchase, and thereafter fulfilling the end-user’s purchase order.” They placed reliance on the FileTek ‘fulfilment only’ transaction in Q3 2010 (VT18) as an exemplar of a standard form of VAR relationship in the industry. In this regard, the Defendants relied especially on Mr Szukalski’s evidence in cross-examination to this effect, and his assertion that an agreement under which the VAR was required to do nothing at all after the VAR sale to it except ‘fulfil’ any end-user sale negotiated and contracted for by Autonomy was “very common practice” . But this carried no weight: Mr Szukalski was in no position to make such an assertion. FileTek carried on no business as a VAR ^{^[279]} , and (though he had worked for several software companies) Mr Szukalski had no evidenced experience of practices in reselling such as to give any credence to his view.

Further, even in the FileTek transaction in Q3 2010 (VT18), which the Defendants promoted as the VAR deal which took the Claimants furthest,^{^[280]} there was no evidence that FileTek was intended to fulfil the end-user’s purchase order except possibly in a nominal or mechanical sense and at Autonomy’s direction. Mr Szukalski confirmed that FileTek’s role was simply “holding the paperwork for those 45 days (i.e. pending closure of the USDVA end-user deal) and in exchange we’d earn a certain margin on the deal…So that’s what we were asked to do.”

Neither in the FileTek transaction nor in any other of the impugned VAR transactions was the VAR to play an active role or do anything on its own behalf at all. Even the residual task suggested of fulfilling the end-user’s purchase order was of no substance. If, in any instances, the VAR was left or required to ‘fulfil’ the purchase order (presumably by making available its access code) that would simply be at Autonomy’s direction in fulfilment of Autonomy’s contract and not its own, and all that was required to ‘fulfil’ the purchase order was delivery of the software; and that could be, and I would assume was, achieved electronically on Automater.

Similarly, the Defendants again sought to illustrate their case by promoting the FileTek transaction as showing that even in this transaction the VAR always undertook real risk. They conceded that email exchanges between Mr Szukalski and Mr Loomis made clear that Mr Szukalski was given to understand by Mr Egan that there would be no real risk for FileTek. But they sought to make a virtue of the concession by also contending that even in that context, Mr Egan had made clear that any indication of his did not bind Autonomy and that FileTek’s legal risk was unaffected. Further, they depicted what Mr Egan had said as soft-soap: all that Mr Egan had said was that (as Mr Egan put it in cross-examination) “If, for some reason, the [USDVA] deal was not completed, Autonomy would use all efforts to find another end-user for the same software or some other way to make sure that FileTek would get paid and could then pay Autonomy.” This was referred to as “backfilling” and the Defendants sought to present it merely as in the nature of an assurance of co-operation to substitute a new end-user. “Backfilling” was simply a euphemism for Autonomy trying to establish a direct deal with another party; the question is whether Autonomy would have pursued FileTek (or any other VAR in the impugned VAR transactions) for the purchase price it had agreed to pay, and the answer is, in my judgment, plainly ‘no’.

What shone through the evidence, in my judgment, was that FileTek was given assurances which satisfied Mr Szukalski and Mr Loomis that FileTek would never in fact be called for the purchase price. That was, as it were, the “bottom line”. Mr Szukalski admitted in cross-examination that but for those assurances, FileTek could not have proceeded: the amount of the deal was far above its ability to pay, and any real risk would have posed an existential threat.

Thus, in important respects the impugned VAR transactions did not satisfy even the Defendants’ formulation in paragraph 2155 above. On the contrary, the analysis that the argument necessarily occasioned demonstrated that the VARs did not expect to be nor were further involved with either the end-user or any onward sale, or indeed to do anything at all except co-operate with any arrangements that Autonomy might make to ensure that the VARs were not left “on the hook”, and receive payment of a MAF which in most cases was the only real reason that the VARs were prepared to accept the legal risk.

In my judgment, (a) the impugned VAR transactions manifest both aspects of the pattern which the Claimants have alleged and (b) the pattern provides further confirmation that in reality (i) the VAR sale did not impose any obligation on a VAR of any intended economic consequence (ii) the only real sale was where one eventuated pursuant to Autonomy’s negotiations with an end-user, and (iii) the only source from which Autonomy ever expected any actual payment (otherwise than out of its own funds) would be an end-user sale.

Relevance of Deloitte ’s approval of revenue recognition

In this context, as in all others, the Defendants relied on the approval given by Deloitte and the Audit Committee as showing that the transactions were within the margins of propriety and that, even if they were not, the Defendants were entitled to assume that they were.

The Defendants instanced paragraphs 82 to 84 of a letter from Taylor Wessing LLP on behalf of Deloitte to Travers Smith LLP, written in response to a letter of claim setting out the basis for a claim in negligence by Autonomy (now controlled by HP) against Deloitte, where it was stated as follows:

“ Deloitte did in fact carefully consider and review each of the situations in which a contract with a VAR was cancelled and the deal was completed directly with the end-user, in order to understand the reasons for the cancellation and entry into a direct contract with the end-user. This included ensuring that the sale had not been double counted. With the exception of the Capax / Kraft transaction, which had no impact on the 2009 year-end results, and which was the first time Deloitte had seen such a cancellation, these transactions were also all raised with the Audit Committee.

In each case, as at the time of considering the original transaction with the VAR, there was a binding and non-cancellable agreement in place between Autonomy and the respective VAR, and under that contract the risks and rewards had passed to the VAR.

[Underlining as in the original]

As foreshadowed above, however, what the Defendants’ submissions did not grapple with was what was missing from the information available to Deloitte, and what they were not told, even when they raised concerns. This included, most importantly:

(1) Deloitte were never told or aware that it was never intended by either Autonomy or the VAR that the VAR would have any role or say in the negotiations with the end-user;

(2) Deloitte were never told or aware that in the negotiations between Autonomy and the end-user after each impugned VAR sale Autonomy was at best indifferent as to whether any sale to the end-user should be by the VAR or by Autonomy. Deloitte were left with the impression that ‘direct ’ deals were exceptional rather than the planned norm, and were unaware, since Autonomy did not mention the VAR sale and appeared to be acting as principal, that the end-user would always be most likely to contract (if at all) with Autonomy as the person it had always dealt with and the supplier of the software;

(3) Thus, for example in the Kraft deal (see paragraphs 106 to 121 of the Schedule of Impugned VAR Transactions), Deloitte were never told or aware that the VAR had never been involved in any negotiations and a direct deal had always been envisaged; on the contrary, the purchase order which was the source of the information given to Deloitte stated in terms that (a) the end-user and VAR were anticipating entering into a license transaction with one another and that (b) any direct deal, though possible, was an “unlikely event” .

(4) Autonomy ’s response to Deloitte ’s expression of concern (in its Q1 and Q2 2010 reports to the Audit Committee) about a possible pattern of direct deals, and their warning that this might jeopardise Autonomy ’s ability to recognise revenue at the point of sale to the VAR, was to represent in Q3 2010 (as was in Deloitte ’s Q3 2010 report) that there had been no further such direct deals; whereas in fact Autonomy had entered into a direct deal that very quarter (Q3 2010) with the FSA (VT10) which it concealed from Deloitte until much later.

(5) Deloitte were repeatedly misled by Autonomy as to the ability of ‘friendly ’ VARs to meet out of their own resources the legal obligations they assumed, and they were unaware that such VARs were reliant upon the assurances given by Mr Egan that Autonomy would never require such payment and would engineer another means of funding the obligation.

(6) Finally, and in summary, Deloitte were never told, and certainly were not given any full or proper understanding, of the assurances given to the relevant VARs by Mr Egan, nor therefore of the shared understandings between the parties which rendered every impugned VAR sale artificial and/or merely a device to give the appearance of a transaction which would give rise to revenue whereas in reality it would never do so.

(6) The Defendants ’ knowledge of improper accounting

That brings me to the issue of the Defendants ’ knowledge: in other words, whether the Defendants knew that the VAR transactions lacked substance, that the only real sale would be if and when one eventuated between Autonomy and an end-user, and that the recognition of revenue at the amount and time of the VAR sale was improper.

Obviously, Mr Hussain was more directly and routinely involved than was Dr Lynch; and it is appropriate to deal with them each in turn.

Mr Hussain ’s knowledge

The Claimants summarised in six points their case that Mr Hussain was well aware of the true nature of the impugned VAR transactions and their improper accounting treatment and presentation.

First, Mr Egan ’s evidence was that it was Mr Hussain who was the architect and “defined the parameters” of the practice of using VARs for “at risk deals” or (as Mr Hussain called them) “acceleration deals” to accelerate revenue recognition ^{^[281]} ; and it was Mr Hussain who selected which deals, and in what amounts would be taken to a VAR, and which VAR to approach and who directed Mr Egan to give the assurances to the VARs that they would not be left “on the hook” or “holding the bag”.

Mr Egan added that Mr Hussain:

(1) gave him “specific instructions to follow so that these deals would be accepted by Autonomy’s auditors (Deloitte)” ;

(2) “provided guidance to [him] regarding what was, and was not, acceptable to communicate in my conversations with VARs” ,

(3) “laid out explicit rules about what could be offered as incentives to the VARs, what was required of the VARs, and what could not be part of any deal”,

(4) instructed him to say that it would be Autonomy which “would continue its efforts to sell to the end-user” and that “ the VAR was not expected to participate in these sales efforts”, and

(5) directed him to assure the VARs that, though nothing could be put in writing, “if Autonomy was ultimately unable to close a deal with an end-user, there were various options that Autonomy had to “fix” the situation for the VAR so that it would not end up having to pay for the software from its own resources”.

According to Mr Egan’s evidence, Mr Hussain was thus well aware of the assurances given. Mr Egan accepted in cross-examination that, as far as he was aware, Mr Hussain was not involved in negotiations or discussions with VARs prior to any impugned VAR sales. However, when it was suggested to Mr Egan (by reference to one of the Capax Discovery sales) that it was “just an arrangement between you and Mr Baiocco and it wasn’t something that Mr Hussain knew about” he was very clear and adamant in his insistence that he knew “first-hand that Mr Hussain knew about it”. Furthermore, he told me that Mr Hussain:

“had much more interaction with the VAR after a deal, particularly if a deal required some form of unconventional conclusion; like if a deal ultimately had to be cancelled and dealt with in a different way, in those circumstances Mr Hussain was often more involved.”

As indicated in the Introduction, Mr Egan ’ s witness statement appeared to me to be over-lawyered; and he was, as also noted previously (see paragraph 1977 above), restricted in what he could say without peril, having agreed with the DoJ not to offer any contradictory evidence or arguments on pain of immediate prosecution without defence for breach. But when he gave his oral evidence (over a video-link, which was adequate) he struck me as substantially reliable in what he said about the institution and development of the “ at risk ” VAR selling programme and Mr Hussain ’ s part in it.

His cross-examination elicited answers supportive of the defence based on the notion that the arrangements with the VARs did not affect their legal liability to pay, and the transactions fell to be accounted for in accordance with their contractual definition. He agreed repeatedly and with enthusiasm that (a) his understanding was that what happened was legal because the assurances given to and understandings established with the VARs were not legally enforceable, did not in any way affect the VAR ’ s legal obligation to pay and thus were not, as he understood it, “ side agreements” in a legal sense, and (b) he saw nothing wrong in paying the MAFs as compensation for the VAR ’ s agreement to take on the legal liability to pay. But his cross-examination did not, in my judgment, unsettle materially his account of the derivation, development and implementation of the post-2008 VAR “ at risk ” sales strategy.

Secondly, it was the unchallenged evidence of Mr Welham that Deloitte was never told of the existence of any side agreements. As an accountant, Mr Hussain well knew that information about any side agreements or understandings needed to be provided to Deloitte. Indeed, Mr Hussain signed management representation letters to Deloitte confirming the absence of any side agreements. Given Mr Hussain ’ s actual knowledge of Mr Egan ’ s assurances to the VARs, it follows that Mr Hussain deliberately misled Deloitte.

Thirdly, Mr Hussain knew, from his own close involvement in the transactions, that Autonomy did not intend the VAR to do anything to achieve an onward sale or to make payment from its own resources. Mr Hussain knew that Autonomy would continue to try to achieve a direct end-user deal, as though the VAR transaction had not happened.

Fourthly, Mr Hussain was intimately involved in the steps taken to unwind the VAR transactions in accordance with the fundamental understanding that the VAR would never be left “ on the hook” or “ holding the bag” , as was his deputy, Mr Chamberlain.

Fifthly, Mr Hussain was a qualified accountant. He can be taken to have understood that the effect of the oral agreements or understandings between Autonomy and the VARs and the parties’ intentions in respect of the same was that which has been explained by Mr Welham, Mr Holgate and Mr MacGregor. There was no evidence from Mr Hussain to suggest that he made an innocent mistake about this.

Sixthly, Mr Hussain ’ s knowledge of the improper accounting for the VAR transactions is reinforced by his involvement in misleading Deloitte about the relevant facts. An example emphasised especially by the Claimants was his conduct after Deloitte had warned Autonomy management and the Audit Committee in its Audit Committee reports for Q1 2010 and Q2 2010 that a pattern of VAR agreements being replaced by direct sales to end-users might jeopardise Autonomy ’ s ability to recognise revenue at the point of sale to the VAR.^{^[282]} Mr Hussain assured Deloitte, and Deloitte ’ s Audit Committee report for Q3 2010 noted that there had been no further reversals of VAR transactions in the form of direct deals between Autonomy and end-users, thus supporting Autonomy management ’ s policy of continuing to recognise revenue at the point of sale to the VAR. Yet Mr Hussain knew, but concealed from Deloitte and the Audit Committee, that in that very quarter (Q3 2010), another VAR transaction (namely VT10, Capax Discovery/FSA) had in fact been replaced by a direct agreement between Autonomy and the end-user. Mr Hussain plainly did this in order to avoid the consequences of which Deloitte had previously warned, namely that Autonomy ’ s ability to recognise revenue at the point of sale to the VAR would be imperiled. This same phenomenon recurred in later quarters, right up until Autonomy ’ s acquisition by HP, but Mr Hussain continued to conceal this from Deloitte.

Further, Mr Hussain ’ s knowledge of the falsity of the accounting treatment can be inferred from the numerous other respects and occasions in which Mr Hussain misled Deloitte about relevant facts, including as to the rationale for the reciprocal purchases that were made to put the VARs in funds to pay down their debt to Autonomy.

I accept these points, which are in my judgment, amply illustrated and supported by a review of the individual impugned VAR transactions to which I turn below. In my judgment, Mr Hussain had ‘guilty knowledge ’ of the artificiality of the impugned VAR transactions and their improper accounting treatment and presentation.

Dr Lynch ’s knowledge

Mr Egan said Dr Lynch was not involved in dealings with VARs. He told me that he was not suggesting that he could give evidence about what Dr Lynch thought or knew. There was not the transaction-specific or direct evidence against Dr Lynch that there was against Mr Hussain.

Nevertheless, the Claimants’ case is that the body of evidence as a whole establishes Dr Lynch ’s knowledge of the artificiality and improper accounting of the impugned VAR transactions.

First, they submitted that it is clear on the evidence that Dr Lynch had personal knowledge of many of the end-user negotiations and was well aware from the inception that, where no end-user deal had been executed by quarter-end, Autonomy used a VAR to recognise the revenue instead.

Thus, for example, in relation to one of the very first impugned VAR transactions, Capax Discovery/Kraft (VT3), Mr Hussain emailed Dr Lynch on 25 September 2009, just before quarter-end, saying “ I have an idea on kraft ”. Dr Lynch ’s omission to address this email, which was expressly pleaded, in his witness statement was characteristic of a notable evasiveness on all questions on the issue of his knowledge. His oral evidence in cross-examination was as follows:

“Q. Mr Hussain would have spoken to you to tell you what the idea was, yes?

A. Well, there would have been some communication I assume, yes.

Q. That idea, I suggest, which was the second line there, that idea, his idea on Kraft, was to use a value added reseller to take over the Kraft deal and so enable revenue to be recognised in Q3 2009, correct?

A. I suspect so, yes.”

He at first sought to deny any specific knowledge and asserted that he was “generally not involved”. He then accepted that he would usually come to know, but that he would only “find out what had happened once the dust had settled ”. The following passage from his cross-examination captures this:

“Q. So you would have known, I think you’re saying, at the latest very shortly after quarter end that there had been a deal with a reseller?

A. Yes, I would probably have known -- again I don’t know but I would probably have known shortly after quarter end. That’s the most likely.”

My impression was, however, that Dr Lynch was seeking, by this formulation by reference to a time by which he would have known of the transactions, to avoid accepting that he knew of their progress well before they were executed as VAR sales. I note also, for example, that on occasion, such as in Q1 2010, it was Dr Lynch who added the VAR ’s name to the revenue route-map for that quarter. I have concluded that Dr Lynch kept a careful eye on the transactions and was well aware of their progress, even if he may not always have recalled more than the headline amount involved. The principal factors that have led to my conclusion are summarised below.

Email evidence confirms that Dr Lynch had a keen interest in the typically large VAR deals which became necessary to fill the gaps, and I do not accept that he only came to know of such matters after the end of the quarter. A comparison between revenue achieved and revenue forecasted was supplied regularly by Mr Hussain to Dr Lynch in schedule form; and the need for recognised revenue would not have been lost on him. I do not accept his evidence that the situation was so frantic and “dynamic” that he had no time to keep track.

The size of the deals, and their importance in the achievement of revenue forecasts, also would have been likely to capture his attention. It is highly unlikely that Mr Hussain would not have kept Dr Lynch carefully and continuously informed. In this regard:

Many of the impugned VAR transactions were amongst the largest, by revenue, in a given quarter.

(1) As illustrated in the relevant slide, in each of Q4 2009, Q1 2010, Q3 2010, Q4 2010, Q1 2011 and Q2 2011, the disputed VAR transactions accounted, in aggregate, for over $20 million of reported revenue in that quarter, and they were key to Autonomy meeting its revenue and earnings per share targets.

(2) The pattern as regards the principal ‘friendly’ VARs in the Relevant Period is also striking and would surely have measured with Dr Lynch:

i. Capax Discovery entered into 10 impugned VAR transactions generating licence revenue of over $48 million. It did not enter into any other VAR transactions with Autonomy above $1 million.

ii. MicroTech entered into 9 impugned VAR transactions generating licence revenue of over $43 million. MicroTech entered into only one VAR transaction with Autonomy above $1 million (in an amount of $1.1 million), which is not impugned.

iii. DiscoverTech entered into 8 impugned VAR transactions generating licence revenue of over $39.5 million. It did not enter into any other VAR transactions with Autonomy above $1 million.

iv. FileTek ’ s sole VAR transaction was for $10 million.

(3) The impugned VAR transactions accounted for 88% by value of Autonomy ’ s VAR transactions above $1 million in the Relevant Period generally.

Secondly, the Claimants submitted that though he professed to be unaware of the detail, Dr Lynch knew that it was not intended that the VAR would participate in the end-user negotiations.

Again, Dr Lynch’s evidence in cross-examination on this aspect struck me as contrived. Ultimately, he appeared to be taking a legalistic position that because the VAR had purchased the goods, and thereby acquired control of them, Autonomy could do nothing to stop it being involved in negotiations with an end-user for onward sale, and that it followed that anything done by Autonomy was at the behest of and in conjunction with the VAR and not unilateral or exclusive of the VAR. But all this was theory and not reality, as I think Dr Lynch well knew.

The reality, I have concluded, is that Dr Lynch would never have expected or tolerated a VAR taking over control of negotiations commenced by Autonomy with one of Autonomy’s customers, and would have operated on the (correct) understanding that none of the VARs had the slightest intention or inclination to do so (being content to receive a large MAF for saying and doing nothing).

I agree with the Claimants’ contention that it would have been quite contrary to Dr Lynch’s controlling nature to suppose that he expected a VAR to be involved in negotiations, which hitherto had been conducted by Autonomy alone and which in truth Autonomy expected to result in a direct deal (or no deal). That is supported by, for example, an internal email from Dr Lynch dated 7 August 2010 in the context of the USDVA end-user deal, which illustrates his own conception of the negotiations being entirely a matter for Autonomy and of vital interest to it. He reminded all concerned that this was “the biggest deal in the quarter” , that “ALL of you own this” and the responsibility was theirs to bring the end-user transaction to a successful conclusion. He required to be told about any problem in securing an end-user deal “ IN A F***ING MILLISECOND ” . I agree with the Claimants’ point that the idea that Dr Lynch expected FileTek - a small company which had never previously acted as a VAR (whether for Autonomy software or otherwise), never spoken to the USDVA before, and knew nothing about the negotiations hitherto - to take over or participate in those negotiations is simply not credible.^{^[283]}

Thirdly, the Claimants submitted that once it is concluded that Mr Hussain knew of, and indeed authorised, the giving of assurances to the VARs by Mr Egan, the natural inference is that Mr Hussain shared what he knew with Dr Lynch. Dr Lynch did not contend that Mr Hussain concealed material facts from him. I agree, and the inference is further supported by a more general consideration of the way Autonomy was controlled and directed, as I have described in the introduction.

In my judgment, it is likely that Dr Lynch would have worked out for himself, if he was not the architect of the strategy, the basis on which the VARs were contracting with Autonomy, taking on apparently unconditional liabilities for millions of dollars on or around the last day of the quarter, with no insight into the end-user’s requirements, and with no intention of themselves seeking any onward sale to the end-user. The balance of probabilities is, in any event, firmly in favour of Mr Hussain having told Dr Lynch of the reality that the VARs had proceeded only on the basis that they would never be left “holding the bag” .

I consider and find that Dr Lynch was aware that it was not intended or expected that the VAR would ever be called upon to pay out of its own resources, and that instead, the understanding was that if and when no end-user sale eventuated, Autonomy would either arrange some form of funding or credit note, or failing that, simply write off the apparent debt. That is what happened: and I would expect Dr Lynch to have been aware of that, either as part of his general supervision or because it would have been reported to him by Mr Hussain or another in the cabal.

I accept in that regard the Claimants ’ submission that it is improbable (the Claimants said “inconceivable” ) that Autonomy released its legal entitlement to be paid millions of dollars by the VAR without first obtaining Dr Lynch ’s authorisation, or, at the least, being entirely confident of his approval, if not on a transaction-specific basis, then at least on the basis of Dr Lynch ’s approval of a general policy. In this context, the Claimants pointed to the fact that neither of the individuals whom Dr Lynch supposed were involved in the decision, namely Mr Hussain or Mr Kanter, had been called to give evidence, and nor was Mr Chamberlain (who became increasingly closely involved); they invited the court to infer that they could not have given truthful evidence to support Dr Lynch ’s case.

The Claimants als o submitted that Dr Lynch was personally involved in many of the steps taken to ensure that no VAR would be left “holding the bag”, and, for example, that he:

(1) Personally gave approval for Autonomy to make purchases from the VARs in order to put them in funds to pay down their debt to Autonomy, such as the purchase of the ATIC licence and the Federal Cloud Platform from MicroTech, the DiscoverEngine purchases from DiscoverTech and the further StorHouse purchases from FileTek; and

(2) Knew of and authorised the further steps taken by Mr Hussain in Q3 2011 to make the VARs whole prior to HP’s takeover of Autonomy by means of very considerable write-offs and credit notes to forgive or set off their indebtedness.

I deal in more detail with Dr Lynch’s personal involvement in the impugned VAR and reciprocal transactions which I address in the next main section of this judgment.

As to (2), in cross-examination, Dr Lynch sought to distance himself from these steps by asserting that they were a frolic on the part of Mr Joel Scott ^{^[284]} , without authorisation from the UK team. He was adamant when cross-examined that he:

“…was not involved at all in any programme to make VARs whole, and [he] was not involved in any of the write-down or reversal decisions…[and] in fact…didn’t know about them at the time.”

Dr Lynch contended that Mr Scott’s contrary evidence in the US criminal proceedings, which was in evidence before me under a hearsay notice, should be given no weight. He made the point that Mr Scott had been granted immunity by the US authorities (for which he implied suitably supportive evidence for the prosecution was the price) and had not provided a witness statement or attended to be cross-examined in these proceedings. Further, he suggested that contemporaneous emails did not support Mr Scott’s position that he had been directed to do what he had done by Mr Hussain, directly or through Mr Chamberlain.

I have not felt able to accept Dr Lynch’ argument. The documents make clear that the exercise of writing off debts was overseen from England by Mr Chamberlain, Mr Hussain’s deputy, and a member of the small group or cabal of managers with and through whom Dr Lynch managed and directed the Autonomy group (see paragraphs 1093 to 1094 above). Ms Harris agreed that she would have expected Mr Chamberlain to have discussed write-offs with Mr Hussain. I consider it much more likely than not that Mr Hussain was involved; and I agree with the Claimants that Dr Lynch would have been kept informed, and such large write-offs probably needed his approval.

Fourthly, as to Dr Lynch’s knowledge that the accounting treatment of the VAR transactions was incorrect and improper, the Claimants submitted that Dr Lynch had approved payment of MAFs on the terms of MAF letters which sought to justify the payments by reference to the purported performance by the VAR of services which Dr Lynch must have known the VAR was never intended to or never did provide. The Claimants cited as typical the MAF letter (dated 30 September 2010) for the Capax Discovery/Eli Lilly transaction, which included the following:

“To formalize our prior discussion, Autonomy Systems Limited (“Autonomy”) and Capax LLC (“Referral Partner”) agree to the terms and conditions of this letter agreement (“Agreement”) as follows:

Referral Partner will: (1) introduce Autonomy into the deals with Eli Lily (“End-User”); (2) obtain quotes from Autonomy on behalf of the End-User; and (3) work with the End-User to assist in executing purchase orders and contracts with Autonomy.

Autonomy will: (1) pay Referral Partner commissions in the amount of US$629,000, as a result of Referral Partner’s direct and proximate participation in the account; (2) deliver products directly to End-User; and (3) use reasonable efforts to provide mutually agreed upon sales assistance. …”

The Claimants invited me to read this and other similar MAF letters as plainly designed to give the impression that it was formalising prior discussions as to what role the VAR would perform, and that, in return for the performance of that role, Autonomy would pay the commission. They described this as simply a false paper trail.

Dr Lynch told me in cross-examination that he did not know whether he approved this letter; but to address the obvious difficulty that it was a characteristic of Autonomy’s true relationship with the VARs in the impugned VAR transactions that they certainly were not intended to and never did have any “direct and proximate participation in the account” , it was contended in his written closing that this was “standard form language and was capable of being used” and also that the “agreement also had forward-looking obligations including a restriction on competition from the reseller.” His closing submissions dismissed the Claimants’ point as “a forensic rather than real one.”

The Claimants also relied on the fact that the MAF letter agreements were drafted and signed by Mr Kanter, a member of Dr Lynch ’ s inner circle. During the oral opening, the Claimants signaled that this would be a matter for Mr Kanter ’ s cross-examination: “ We will have to ask Mr Kanter when he comes to give evidence on whose instructions he does this. ” In the event, Mr Kanter did not give evidence. The Claimants were thereby deprived of the opportunity to cross-examine Mr Kanter. The Claimants submitted that it should be inferred that Mr Kanter was not called because (i) he could give no honest explanation for the misleading way in which the MAF agreements were drafted and (ii) he would have implicated Dr Lynch.

Plainly it would have been interesting to hear any explanation from Mr Kanter. However, it seems to me likely that he would simply have reiterated Dr Lynch’s explanations of the MAF letters and their content. As explained in paragraph 2147 above, I prefer to determine the issue without relying on the inferences contended for by the Claimants and simply on the basis of the balance of probabilities revealed by the evidence available to me.

I consider it plain that the MAF letters did not reflect the reality: the reality was that the VARs were not expected to do anything, and the MAFs were the reward for their co-operation and assumption of a risk which, though in my view not intended to be enforced, was nevertheless legally enforceable.

I accept and find that the MAF letters were contrived in order to give the paper impression, especially to Deloitte, that the payments were not for the VARs’ co-operation in accepting legal risk and in signing confirmation letters that this legal risk was not subject to any side agreement, but for their contribution to the onward sales.

The Claimants’ sixth point was in two parts: (a) that Dr Lynch worked closely with Mr Hussain, a qualified accountant, and must have understood from him that Autonomy’s accounting was wrong, which is why Dr Lynch “supported the steps taken by Autonomy’s management to mislead Deloitte and the Audit Committee about the true facts”; and (b) that Dr Lynch himself had a sufficient grasp of accounting and the accounting standards to know that recognising revenue on the impugned VAR transactions was wrong. Both parts relate to and invite a determination of the question whether in the case of the impugned VAR transactions Mr Hussain and Dr Lynch did know that, on the true facts, revenue recognition was improper.

I address that question by reference to the individual impugned VAR transactions in the Schedule of Impugned VAR Transactions that accompanies this judgment. However, it is convenient to mention now four points advanced by the Claimants in support of their case as to Dr Lynch’s own understanding of the accounting standards which are not fact-specific and are of general application, and which raise an important question as to what really was Dr Lynch’s subjective understanding by which of course he must be judged:

(1) The first is that Dr Lynch accepted in cross-examination that he knew that the correct question to ask was whether Autonomy had transferred to the buyer the significant risks and rewards of ownership of the goods.

(2) The second is that Dr Lynch accepted that, if the reseller/VAR only had to pay for the product if the end-user bought the product, the transaction did not qualify for revenue recognition, and that “the way [he] would interpret it” the reason was that in that scenario the reseller had not taken on the unconditional risk of having to pay for the goods it had bought.

(3) Although Dr Lynch asserted in his oral evidence that he understood that any assurance had to be legally enforceable in order to affect revenue recognition, that assertion should be rejected. Dr Lynch accepted that Deloitte never told him that everything turned on what the written contract said. Dr Lynch also accepted that he was familiar with the concept of substance over form.

(4) Dr Lynch accepted in cross-examination that he knew that there was “some sort of managerial control test to do with having control of the goods” and “that the customer had to have control of the goods”.

In my view, it is fair to add that in giving those answers Dr Lynch had made clear that he was not an expert and could only speak to his “limited understanding”. It is also fair to add that in his answers Dr Lynch never stated expressly (as the Claimants stated he had) that his understanding was that “any assurance had to be legally enforceable in order to affect revenue recognition”. It was the Claimants who (both in their questioning and in their written closing submissions) repeatedly added the word “legally” before the word “enforceable” in restating Dr Lynch’s answers. Dr Lynch in his answers seemed to me to shy away from expressly asserting that only legal enforceability sufficed. He preferred to instance examples of what would not suffice to constitute an impermissible side agreement or assurance. The following answer exemplifies how he sought to explain his approach:

“A . …if someone gives someone general industry patter about, you know, “ We look after our partners”, but that is not enforceable, then I don’t think that’s a problem.

Q. What if it goes beyond that and it’s an assurance that that will not happen, in other words you will not be left –

A. That’s my point. If it’s enforceable then I would be in agreement with you, if it’s not enforceable then I’m not.”

It is of fundamental importance to distinguish, as sometimes to my mind the Claimants failed to distinguish, between on the one hand, Dr Lynch’s own genuine but subjective understanding of the applicable accounting principles and standards and, on the other hand, an objective determination of their meaning (though the latter is also necessary). It is therefore necessary to establish what he meant on this important point.

In my judgment, Dr Lynch was aware that something less than a legally enforceable agreement would ordinarily suffice. His point was that he thought that the provisions of every one of the VAR agreements expressly stating that no side agreement would be legally enforceable did effectively negate any purported side agreement or assurance. The following exchange later in his cross-examination, focusing on the difference between substance and form, seems to me to illustrate this:

“Q. Are you familiar with the concept of substance over form?

A. Yes.

Q. And so if the substance of the transaction reflects the fact that there is an arrangement, regardless of what form is, regardless of what the legal term says, if the substance is there is an arrangement under which the VAR is not at risk, you would have known that revenue couldn't be recognised?

…

A. If obviously the whole thing was a sham, then I agree, but if it's a real situation, surely one has to look to the contract to see what it actually is.

Q. I think you’re accepting you don’t just look at the contract, you look at substance over form. The contract may suggest the form, but if in commercial substance there is a wider arrangement, then that’s what you look at, correct?

A. Again, I feel like we - you know, I ’m in a room full of lawyers, you're asking me what legal terms are, and then there's a roomful of accountants. I agree with you that if you had a contract but the reality was the whole thing was cooked up and was a complete sham, then obviously. But in any practical situation, when there's a real contract and it's clear, then surely that is the commercial reality.

Q. I'm asking you for your understanding at the time?

A. I think that’s my understanding at the time that I'm giving you.

Q. You say if you had a contract but the reality was the whole thing was cooked up and it was a complete sham. That's not what I'm suggesting. I'm suggesting you had a contract and a side arrangement attached to that contract which makes it clear that the VA R is not on risk, so that the substance –

A. Yes, but then you have to look at the legal agreement and if it says that whatever you're calling a side agreement is worthless and therefore not enforceable, then the legal term is showing that you have a real deal.”

That was the nearest that Dr Lynch came to asserting that only a legally enforceable side agreement would alter the substance of the transaction in accounting terms. It is of obvious importance to the VAR case more generally to determine whether that was truly his understanding and belief (as is his case) or whether in truth he appreciated that the substance of the arrangements had to be viewed as a whole, had to be determined according to the real intentions and expectations of the parties, and that determination would, in turn, be determinative of its proper (and mandatory) accounting treatment. That is to my mind, indeed, the single most important question to be answered on this part of the case.

It seems to me that Dr Lynch persuaded himself that the contractual wording gave sufficient protection for what he and Mr Hussain had decided needed to be done in order to preserve Autonomy’s reputation as a company which set and met growing revenue forecasts, but did appreciate that words on the page of a contract could not change reality. He knew that a transaction had to be accounted for on the basis of its substance rather than its form, and that substance had to be measured according to the true intentions and expectations of the parties. If, as I find, he knew that the VARs were proceeding on the basis that whilst the contract would be enforceable it would never in reality be enforced, he must have appreciated, as a highly intelligent and experienced CEO, that the substance of the transaction was different from its legal form.

In short, in my judgment, Dr Lynch knew that the contractual wording did not define the contracting parties’ relationship but rather obscured the substantive reality that the VARs were placeholders taking on a commercially illusory legal liability in expectation (though there was no contractual provision for one)^{^[285]} of a substantial fee.

The seventh and last of the points relied on as establishing the fact and extent of Dr Lynch’s knowledge was again in two parts: (a) that Mr Hogenson had specifically raised concerns with Dr Lynch about Autonomy’s accounting for VAR transactions, which Dr Lynch had purported to have fully investigated but in fact had succeeded in smothering; and (b) Dr Lynch contributed to and approved Autonomy’s letters to the FRRP, following and prompted by the Hogenson episode, which were highly misleading, as he must have known.

In relation to (a) in the preceding paragraph 2220 it is the Claimants’ case [addressed further in paragraphs 2232 to 2289 below] that rather than investigate dispassionately the concerns expressed by Mr Hogenson, Dr Lynch’s reaction was to push them into the long grass, and then find a reason to eject him from the company at the earliest possible opportunity. The Claimants contended that this is not how an honest CEO would have reacted, and that his conduct betrayed his appreciation that Mr Hogenson’s concerns, if not smothered early, would reveal the truth. I address Mr Hogenson’s concerns and how they were dealt with in greater detail later.

It is sufficient for present purposes to note that those concerns were, for the most part, well-founded; that the way they were dealt with was contrived to ensure their dismissal without proper investigation; that Dr Lynch was closely involved in that process; and that the “Hogenson episode” further demonstrated that the accounting treatment of the VAR deals might not survive scrutiny of what Mr Hogenson set out.

Similarly, the Claimants contended that Dr Lynch approved Autonomy’s response (dated 3 March 2011) to the FRRP’s letter dated 2 February 2011 seeking information about a number of the VAR transactions, particularly Capax Discovery/Kraft (VT3) and Capax Discovery / Eli Lilly (VT4), and that the response was materially misleading in a number of respects. They submitted that Dr Lynch can only have sought to mislead the FRRP if he appreciated that stating the truth would have caused them concern and/or undermined the presentation and accounting treatment of the transactions under review.

As to (b) in paragraph 2220 above, Dr Lynch accepted in cross-examination that he appreciated that Autonomy’s response to the FRRP needed to be accurate and complete. It was apparent from Mr Kanter’s email of 2 March 2011 attaching a draft that he had looked at Autonomy’s draft response and was content with it in final form, as he ultimately accepted in cross-examination. It is, as I find later in this judgment, equally clear that the responses given to the FRRP were fundamentally misleading.

I address in more detail the FRRP’s requests and Autonomy’s answers at paragraphs 2290 to 2336 below. It is sufficient for present purposes to note that a central theme of Autonomy’s responses was that it was unusual (indeed “almost unique” ) for Autonomy, having sold a licence to a VAR for ostensible on-sale to an end-user, to enter into a direct agreement with the end-user. As the Claimants contended that was highly misleading, as Dr Lynch (as well as Mr Hussain who was also involved) must have known.

As at 3 March 2011, this had happened 13 times. Far from being “almost unique” it was an invariable characteristic of every one of the impugned VAR transactions which resulted in any end-user deal at all. Deloitte, believing there had been only isolated incidents, and in ignorance of the extent of the practice or pattern, had warned in their Q2 2010 Report to the Audit Committee that “significant evidence of such further revenue reversals may jeopardise management’s ability to recognise revenue at the point of sale to the reseller.” The Defendants knew full well that the practice was incompatible with revenue recognition.

For the reasons elaborated later, I accept the Claimants ’ case that if Dr Lynch (and indeed Mr Hussain) had honestly believed that Autonomy ’ s accounting treatment of the impugned VAR transactions was proper, they would have had no reason to withhold the true facts from the FRRP, let alone actively mislead the FRRP through false statements and implications. That the Defendants were prepared to draft and approve letters to the FRRP (part of the FRC, and in effect its Regulator) in these terms reinforces the conclusion that they appreciated that Autonomy ’ s accounting treatment of the impugned VAR transactions was improper.

Summary of my conclusions on the VAR case

It may be helpful to summarise now the overall conclusions I have reached on the Claimants ’ VAR case. I confirm that these conclusions, though stated now to assist digestion of my analysis of the individual VAR transactions that follows in the Schedule of Impugned VAR Transactions that accompanies this judgment, have been reached after my assessment of each of the 37 impugned VAR transactions.

In summary, in my judgment:

(1) The impugned VAR transactions follow a pattern, of which the characteristics were that (a) they were all undertaken right at the end of a quarter (b) they were all directed and/or encouraged by Mr Hussain with the objective of making good shortfalls (compared with quarterly forecasts) on the recognised revenue from other activities; (c) their value of course varied, but all were of considerable size; (d) almost all were negotiated by Mr Egan; (e) in every case, an enforceable VAR sale contract was made or treated as made before the end of the quarter so that the revenue from it could be recognised in the accounts for that quarter; (f) in every case, the reseller confirmed that there was no side agreement or understanding such as to qualify their contract; and (g) in every case the relevant software was delivered electronically on Automater; but equally, (h) in every case and however transmitted between the reseller and Mr Egan (or whoever was arranging the VAR sale on behalf of Autonomy) the shared intention and expectation of the contracting parties was that the contractual terms, though they had to be seen to be agreed and confirmed, would not ever be enforced, so that the reseller would not seek to use, or negotiate with the prospective end-user to sell, the software (unless and until so directed by Autonomy) nor would Autonomy require payment of outstanding payment obligations until the VAR could pay either out of the proceeds of an end-user sale negotiated by Autonomy or funds from a transaction arranged to generate funds for the reseller to use to pay it and that instead and in the meantime, (i) the VAR’s role would be to accept a nominal legal risk (sometimes called “holding the papers”) but otherwise do nothing further (save if and as required by Autonomy), in return for which (j) the reseller would be paid a MAF, purportedly for its services as a reseller but in fact as a quid pro quo for its passive role. No reseller was to be left “on the hook” or “holding the bag” (as the understanding and intention that all should be insulated against their legal liability was variously described).

(2) Deloitte did not see the pattern; alternatively Deloitte preferred to accept reassurances that ostensibly negated its true purpose. In any event, neither the approval of Deloitte nor that of the Audit Committee was fully and properly informed, and the fact of it does not avail the Defendants, who knew that.

(3) The strategy which the pattern was designed to implement was to ensure that Autonomy continued to appear to be a company which met its forecasts out of the sales of IDOL and related software, and thus maintain Autonomy’s market following and share price even in difficult trading conditions after the financial crisis in late 2008 and early 2009. The VAR strategy became of additional importance (and increased in volume) after the supplier of hardware originally relied on by Autonomy, EMC drew back from its association with Autonomy. VAR sales, like hardware sales, were increased or decreased according to the revenue shortfall at the end of each quarter. There is a correlation between the throttling back of revenue from hardware sales and the acceleration of revenue from VAR sales in and after Q3 2009 (and most markedly in Q4 2009 and Q1 2010) which speaks to the objectives of both.

(4) I see no reason to doubt Mr Egan’s evidence that the strategy was first conceived by Mr Hussain; and it also seems clear from that evidence and other documentary evidence that it was directed by Mr Hussain and encouraged and presided over by Dr Lynch, both of whom knew that the transactions were not being accounted for according to their true substance, and both of whom knew that the recognition of revenue on the sale to the VAR was improper, and that the accounts were thus false.

(5) Both Defendants thus had “guilty knowledge” of untrue or misleading statements in Autonomy’s published information for the purpose of FSMA and are liable accordingly. The Claimants’ VAR case succeeds.

That conclusion, and the strategy informing and the pattern revealing the impropriety of the impugned VAR transactions, is supported and illustrated in the Schedule of Impugned VAR Transactions that accompanies this judgment in which I address each of the 37 impugned VAR transactions in turn. (I have incorporated these in a Schedule to try to make this judgment a little more manageable. But I should stress that the findings there should be taken to have the same intended status as if in the main judgment itself.)

The variations illustrate the adaptability and incremental development of the pattern. The growing number of impugned VAR transactions which were never followed by any end-user sale at all, and where Autonomy had, in effect, to make good the money it had paid to the VAR whilst leaving the VAR with the benefit of the MAF, provided the most lurid illustrations of the use of the strategy to generate recognised revenue at very considerable ultimate cost.

The ‘ Hogenson episode ’

I have mentioned the ‘Hogenson episode’ briefly in various places above. It follows incidentally from my conclusion on the VAR case and the individual impugned transactions (including all those which caused Mr Hogenson concern and prompted his inquiries) that Mr Hogenson’s questions and concerns about the way Autonomy accounted for VAR transactions were justified.

Before addressing each of the impugned VAR transactions to illustrate further the pattern identified by the Claimants and which I have accepted, I turn to consider in more detail the ‘Hogenson episode’. Partly this is in deference to the detailed submissions made by the parties. Partly it is because an understanding of the way Mr Hogenson’s concerns were dealt with assists an understanding of the questions raised by the regulators after the episode. And partly it is because the way Dr Lynch dealt with the concerns (which, as will appear, included the sacking of Mr Hogenson) seems to me to fortify my conclusions on the issue of Dr Lynch’s knowledge.

Although the Hogenson episode gave rise to no claim, it was treated by the Claimants as the most telling demonstration of the fundamental dishonesty of the Defendants and their knowing participation in the improper transactions and accounting which had led to the Claimants’ losses. They concluded the separate chapter they devoted to the episode in their written closing submissions as follows:

“The Court should find that the explanation for the Defendants’ conduct is that they knew that Mr Hogenson was on the scent of fraud and were determined to get rid of him, come what may.”

An earlier passage identifies who was involved:

“Dr Lynch’s reaction, in concert with Mr Hussain, Mr Kanter and Mr Chamberlain, was to take all steps necessary to undermine, discredit, retaliate against and ultimately eject Mr Hogenson from the company. This is compelling evidence that Dr Lynch felt threatened by Hogenson’s allegations, because he knew them to be well-founded.”

The Claimants also and more generally relied on the spotlight shed on them by the Hogenson episode as having provoked Autonomy suddenly to shift away from VAR transactions and the impugned ‘reciprocal transactions’ and to resort to other different (but so the Claimants contended, equally objectionable and surreptitious) means of enabling it to appear to generate recognised revenue and cover shortfalls in actual software sales.

In particular, the Claimants pointed to a sudden decline in VAR transactions after Mr Hogenson had raised his concerns, accompanied by a sudden increase in “pure hardware” sales (from $11.8 million in Q1 2010 to US $31.1 million in Q2 2010), together also with an acceleration of hosting revenues through the sale of licences (which rose to $32.5m in Q2 2010).

Mr Hogenson had been well regarded in Autonomy. He had been Vice-President, Finance at Interwoven Inc before its acquisition by Autonomy in 2009 and had made rapid progress in Autonomy. Although Dr Lynch said his title of ‘CFO in the Americas’ was overblown and that he was really an operational manager, he was described by Mr Anthony Bettencourt (as he was then) (in June 2009) as “spot on in the details, digs into the revenue and takes this all very seriously” and by Mr Hussain as “Definitely worth keeping, trustworthy and revenue driven” .

In Q2 2010, Mr Hogenson emailed Dr Lynch (dated 23 June 2010 but in fact sent on 22 June according to West Coast/Pacific time) stating that he had become concerned that Autonomy might have engaged in a number of VAR transactions, and at least one of what the Claimants label reciprocal transactions, which had been wrongly accounted for, leading to a material misstatement of revenue in Autonomy’s published information.

The introduction to the email carefully set the context: explaining that the concerns it expressed followed on from a review of financial accounts and account reconciliations across all business units in the Americas which Mr Hogenson had instructed the newly consolidated Autonomy Americas finance team to undertake in Q2 2010.

Mr Hogenson expressed these concerns as follows:

“During this review I became concerned that Autonomy may have engaged in multiple material transactions with resellers of our software where the available information suggests that we may have materially misstated revenue and income within our reported financial statements in 2008, 2009 and Q1 2010. The evidence that I have gathered suggests that members of your senior management team and other employees may have been engaging in seven and eight figure transactions with resellers where 1) there was no end-user and the reseller does not have the ability to make the payments under the agreement; 2) there appear to be resellers with related party relationships on revenue transactions; and 3) there appear to be barter transactions where the economic benefit on both sides of the transaction appears to be materially overstated.”

[Underlining, for emphasis, supplied by me]

Mr Hogenson went on to say that he considered this to be “a qualifying disclosure made in good faith based upon the evidence currently available to me” (which was no doubt important to him because that categorisation of the disclosures insulated him under California law from adverse employment repercussions). He requested:

“a meeting with the audit committee prior to the release of Autonomy’s Q2 2010 financial statements to review these matters and ensure that material misstatements, if any, are corrected prior to releasing future financial statements” .

That initial email gave no particulars or examples: it mentioned no specific deals, counterparties, dates or transactions. However, Dr Lynch might be expected to have appreciated the seriousness of a concern about potential wrongdoing by senior management, especially when expressed by a senior employee with finance management responsibilities. It was also clear that Mr Hogenson wanted a meeting with the Audit Committee.

Dr Lynch accepted in cross-examination that he understood that Mr Hogenson was expressing concern about potential wrongdoing by management and (from the request to meet with the Audit Committee) that Mr Hogenson wanted to escalate those matters above Dr Lynch’s head. He accepted also that this was “a very serious allegation.”

Dr Lynch’s first answer to the concerns raised, however, was that, far from being a genuine whistleblower, raising concerns in good faith to ensure that Autonomy’s financial statements did not include inaccuracies capable of misleading the market, Mr Hogenson’s letter “out of the blue” was prompted by another objective: to deflect attention away from serious process problems with Autonomy’s payroll systems, which had resulted in a substantial fraud, in his own department, which he feared could put his future at Autonomy in jeopardy.

The Defendants based this suggestion principally on the fact, revealed by email exchanges shortly before Mr Hogenson’s email to Dr Lynch late on 22 June 2010, that an issue had arisen, which should have been caught during the reconciliation process which Mr Hogenson had directed upon becoming CFO in the Americas, about certain discrepancies in relation to Autonomy’s US payroll systems (which it was part of Mr Hogenson’s job to oversee). The Defendants contended that what Mr Vaidyanathan described in an email to Mr Hogenson and Mr Chamberlain dated 22 June 2010 as “more payroll stuff crawling out of the woodwork!” (which Mr Vaidyanathan reported he was going to investigate in detail in San Francisco the next day) was the catalyst for Mr Hogenson’s sudden late email to Dr Lynch.

I do not accept this. By 22 June 2010, the problem was identified and needed review; but contrary to the Defendants’ case no serious process problem had yet been revealed. Furthermore, Mr Hogenson’s concerns as expressed in his 22 June email to Dr Lynch were of longer standing, and I was not persuaded that they were not honestly held. I am satisfied that when he raised them with Dr Lynch, he did so in good faith and not for any collateral purpose. I am satisfied also that, whether or not there had been a payroll fraud in Mr Hogenson’s department (in which it was not suggested Mr Hogenson was at all involved), that was not a basis for not treating his concerns properly.

I am satisfied, however, that Dr Lynch instead determined to do his best to push Mr Hogenson and his concerns into “the long grass”; and instead of seeking to facilitate access to Autonomy’s Audit Committee he sought to put obstacles in the way.

A considerable amount of evidence was put forward to show a very detailed sequence of events, and factual intricacies on which the Claimants made repeated submissions that Dr Lynch was lying about the substance of the matter and lying about matters of fact in the process he adopted after receipt of the letter. There were curiosities and instances when I cannot accept Dr Lynch’s account.

For example, Dr Lynch said that after receipt of Mr Hogenson’s initial email, he contacted Mr Kanter and subsequently tried to get in contact with Mr McMonigall, then the senior member of the Audit Committee, before finally having a phone call with Mr Hogenson himself. Dr Lynch said that, during this conversation with Mr Hogenson, “ I asked him which members of management he was referring to and he told me Mr Hussain and Mr Egan ”. He had not mentioned this in his witness statements and I doubt that this is accurate. In fact, the documentary evidence suggests that Dr Lynch consistently warned Mr Hogenson against identifying specific wrongdoers. It would be inconsistent with Dr Lynch’s own email dated 24 June 2010 to Mr John McMonigall, in which Dr Lynch had reported that Mr Hogenson “ refuses to name ” the wrongdoers. Dr Lynch’s subsequent suggestion, to seek to explain the email, to the effect that he had relayed the two names to Mr McMonigall in advance of sending this email, and that he had refrained from stating the names in the email itself for fear of leaks, was not persuasive. It might, I suppose, have been given credence if Mr McMonigall had confirmed it; but Mr McMonigall could not attend for undisclosed medical reasons; and no witness statement was provided under hearsay notice.

However, since it is not necessary or possible to delve deeper, I shall assume that either there were factual details which resolved the explanation or that this was a failure of recollection. Even in a judgment of this length, it would be disproportionate to rehearse all the details of this and other instances put forward by the Claimants; and it is not necessary, since even on the assumption that the inconsistency was apparent but not real for whatever reason, the conclusion I have reached on the substantive issue is that the Hogenson episode as a whole further demonstrates that the Defendants did know that the VAR transactions impugned in these proceedings were concocted means of revenue acceleration. I can summarise the salient events and features as follows.

It is apparent that Dr Lynch forwarded Mr Hogenson’s email of 22/23 June 2010 to Mr Kanter soon after its receipt, without any explanatory message (which may suggest an earlier telephonic explanation), but attaching a draft response, replete with spelling mistakes characteristic of Dr Lynch.

It seems that Dr Lynch also tried to get hold of the senior member of the Audit Committee, Mr McMonigall, initially without success because Mr McMonigall was offshore sailing. Dr Lynch could not recall exactly when he had first managed to speak to Mr McMonigall; but it seems that he had done so by 24 June 2010 (and see paragraph [2269] below). Dr Lynch’s evidence was that he also spoke to Mr Knights at Deloitte; and an email dated 28 June 2010 ^{^[286]} from Mr Knights to Mr Hogenson, Mr Knight and Mr Mercer, promising to respond to Mr Hogenson’s concerns “directly to the Audit committee…when we have finalized our review”, supported that.

At 5pm on 23 June 2010, and after some small amendments to the draft he had attached in his email of that morning to Mr Kanter (presumably reflecting their conversations after that), Dr Lynch sent an email, with the finalised draft letter attached, to Mr Hogenson, copying in Mr Kanter.

The letter expressed surprise about the concerns Mr Hogenson had generally identified, and asked that “as a first step” more information, identifying particular transactions of concern, should be provided. Whilst confirming that “it is clear that in any case you are acting in good faith and I appreciate your efforts” (thus affording Mr Hogenson the “whistleblower” protection under California law that he had wanted), Dr Lynch also raised an issue as to whether the concerns might be based on a misunderstanding of differences between US GAAP and IFRS, or partial information, stating:

“In order to provide an accurate statement of the questions as any part of any review, it is important to quickly ascertain whether your concerns have arise [sic] due to:

a) Misunderstanding of IFRS tests versus US GAAP tests;

b) The fact that (as you correctly state) you have access to only half of the jigsaw puzzle and whether other pieces have already been given to the auditors and Audit Committee (which, as you state, you have not been a part of) and they have taken them into account and already made suitable decisions;

c) An accounting policy weakness;

d) False information being provided to Autonomy or you have information Autonomy does not have which changes its current view; or

e) The fact that matters you raise have already been identified and considered as a failing by the Audit Committee and have been addressed.”

Dr Lynch also identified particular areas which he suggested required further elaboration and evidence, and in particular Mr Hogenson’s apparent concerns (as stated in Mr Hogenson’s initial email) about (a) there being “no evidence of sell-through by a reseller” ; (b) the use of VARs with “an inability to pay”; (c) what Mr Hogenson had called “Related Party Transactions” ; and (d) “Overstated barter deals.” Of course, on the conclusions I have reached, Dr Lynch knew the answers already, as did Mr Kanter and Mr Hussain.

However, Dr Lynch warned against:

“involving Sushovan and the rest of the finance team to preserve the integrity of the review. Thus please do not discuss this matter with them.”

He also emphasised the need for speed and invited a response by close of play, stating:

“ in order to quickly get a prima facie analysis of the situation I suggest we start with Q1 2010 as the large transactions in that quarter are fresh in the minds of those involved. Please by return can you list the deals you have concerns with along with the reasons and Andy and I as a first step will investigate what information was known and considered already but [sic] the auditors and share this with you”…

Lastly, Dr Lynch sought to reassure Mr Hogenson that he was “maintaining an open mind” and that “[i]f once we have compared the missing pieces we still have issues we will rapidly take this to the next level” ; but he also sounded a warning:

“I would ask that until we have performed the next step that the language you use is a little more moderate as until we confirm your concerns are valid and not just an artefact of partial information I would not want to inadvertently run the risk of accusing the innocent as emails have been known to escape. I would suggest you use the usual encryption.”

It is now clear that within minutes of sending the email attaching that letter, Dr Lynch also telephoned Mr Hogenson from his mobile telephone. Dr Lynch had not mentioned the call in his witness statements, and there was no evidence of it in the trial bundles. When he was first cross-examined about that telephone call, the Claimants questioned him with evident scepticism as to whether it had taken place at all. However, Dr Lynch’s recollection of there being a recording of it in disclosure proved to be correct: a recording was found showing that the call had commenced at 17:06.

In summary, Dr Lynch:

(1) Reiterated the point made in his letter that the concerns might reflect a misunderstanding of the important differences between US GAAP and IFRS;

(2) Warned against the danger of making an accusation of dishonesty against anyone without very careful checking, especially under English law, to which Mr Hogenson responded that he had worded his email very carefully “not to accuse anybody, just suggest that the evidence provides information around” ;

(3) Suggested that the most important thing was to “make sure that we’ve understood all the things and dealt with all the possibilities” , and asked Mr Hogenson to provide a few examples “to start with” and to determine whether they showed any real problem;

(4) Asked Mr Hogenson whether he was “happy with this as an approach” to which Mr Hogenson replied that he thought “it’s a fair approach, absolutely.”

With hindsight it seems to me to be clear that in fact the approach was not intended to be “fair”: it was intended to promote Dr Lynch and Mr Kanter’s strategy of keeping control of the process and what the Audit Committee was told.

Late in the evening on 23 June 2010, and in compliance with Dr Lynch’s rather peremptory request, Mr Hogenson sent a letter responding to Dr Lynch’s email earlier that day and giving more factual detail about, and examples to illustrate, his concerns in each of the areas identified by Dr Lynch. He sought to tie these examples in to the provisions of IAS 18.14. He also reiterated at the commencement of his letter his “desire to speak directly with the Audit Committee prior to the release of the Q2 2010 financial statements.”

Mr Hogenson identified the following concerns, each of which I have found to be justified in this judgment:

(1) The Capax Discovery / Eli Lilly deal ($6.3 million), on the basis that there was apparently no end-user and the reseller not having the ability to make the payments under the agreement;

(2) The MicroLink acquisition, as being between related parties without a proper valuation; and

(3) Purchases from FileTek, on the basis that they appeared to be barter transactions where the economic benefit on both sides of the transaction appeared to be materially overstated.

There followed a telephone conversation in which Dr Lynch probed for more details and emphasised that “sell through” (that is an end-user deal) was not required under IFRS accounting (unlike US GAAP with which Dr Lynch supposed Mr Hogenson was more familiar). Dr Lynch also promised to provide more detail on the FileTek fair value point. The call concluded with Dr Lynch saying:

“Alright, well look, I really appreciate the effort. I hope you've found this useful. We'll send you the document now if you need any more let us know. And, you know, I think once you've done the work you know, I'm very happy to set up a meeting with you, for the audit committee and let me know what you want to do. Obviously, if you've only got one barter and one related party given that those are very easy because we can look up the audit committee notes to see whether the tests were done we can knock those ones out very quickly. Alright, and Brent, I know this a lot of work and I really appreciate the work you've put in.”^{^[287]}

A sourer note was struck by Mr Kanter when emailing Mr Hogenson after that telephone call under the subject heading “Follow up” and attaching documents going over the matters discussed: in explaining the attachments, he added this:

“One thought that did occur to me. We know of a couple of rather dishonest characters who are fronting hedge funds, and by coincidence a couple of phrases you used were reminiscent of them. Please can I ask that if you are getting outside IFRS advice please be very mindful of the need for discretion and please be careful who you consult with. We are happy to arrange any independent training courses you may wish to take on IFRS.

Please note that all of this information, especially the board minute and audit pack extracts, are highly confidential and must not be shown to a third party without our written permission, but I hope you will find them useful.

I look forward to speaking again soon. Please call me any time with any questions…”

The call was followed by a joint letter to Mr Hogenson from Dr Lynch and Mr Kanter, sent as an attachment to an email dated 24 June 2010, setting out again Autonomy’s position on VAR transactions. The letter, signed by “Mike and Andy”, described Mr Hogenson’s original email in its second paragraph as having “made some potentially serious accusations against members of the senior management team” and in one of its closing paragraphs as containing “some very strong accusations against members of the management team” . It gave the impression, however, that nevertheless his concerns would be dealt with in a measured and detailed way, with a view (it stated) to enable him to understand more fully matters “to which you have not been privy, regarding revenue recognition and audit decisions.” The basic message was that (a) it was understandable that Mr Hogenson might have “jumped to some of the concerns you have expressed” but (b) there was so far no new evidence or insight to undermine or change the decisions made by the auditors and the Audit Committee at the time. The letter added:

“Obviously whilst I hope you agree that these particular points, although we may learn good policy lessons from them, once all the information is available do not seem to fit with the original accusations. I am course open to continue investigating any other issues you may have as an open, ongoing process. Consequently as we have now been through this exercise on Q 1 and you can now see the effect of the missing information I think it would be fruitful for you to look at 2008. If you could please classify any issues under your concern headings (bulleted below), I will be happy to continue with this exercise in collaboration with you.”

The letter concluded stating that if Mr Hogenson should find wrongdoing by members of the management team this should be referred to the Audit Committee; but warned that “as under such circumstances it would be required for you to name those at fault” it would be “negligent for us to have not considered all the information rather than just half of it before presenting it, especially under UK law.” Its last paragraph plainly emphasised the real message which was that Mr Hogenson should assemble and share whatever information he had, not with the Audit Committee but with Mr Kanter:

“Lastly, we have contacted the senior member of the audit committee who has suggested that we continue the process of merging each others information to see what we find. He has asked to be keep updated on progress, which will be done by Andy, as part of the standard process.”

On the evening of 24 June at 8.46pm, Dr Lynch emailed Mr McMonigall with what he described as a “quick update on the current status” . As previously explained, it seems to me that Dr Lynch had already had a brief discussion with Mr McMonigall, having tracked him to Nice; but this was the first occasion on which Dr Lynch went into any detail.

The Claimants emphasised various questionable features of what Dr Lynch told Mr McMonigall in that email ^{^[288]} , a draft of which Dr Lynch accepted in cross examination he had “ run past” Mr Kanter (to whom the email was also copied), and in particular:

(1) The description of Mr Hogenson as “a US financial employee of Interwoven…who had access to the US PO books but is lacking the other rev rec information” : which undoubtedly minimised his status and although not inaccurate did not reflect fairly that he was CFO of Autonomy in the Americas as well as President of Interwoven, and the highest ranking finance person in Autonomy based in the US (Autonomy’s largest market by far);

(2) The assertion that Mr Hogenson was refusing to name the senior members of Autonomy’s management team he thought might be responsible for accounting impropriety: which was inconsistent with (a) Dr Lynch’s oft-repeated assertion that Mr Hogenson had named Mr Hussain and Mr Egan, (b) Dr Lynch’s recollection of having told Mr McMonigall their identities also, and (c) Dr Lynch’s strict admonition to Mr Hogenson not to name them formally or in writing unless and until his concerns were fully researched, and furthermore, was calculated to give the impression that Mr Hogenson had not the conviction to name names;

(3) The conclusory tone and substance of the report, including the dismissive and patronising conclusion “At this stage nothing seems amiss and the guy is learning a lot but somewhat confused” ; which may have reflected Dr Lynch’s own assessment but was also calculated to trivialise the issues raised and condition Mr McMonigall’s mind on his first learning of the nature of the concerns.

(4) The final assessment that “He appears to be somewhat backing off his original statement but we shall see…” : which the Defendants breezily dismissed as “a fair comment in light of the conversation which Mr Kanter and Dr Lynch had had earlier that day with Mr Hogenson” but which was far from a complete and accurate depiction of the way the matter was left with Mr Hogenson (and was certainly not borne out by immediately subsequent events).

On 25 June 2010 at 9.43am, Mr Hogenson responded to Dr Lynch’s and Mr Kanter’s letter of the previous day. The letter said that he was “ confused by a few of your statements in the return mail ” and that this was probably due to the fact that they were “ not both accountants speaking with a history of dealing with phrases such as Fixed and Determinable ”. The letter continued:

“ As the CFO of Americas, with responsibility for a large portion of the consolidated financial statements as Americas makes up approximately 70% of total revenue, it may make sense that I have a direct conversation with Richard Knights our D&T Audit Partner ”

Dr Lynch forwarded that letter to Mr Kanter and in the early hours they worked on a response, which went through at least two drafts, the first of which was exchanged at 01.34am on 26 June, with further exchanges of minor details early in the morning before 9am. Dr Lynch sent the final response to Mr Hogenson at 08.39am on Saturday 26 June 2010, copying in Mr Kanter.

The response stated that Dr Lynch had contacted the senior member of the Audit Committee running the process and that the mandate from him was clear:

“ He wants you and I to work together to look at these specific examples that have given you concern and add the half of the information that you are missing. We then need to understand from you whether (on the basis of this full information) you are still concerned, and whether these concerns support your original statement about certain individuals, and once we have done this to then look at the overall materiality of these potential issues of concern over the periods you cite. I must stress the issue at hand is to investigate your original statement. I am also interested where you believe you have more information than was given to the company's auditors at the time of them considering these transactions ”.

The letter continued, “ To provide the senior director the information he needs there are a series of actions ”. Dr Lynch then set out a series of action points including:

“1: Please can you send any evidence you have of information that the auditors should have seen during the periods covered by their past audit operations that they did not see, i.e. matters arising pre-April 20, 2010.

2: Please can you provide this direct language from the master IFRS definition that requires the examination of the end-user contract between the reseller and its customer.

3: From what I understand, Interwoven, which you have managed the finances of for years, has always done a lot of its business through resellers, for example via the ishop purchasing process. In these cases I believe that no end-user contract is viewed or obtained from these resellers yet the revenue is recognised. Can you help me to understand why is this case a different one to the above?

4: Please send the other examples you said you have from 2008.”

The letter concluded:

“ Brent, I really appreciate your effort in this process, and although we must remain focused on your original assertion, as a separate process the two of us seem to be learning useful things that we should keep note of for future improvements in policy. I think it is important that whilst there may be many options open to the senior member of the audit committee for different things to do, at this stage I obviously must follow the mandate and our official policy for dealing with these matters. I look forward to continuing this discussion ”.

Although the letter referred to Dr Lynch having “touched base with the senior member of the Audit Committee”, and the request for information is couched in terms of being what that person has specified to be required, there is no evidence of any conversation between Dr Lynch and Mr McMonigall between 24 June and 26 June 2020. It may be that they had a conversation after Mr McMonigall’s receipt of Dr Lynch’s 24 June email; or it may be that the reference was simply to the first conversation where Mr McMonigall asked Dr Lynch to gather information.

Either way, I consider it more likely than not that the reality of the matter was that Dr Lynch (with Mr Kanter) were in truth operating with little, and most probably no, specific directions from Mr McMonigall, and the presentation to Mr Hogenson of the requests for information being driven by the senior member of the Audit Committee was not correct. It may be that this impression might have been corrected if Mr McMonigall had given evidence, but he was not called. The Claimants invited me to infer that Mr McMonigall could not have given truthful evidence to support Dr Lynch; I need not go further than to note that the impression I formed was in the circumstances not corrected.

Whether Mr Hogenson sensed any of this, or had other motivations, is a matter to which I return later. In any event, Mr Hogenson did not respond to Dr Lynch’s letter, which was the last substantive communication between them on the accounting issues. In the evening of 26 June 2010 (at 19.36 BST) Mr Hogenson sent an email to Deloitte headed “IMPORTANT AUDIT RELATED QUESTIONS” detailing his concerns to Deloitte and also noting that he had identified questions to Dr Lynch and had further conversations with him and Mr Kanter over the past week beginning 22 June 2010, but that despite his requests he had not been able to have any direct contact with the Audit Committee.

Mr Hogenson’s decision to write directly to Deloitte put paid to any idea that he was becoming persuaded of Dr Lynch’s explanations and preparing to back down. Dr Lynch’s response was not satisfactory.

Again, much detailed evidence was submitted by both sides which it would be disproportionate to recite at length. Suffice it to say, that having carefully considered it, I am satisfied on a balance of probabilities that the objective of Dr Lynch and Mr Kanter after Mr Hogenson had sent his email to Deloitte (which obviously brought an end to the efforts to kick his concerns into the long grass) became (a) to ensure that they were the persons in contact with Deloitte and that the Audit Committee were largely kept out of the loop; (b) to present to Deloitte the attitude of the Audit Committee (which had in reality not been involved substantively at all) as being that whilst of course an investigation would be needed, it should be confined to what was necessary to answer Mr Hogenson’s questions from an accountancy point of view, (c) to keep Deloitte well clear of any investigation of the other aspects of Mr Hogenson’s concerns, which in a nutshell were that the transactions might involve conscious impropriety on the part of some of those involved in the VAR sales in the USA (such as Mr Egan) and (d) thereby to procure that any investigation became an exercise for Deloitte of double-checking their previous auditing approach, and duly repeating it.

Thus, for example, Dr Lynch continued to be the person dealing with Deloitte on the part of Autonomy, and continued to relay the message that the Audit Committee considered the issue to be a dry issue of accounting; in the evening of 27 June 2010, Dr Lynch wrote to Mr Knights, copying Mr Kanter and Mr McMonigall:

“I spoke to John [McMonigall] who is sailing.

Due to his difficulty in communication he asked me to pass the following request to you. I understand when possible he will be intouch [sic] directly.

He would like D&T to consider the questions sent by Brent and rapidly revert back to him. He would only like to focus on material matters.”

The Defendants were assisted in this by the fact that Mr Hogenson did not include in his correspondence with Deloitte the references he had initially made to concerns about potential wrongdoing, and he framed his email in terms of questions rather than concerns: his correspondence was in terms that plainly reflected the warnings Dr Lynch had issued. In the event, Deloitte treated the letter as raising issues of accountancy judgement; they did not treat it as raising issues of potential wrongdoing, and did not regard Mr Hogenson as a “whistleblower” in that sense.

Deloitte’s own perception of its role and approach was summarised in the Deloitte Defence in the FRC proceedings (which was endorsed by Mr Welham in his evidence):

“307. The audit team (reasonably) did not regard Mr Hogenson’s questions to be “whistleblowing” properly so-called (that is, implicitly confidential allegations as to management wrongdoing or fraud made by an anonymous whistleblower) or as raising any allegation of wrongdoing on the part of the audit team. On that basis Mr Hogenson’s questions did not require an independent corporate investigation as might be required into a fraud and there was no need for ethical walls to be erected between the investigating team and the 2009 audit team.

308. It is averred that Mr Hogenson’s questions (which were addressed to Mr Knights and Mr Knight) could quite properly have been investigated by them alone as no allegation was made of wrongdoing on the part of the audit team.

309. That said, it was apparent that Mr Hogenson had raised potentially important questions as to, for instance, Autonomy's approach to revenue recognition and he had done so outside the usual internal Autonomy channels. Given that the audit team had accepted those judgements, there was an obvious benefit in involving personnel in the investigation who were not a party to those audit judgements. It was therefore decided in consultation with Deloitte's Head of Audit Risk, Simon Letts, that the judgements on the conclusions of the investigation would be made by senior members of the firm who had not been involved in the relevant audits. These were the new audit engagement partner, IRP and PSR director.

310. In consequence the Deloitte team which investigated Mr Hogenson's questions was different from the historical audit team. The historical audit team (that is, the team which conducted the 2009 year-end audit) comprised Mr Knights (engagement partner), Mr Henderson (EQAR), Mr Robertson (IRP) and Ms Bennett (PSR). Other team members included Mr Welham (Senior Manager), Ms Anderson (Manager) and Mr Murray (Assistant Manager).

311. The Deloitte team which investigated Mr Hogenson's questions comprised Mr Mercer (engagement partner) who led the investigation, Mr Brough (IRP), Mr Robertson (EQAR) and Mr Lumb (PSR director). Other team members who investigated or assisted included Mr Knights, Mr Knight (Director), Mr Welham (Senior Manager), Mr Milburn-Smith (Senior Manager), Ms Anderson (Manager) and Mr Murray (Assistant Manager).”

I discuss later the Regulator’s subsequent concerns in this regard. It is clear, however, that Dr Lynch and Mr Kanter had succeeded to a large extent in their objectives.

Two important disputed factual matters also deserve brief mention. Once more I shall not recite the factual detail of the extended dispute but I am satisfied that (a) as Mr Welham stated in his witness statement, Deloitte were not provided at the time with Mr Hogenson’s initial emails to Dr Lynch mentioning potential management wrongdoing and (b) nor was Mr McMonigall or the Audit Committee: or if they were, no one appears to have focused on them (see below).^{^[289]}

As regards the latter point, the email which Dr Lynch had earlier sent to the Audit Committee (cc Mr Kanter) forwarding Mr Hogenson’s email to Deloitte, which also was belittling of Mr Hogenson’s standing and dismissive of his concerns, seems to me to confirm that the Audit Committee had been told next to nothing and had either not been shown or had not focused on Mr Hogenson’s initial email:

“John

FYI

There is some background on this which we can discuss when you ring. The guy (US finance person from Interwoven) has played no role in the audit process and is missing a lot of the information he would need to understand all this. He seems to confuse USGAAP and IFRS.

Our current analysis: We just received this but a quick read’s conclusions:

Sushovan has seen the questions and believes there are no issues here. A few of the facts are just wrong or have perfectly good commercial explanations which were considered already by the AC, his understanding of IFRS differs from Deloitte’s (we believe Deloitte is correct) and all of the things he asks if Deloitte (and in most cases the audit committee) have considered and seen were indeed considered. We do not believe there is any new information here relating to any of the past auditing decisions”.

I acknowledge that, subsequently, Mr McMonigall stated in an email dated 8 July 2010 to Mr Hogenson (in fact Mr McMonigall’s first direct contact with him) that he had “seen all of your correspondence on this matter, going back to the first email to Dr Lynch, Autonomy’s CEO.” Naturally the Defendants much emphasised this. But Mr McMonigall was in reality detached from the process. The email was drafted for him by Mr Kanter and sent out without apparent review.

The subsequent review by Deloitte, the issues arising as to its independence and the inquiries on the part of the FRRP/FRC that followed are separate matters which I need not discuss here. Apart from mentioning that the result of all this from the point of view of Mr Hogenson was that he was sacked, purportedly by reference to the payroll fraud and “other irregularities”, made a settlement to forestall claims against Autonomy for alleged breach of statutory whistle-blower protections, and in exchange for $750,000 provided an affidavit calculated to give the impression that he had recanted and ensured that the matter could not be aired publicly, I need go no further.

The point in summary is that, in my judgment, the ‘Hogenson Episode’ provides further support for the conclusion I have reached that both Defendants knew that the VAR sales were illusory and improper, and were determined to avoid investigation that would reveal this.

The FSA and FRRP correspondence

I have been fortified further in my conclusions by the lack of candour, and in some instances dishonesty, of the way the VAR sales were presented to the FSA and the FRRP (to which I shall refer as “the FRRP”, since only the FRRP had jurisdiction) in response to their letters seeking clarification of their nature and purpose after Mr Hogenson and others (including Mr Tejada) had raised serious concerns about their impropriety (and see paragraphs 2232 to 2289 above as to the “Hogenson episode” ).

I have concluded that these responses were misleading in circumstances where candour and truth would have exposed the impropriety of their accounting treatment.

Although my focus in this section is intended to be on the falsity of what the FRRP were told, and whether the Defendants had knowledge of such falsity, it is necessary first to explain the genesis and content of the exchanges in question.

Mr Hogenson ’s concerns about Autonomy ’s accounting treatment of VAR sales

On 26 July 2010 Mr Hogenson had emailed the FSA to alert them to various questions and concerns he had raised with Autonomy about certain of its VAR sales and, in particular, their accounting treatment and what he thought might be associated “barter sales”. Mr Hogenson had presented his purpose as being:

“to provide assistance in ensuring that the Autonomy financial statements provided to investors are materially correct and are not misleading current or potential investors.”

Mr Hogenson had justified his direct approach to the FSA on the basis of the “nature and lack of response” to his questions when raised with the Audit Committee and with Mr Kanter (as its COO) and the appearance to him that his

“questions were closed by Autonomy without a serious review or even discussion with me.”^{^[290]}

Mr Hogenson had copied his email to Mr Robert Webb QC, who had become Autonomy ’s (non-Executive) Chairman in mid-2009. Receipt of the email appears to have prompted Autonomy to “reach out” to the FSA by letter dated 30 July 2010 “in the interests of open communication” (as Mr Kanter explained it in an email to the Audit Committee ^{^[291]} dated 2 August 2010)^{^[292]} . The Claimants took a different view of its purpose: in effect (though the words are mine), being to spike Mr Hogenson ’s guns and introduce doubt as to his reliability.

Autonomy ’s 30 July 2010 letter, which was signed by Mr Kanter on behalf of Autonomy (and given its nature, and the unlikelihood that Mr Kanter would have signed such a letter without their approval, must in my judgment be taken to have been approved by both Defendants) was presented as providing “background which may assist you in the matter” and included the following passages:

“In summary, Mr Hogenson raised a series of questions to Autonomy's governance bodies. The questions were thoroughly and independently investigated, and the matter concluded. It appears to us that Mr Hogenson is not accepting the conclusions on these matters, and is thus asking the same questions despite the answers being independently confirmed.

Investigation

Mr Hogenson served as Autonomy's CFO of the Americas for the last year or so. In late June Mr Hogenson raised unspecified accounting questions. The Autonomy directors took the questions raised by Mr Hogenson extremely seriously, and significant time and cost was expended to reach a satisfactory conclusion that the accounts are accurate. The matter was passed to Autonomy's Audit Committee and independent auditors for their review. In working with him via Autonomy's Audit Committee, the deal issues were defined and he was provided group-level information he would not have had access to in a regional role. Mr Hogenson was given the opportunity to deliver any additional information but declined.

Autonomy's independent auditors Deloitte, after an investigation of the points using an independent Deloitte team, reported to the Audit Committee there was no new information in Mr Hogenson's queries, and there are no areas that require change or would have been material in relation to prior periods. Because there was no new information provided by Mr Hogenson, ultimately the questions involved different levels of understanding of IFRS by Deloitte and Mr Hogenson (Mr Hogenson is experienced in US GAAP, not IFRS).

…

Conclusions

In short, a full and proper procedure has been followed to investigate Mr Hogenson's concerns, which involved an independent investigation by a team separate to our normal audit team at Deloitte. The investigation found no issues relating to the points raised and confirmed that Mr Hogenson provided Deloitte with no new information they did not already have in making their original decisions. Despite the questions being thoroughly independently reviewed Mr Hogenson continues to seek to keep the matter alive.”

The 30 July 2010 letter also contained a section entitled “Other Issues” . This, with elaborate but contrived moderation and obvious intent, made reference to a “payroll fraud” and other “unauthorised third party payments” which “coincidentally, at virtually the same time…began to be uncovered in Mr Hogenson ’s department.” It left hanging whether Mr Hogenson was implicated personally (which he was not, though a question was raised whether he should have spotted earlier shortcomings in controls which enabled the unauthorized payments to be made).^{^[293]} It did not mention that Mr Hogenson had already been dismissed, purportedly for gross misconduct. A little later, on 2 August 2010, Mr Kanter sent the Audit Committee an update, asserting that an independent investigation had confirmed over half a million dollars of unauthorised payments by Mr Hogenson “ and other irregularities ”, as a result of which Mr Hogenson had been “ let go ”.

After some equivocation, and only after being shown email exchanges confirming it, Dr Lynch had to accept in cross-examination that he had received and commented upon drafts of the 30 July 2010 letter.

Although its involvement in the process of investigation of Mr Hogenson’s questions and concerns was an important element, Deloitte was not shown the letter before it was sent. According to Mr Welham, they did not see it until many months later.^{^[294]}

It was the Claimants’ case that this omission was inexplicable, except on the basis that Mr Kanter and those who approved the letter knew that the presentation of the Deloitte investigation as having been by “an independent Deloitte team” and that it was an “ independent investigation by a team separate to our normal audit team at Deloitte ” were deliberately misleading. The Claimants contended that if Deloitte had been asked to approve the letter, Deloitte would have pointed out that it was untrue to say that, given the central involvement of Mr Knights, along with Mr Knight and Mr Welham. Their concern in that regard and generally would have also considerably been increased had they seen the letter dated 24 June 2010 from Dr Lynch and Mr Kanter to Mr Hogenson which I have referred to above and of which, according to Mr Welham’s evidence neither he nor his Deloitte colleagues were aware.

That is, of course, relevant to (a) the honesty of all those involved in the letter and (b) the integrity of the process to investigate Mr Hogenson’s concerns. In the present context, however, and as previously explained, my focus is on the specific misrepresentations made to the regulators after receipt of this letter and Mr Hogenson’s email with reference to the VAR sales.

To revert, therefore, to the correspondence with the Regulator, the matter was not pursued by the FSA, given the jurisdictional issue which Mr Kanter had identified, but it was referred to the FRRP ^{^[295]} .

On 2 February 2011, the FRRP sent a letter (“the February 2011 FRRP letter”) addressed to Mr Webb, and copied to Mr Hussain, asking for a further explanation of the issues raised in order to help it determine whether to open a formal enquiry. In the February 2011 FRRP letter, the FRRP posed (amongst others) the following questions regarding the Capax Discovery/Kraft transaction:

“Kraft

9. It has been alleged to the Panel that, in the quarter ended 30 September 2009, Capax gave an order to Autonomy worth some $4m in respect of the supply of licences to Kraft, an end-user, which was recorded in the revenue of one of the company’s US subsidiaries for that quarter. Subsequently, in December 2009, Kraft executed a licence agreement directly with Autonomy for a similar amount. It is understood that, in December 2009, the group relieved Capax of its obligations under the order given to the company and itself paid Capax a fee of some $400,000. The Panel would be grateful for the company’s comments on the truth or otherwise of this account.

10. To the extent that the above sequence of events has been confirmed by the company, the Panel would then be grateful for information enabling it to understand why the company appears to have contracted with both Capax and Kraft for the same licences.

11. The Panel would also like to know when the revenues concerned were recorded in the group’s consolidated accounts. If the revenues were recorded in the consolidated accounts for the quarter ended 30 September 2009, the Panel would then like to know how the group determined that it was in a position to recognise revenue relating to the sale of licences to Kraft in that quarter if the agreement concerned was not executed until December 2009.

12. Finally, the Panel would be grateful for information enabling it to understand why the group paid Capax a fee in respect of these transactions.”

Autonomy ’s response dated 3 March 2011 ( “the 3 March 2011 letter”) stated the following:

“In an almost unique series of events, Kraft sought to enter into a different agreement directly with Autonomy in a subsequent quarter. The latter agreement provided for direct payment to Autonomy, so even though the VAR was already paying Autonomy for its order the original transaction was cancelled …

… For the original order Capax would earn a normal software resale margin of approximately 30% under the terms of its VAR agreement, and in fact could have blocked the cancellation of the binding original order. Given Capax’s involvement in the original transaction, Capax was entitled under written agreements to a Marketing Assistance Fee of approximately 10%. Capax earned less from this revised arrangement than the original transaction.”

These statements to the FRRP in the 3 March 2011 letter were materially misleading in a number of respects:

(1) Far from being “ almost unique ”, by the time of this letter in March 2011, the pattern of Autonomy contracting directly with the end-user was well-established and highly recurrent. As stated in paragraph 2226 above, by 3 March 2011, this had happened 13 times.^{^[296]} The falsity of the representation was especially arresting given that at the time of the June letter to the FRRP, Autonomy had just entered into another sizeable direct deal (further to the Tikit/KPMG transaction, VT26).

(2) The attempt to justify the $400,000 payment on the theory that Capax Discovery could have blocked a direct sale by Autonomy to Kraft was false also. In fact, Capax Discovery was a complete stranger so far as Kraft was concerned. It had provided no “ marketing assistance ,” and it had no contractual or other right to block the direct sale.

(3) Moreover, Capax Discovery had not missed out on any “ resale margin ”. Capax Discovery never even tried to sell anything to Kraft. Capax Discovery’s VAR agreement did not provide for a “ normal software resale margin of approximately 30% ”.

(4) In any event, this transaction could never have involved any actual resale margin. Capax Discovery “ bought ” its licence for the same price that Autonomy was trying to sell a licence to Kraft.

(5) There was no written agreement, at the time of the Capax Discovery purchase order, which entitled Capax Discovery to a Marketing Assistance Fee of “ approximately 10% ” or to any fee.

The falsity of this answer was later compounded in a further letter from Autonomy dated 8 June 2011 to the FRRP responding to a request from the FRRP (in a letter dated 5 April 2011) for clarification as to what was meant by “ an almost unique series of events ” in the explanation they had been given as to why the Kraft transaction had culminated in a direct deal between Autonomy and Kraft. Autonomy gave the following explanation:

“This is a series of events that virtually never happens; in fact management struggles to recall any prior time that this happened, but hesitates out of prudence to give an absolute answer one way or the other. Thus the events were described as “almost unique …”

This response was even more misleading as at 8 June 2011, the Tikit/KPMG transaction had also been followed by a direct end-user deal (VT26). Autonomy had also, in the meantime, entered into three further VAR transactions (Capax Discovery/UBS (VT28), MicroTech/Bank of Montreal (VT32) and MicroTech/Xerox (VT33)) which would in due course end up being followed by direct deals between Autonomy and the end-user in a subsequent quarter.

Dr Lynch tried to justify the statement in Autonomy ’s 8 June letter by suggesting that the words “ any prior time ” in the second line of the paragraph may have meant “ before Kraft ” rather than “ before the date of the letter ”. As the Claimants observed, if that really was the basis on which Autonomy gave this answer to the FRRP, it was pure sophistry.

Dr Lynch ’s alternative explanation was an assertion that there were “ 10,000 reseller deals ” and the transactions at hand were “ still a very rare event out of 10,000 ”. No evidential support beyond this somewhat airy statement was ever provided. I would accept that there was a considerable number of VAR deals: but that must include many of trivial value. Further, Dr Lynch ’s submission in his written closing that the “Kraft situation, described as almost unique, was a rare one” was simply wrong: it was not the exception but the rule.

In any event, I do not accept the explanation offered. It was, to my mind, obviously misleading to present as “ almost unique” a sequence of events characteristic of so many of the deals of the same value and type as those to which the request for clarification related, and which taken together made up at least 88% by value of all VAR transactions above $1 million entered into by Autonomy during the Relevant Period.

Similarly, and also to my mind confirmatory of my approach, the easy and correct answer to another question from the FRRP about whether there were other contracts as at 30 December 2009 in respect of which revenue had been recognised on a VAR transaction, where the VAR had yet to conclude a transaction with an end-user, was not given. There were seven such contracts. But Autonomy ’s response did not say this. Instead, Autonomy ’s answer was an exercise in obfuscation and gave a false impression:

“All agreements between Autonomy and its resellers represent binding contracts whereby the reseller’s obligation to pay Autonomy is independent of whether or not the reseller gets paid by its customers. Moreover the software sold to the VAR is only licensed for a named end-user and thus can not be used for stock. Ultimately provided the reseller is creditworthy and all other revenue recognition criteria are met (e.g delivery, fixed price, etc.) then revenue is recognised upon receipt of the binding contract.

We do not require the reseller to provide us with confirmation that they have contracted with their end-users, as is normal in the software industry for standard product; rather resellers are required to identify the end-user for licensing purposes. It would be impractical in reality for us to confirm each contract with the final end-user. Under this approach Microsoft, for example, would have to visit offices to prove Word was sold-through by its OEMs and resellers. As end-users require ongoing support and maintenance it is not possible for a reseller to supply licensed software to another party without that transaction becoming known. Thus, whilst we may not have the details of when arrangements are signed between a reseller and their customer, the process works smoothly. This is the normal practice in the software industry for standard product.”

The implication of this response was that the VAR would attempt to contract with the end-user but that, for reasons of practicality, Autonomy did not seek confirmation of this fact. This was highly misleading: Autonomy did not intend that the VAR should make any attempt to contract with the end-user.

To add to this litany is Autonomy’s response to a further enquiry by the FRRP about the Kraft deal, the enquiry being the following:

“The Panel would also be grateful for confirmation or otherwise as to whether the “binding original order” between Capax and Kraft referred to in your letter had been signed by the two parties as at 30 September 2009.”

It seems from this that the FRRP had assumed that Kraft had placed a binding order with Capax Discovery but later decided to contract directly with Autonomy instead. Autonomy management, including the Defendants, knew that this was a misapprehension and that Kraft had never placed any order, or had any prior dealings, with Capax Discovery.

Autonomy nevertheless responded to the FRRP’s query as follows :

“The binding original order between Capax and Autonomy was signed by both parties as at 30 September 2009. For the reasons set out above we would not necessarily be aware of when Capax and Kraft signed their agreement.”

This failed to correct the FRRP’s misapprehension. On the contrary, it strongly implied that Autonomy management believed that Capax Discovery and Kraft had indeed “ signed their agreement ”, but that the date of that agreement was not known to Autonomy. In reality, Autonomy management knew that no such agreement would ever have existed.

Furthermore, in its letter of 5 April 2011, the FRRP had also raised the following :

“On page 19 of the Deloitte report, the management response refers to management’s consideration of the ability of Capax to stand by its obligation to Autonomy, irrespective of its ability to onward sell the Autonomy product supplied. The Panel would be grateful for information enabling it to understand what factors the company took into account in its consideration of the ability of Capax to stand by the obligation concerned. The Panel would also be grateful for copies of any financial information available to the company at that time concerning Capax. A recent set of accounts would be particularly helpful if available.”

In its 8 June letter, Autonomy responded to that as follows:

“Autonomy started to do business with Capax in early 2009. At that time we obtained financial statements from Capax (Attachment 1). These financial statements showed the company at that time to be profitable and able to support the payment stream required by purchase for Eli Lilly, part of the company’s revenue recognition criteria.

Capax have had an excellent payment record since that date. As a result there has not been a need to obtain more recent financial statements since their history of cash collection has shown no concerns regarding recoverability. These judgements were considered by our auditors at that time.”

The financial statements at Attachment 1 to Autonomy’s 8 June 2011 letter were those of Capax Global. Autonomy’s letter referred throughout to “ Capax ”. It did not explain that Capax Global was not, in fact, the relevant contracting party. Nor did Autonomy tell the FRRP of Mr Baiocco’s letter disclaiming Capax Global’s responsibility for Capax Discovery’s debts.

The assertions in Autonomy’s letter about Capax’s “excellent payment record ” were also misleading. The FRRP was not told that this payment record was the result of monies channelled by Autonomy to Capax Discovery for non-existent services.

Defendants’ knowledge but alleged reliance on Deloitte

Mr Hussain signed both the 3 March 2011 letter and the 8 June 2011 letter. But Mr Casey said nothing at all about the FRRP letters or Autonomy’s responses in his written closing submissions on Mr Hussain’s behalf, nor in his oral submissions on Day 92 of the trial. On this issue as to the misrepresentations in the letters to the FRRP there was a tension or conflict between Dr Lynch and Mr Hussain. Mr Hussain was also of course operating within the finance department, which Dr Lynch portrayed as having had carriage of the correspondence.

Dr Lynch maintained that he himself had no material involvement in the drafting process. According to Dr Lynch, the drafting was done by Autonomy’s finance department and Mr Kanter working with Deloitte. When cross-examined in relation to the March 2011 letter:

“Q. You were involved in the drafting of this; that's obviously what's going –

A. I wasn't involved in the drafting of it. Mr Kanter and Deloitte were involved in the drafting of it and then I look at it when it's finished and I say "Looks fine".”

There was a similar dispute as to whether Dr Lynch approved the June 2011 letter. He initially denied having done so, although this was not really plausible: there were emails from Mr Kanter sending earlier drafts and referring to “Mike…having a look”, and on 25 May 2011, Mr Kanter had sent Dr Lynch a further draft and asked him to “ Ring me when you have a moment ”. It seemed to me unlikely that having engaged him in the process, and apparently awaited his replies on each occasion, Mr Kanter would have gone ahead without Dr Lynch ’ s approval.

Dr Lynch eventually refined his position a little, and told me in cross-examination that he did not know if he had seen the final version, but did not think he had; and if he had, he had not checked or verified it, citing pressures on his time, the necessity to delegate and the reasonableness of relying on the finance department and Deloitte:

“…if we just come back to reality here…I ’m running a company, I ’m doing all the things that have to be done. This is a matter for the legal, accounting and Deloitte departments to deal with. If I saw the letter, I may well have read it, but I would certainly let them get on with answering it and that ’s what happened and legal, finance and Deloitte all worked together in a large amount of back and forth and came up with what they think is the correct answer.”

Ultimately, the main plank of his defence was to the effect that Deloitte had done it. This was the point emphasised in the closing submissions on his behalf. The point was made that Deloitte had undoubtedly been closely involved and had had many chances to comment on the drafting and made points both on the structure and style as well as on points of detail.

Thus, for instance, in relation to the 3 March 2011 letter, Mr Welham had emailed Mr Kanter on 17 February 2011, saying:

“From talking through your draft response with our internal specialist in this area, our overall thinking is that you should transform the letter somewhat from its current state, to a format which is much shorter and concentrates on the specific facts required by the letter from the FRRP.

…

We suggest that you keep your responses concise and concentrate on answering the questions asked and do not give additional information or include emotive language. Where it is appropriate, you should simply cross refer to the Appendix to our Audit Committee report, which helpfully includes detailed management responses from you, to avoid repeating information included in this report. This will assist in making your response to the FRRP shorter.”

The letter then went through various internal reviews at Deloitte with further comments provided before it was finalised. Deloitte were content with the final version, a copy of which was sent to them. Mr Welham confirmed that Deloitte did not consider that there was anything misleading in the letter.

A similar process appears from exchanges of drafts by email to have been undertaken by the relevant departments at Autonomy and by Deloitte for the June 2011 letter. Indeed it appears that Mr Welham referred both the FRRP ’s letter and Autonomy ’s proposed response to Ms Isobel Sharp, one of the Deloitte review team described by Mr Welham as “our internal specialist in this area” , who also (according to the email from Mr Welham requesting her input) had assisted on the previous response as well.^{^[297]}

The Claimants presented a short three-part answer to this line of defence: Deloitte did not know about Mr Baiocco’s letter, Deloitte did not know that Autonomy was paying for non-existent services and Deloitte did not know that Capax Discovery had had no dealings with Kraft.

As to the last of the three points, there was considerable dispute as to what Mr Welham and/or Deloitte did actually know. Dr Lynch emphasised that Deloitte reviewed the Kraft transaction, and were well aware of its context and that (a) Capax Discovery was only introduced because the end-user deal could not formally be completed in Q3 2009 (b) Autonomy undertook the negotiations with the end-user (Kraft) before and after the VAR agreement with Capax Discovery and (c) ultimately, Kraft (as end-user) insisted on contracting directly with and paying Autonomy (with Autonomy thereafter crediting Capax Discovery to avoid double recovery, and paying a MAF).

Dr Lynch’s closing submissions also sought to draw support from a summary of the position in Deloitte’s Defence in the FRC proceedings, which Mr Welham confirmed was accurate, which read as follows:

“…it is admitted that Deloitte and Mr Knights were aware (a) that Capax ’s principal role was to allow Autonomy to complete a sale and for revenue to be recognised in the quarter in which the Kraft deal was negotiated because it could not be formally completed with Kraft in that quarter, and (b) that Autonomy were negotiating directly with Kraft.”

In his cross-examination of Mr Welham on this point, Mr Miles sought to elicit from Mr Welham confirmation that the relevant issue in determining revenue recognition was not whether Autonomy, having sold the software to the VAR, was or was not negotiating with the end-user notwithstanding the VAR sale, nor even whether or not the seller (here, Autonomy) was the principal or even only person negotiating with the end-user after the sale to the VAR; it was whether the seller/Autonomy was under an effective obligation (which did not have to be a legally enforceable obligation) to negotiate with the end-user. Only in the latter case (of the seller being under some form of obligation), Mr Miles suggested (and eventually submitted to me), did the fact of it negotiating impact on revenue recognition, because only then would the fact of the seller negotiating suggest that it/they retained control or risk in the goods. He submitted that the explanation accords with Deloitte ’s report to the Audit Committee on the 2010 interim review :

“We note that management has responded to the concerns raised in Q1 2010 where for the first time we noted instances where deals had been credited and re-sold directly to end-users. If Autonomy is required to maintain ongoing managerial duties in respect of reseller deals or if the reseller cannot demonstrate its ability to pay for goods received then it would not be appropriate to recognise revenue on delivery of the product.” [My emphasis, reflective of the Defendants ]

Mr Welham would not be drawn that far. He agreed that if there was such a requirement or an obligation on the seller then that would be a clear and different case; in other cases, where the seller as a matter of its own choice, was simply “assisting the sales process and the close process” the decision was more nuanced, and would be a matter of judgement as to the extent of the “assisting”. However, he also made clear that if what the seller did went beyond assistance, and its involvement was “extensive” or ( a fortiori ) exclusive that would be “ problematic.” Indeed, my understanding of his evidence was that had Deloitte known that Autonomy had “more than an assisting role” and been in reality “the primary business closing the deal” it would not have approved revenue recognition. Another way of putting this, as it seems to me, is whether Autonomy had de facto control of the negotiations. In every one of the impugned VAR transactions, it did.

Accordingly, although perhaps a little glibly stated, the third part of the Claimants’ rebuttal of the Defendants ’ reliance on Deloitte ’s approval was also, in my judgment, sustained. As I have already determined previously that the first two parts of the rebuttal were proven, in my judgment Dr Lynch ’s reliance on Deloitte, which (apart from his plea that he was not involved, which I confirm I have rejected) was his only resort in seeking to neutralise the fact that Autonomy ’s letters to the FRRP were fundamentally misleading, could not be sustained because Deloitte themselves had been misled.

The admission in Deloitte ’s Defence in the FRC proceedings, which Mr Welham had confirmed was accurate and on which Dr Lynch had relied, does not assist the Defendants either on that analysis (if ever it could have been much help anyway). The admission that “Autonomy were negotiating directly with Kraft” does not touch on the point whether Capax Discovery were involved with them, or (put another way to reflect Mr Welham ’s evidence) whether Autonomy was in an “assisting role” or by itself and in exclusive control.

Summary in respect of the letters to the FRRP

In summary, in my judgment, Autonomy ’s letters to the FRRP did contain fundamental misrepresentations for which the Defendants were responsible. They are not absolved by the involvement of Deloitte, since Deloitte did not know the full facts and had indeed also been misled. The misrepresentations were intentional: they were necessary to prevent the VAR strategy being exposed and revenue recognition being disapproved. The position is the more telling in light of the fact that Deloitte, knowing only half the story, had expressly warned in its Audit Committee Reports for Q1 2010 and Q2 2010 that (as its warning was refined in the Q2 2010 Report):

“In Q1 2020…two deals sold to MicroTech in Q4 2009 were credited and resold directly to the two end-users. In our Q1 2020 report we highlighted that significant evidence of such further revenue reversals may jeopardise management ’s ability to recognise revenue at the point of sale to the reseller. Only one deal has been signed with the reseller MicroTech during Q2 2010 for $270k and the overall level of software deals done this quarter through resellers is significantly reduced.

During Q2 2010, a $6m licence deal originally with the reseller Capax Global from Q4 2009 was signed directly with the end-user, Eli Lilly and a Q1 2010 $4.2 million deal with Discover Technologies LLC was signed directly with the end-user Philip Morris….

We note that management has responded to the concerns raised in Q1 2010 where for the first time we noted instances where deals had been credited and re-sold directly to end-users. If Autonomy is required to maintain ongoing managerial duties in respect of reseller deals or if the reseller cannot demonstrate its ability to pay for goods received then it would not be appropriate to recognise revenue on delivery of the product. Management acknowledges this position and further highlights that there have been no significant software sales to resellers in Q2 2010…”

Of course, these matters do not give rise to a claim; but, with the Hogenson episode, they have confirmed me in my assessment of the impugned VAR transactions and the “guilty knowledge” of the Defendants, as summarised in paragraph 2229 above.

[1] The Claimants’ pleading defined the “Relevant Period” as “the period from (at least) Q1 2009 to Q2 2011”. The Claimants also criticized five transactions that took place in 2008, relating to what were referred to as “hosting transactions”. However, I have understood the Claimants’ claims for loss to refer to published information in Q1 2009 to Q2 2011; and I did not understand the Claimants to advance any claim in respect of any of Autonomy’s Accounts for 2008 or any of its published information in 2008.

^{^[2]} Autonomy’s financial quarters corresponded to calendar quarters and are referred to as Q1, Q2, Q3 and Q4 as appropriate.

^{^[3]} The expression “Pure hardware sales” is not an accounting term or a term of art: it was newly minted by the Claimants. In their RRAPoC the expression is defined as “substantial sales of third-party computer hardware (with or without third party software) without modification by Autonomy and unaccompanied by any Autonomy software.”

^{^[4]} This aspect of the group’s business was the provision by Autonomy or another group company (principally a company called Zantaz) of data storage, archiving, e-Discovery and retrieval services using its own hardware at one of its data centres. Hosting services were typically provided over a period of years, resulting in a reliable revenue stream.

[5] The Claimants acknowledged that the impugned transactions “represent a small fraction of Autonomy’s total transactions by number” but made the point that they were “very significant by value”, the impugned transactions being amongst the largest in terms of revenue in a given quarter.

^{^[6]} including (a) financial institutions (such as CitiBank, Barclays, BoA, RBS, Lloyds TSB, Deutsche Bank and Merrill Lynch, who used Autonomy’s software for archiving and regulatory compliance and litigation data management); (b) government and public-sector agencies throughout the world (including the FSA and SFO in the UK) who used such software to recognise unusual patterns in money flow to identify laundering and for other security and surveillance; (c) manufacturers (including Ford and General Motors) who used the technology to manage engineering know-how; (d) Telecommunications providers (such as AT&T, Ericsson, Cable & Wireless, BT and Vodafone) who used the software for call monitoring and analysis; (e) Pharmaceutical corporations (including AstraZeneca, GlaxoSmithKline and Pfizer) who used the software to keep up with changing regulations, demographic information, general R&D and litigation; (f) Media organisations (such as the BBC, ITN, MTV, Bloomberg, CNN, Reuters and Forbes) who used Autonomy’s software to manage their TV programmes, publish media, archive content, increase website traffic and advertising revenue; (g) eCommerce providers (such as Play.com, FedEx, Forbes and T-Mobile, which used the software to understand patterns in buyers’ behaviour and to monitor customer satisfaction and boost ‘upselling’); (h) Food and Beverage suppliers (such as Nestle, Coca-Cola and Britvic) who used the software to monitor product developments and market opportunities; (i) Intelligence and Defence organisations across the world to monitor and protect against security threats; (j) legal organisations including 75 out of the top 100 global law firms, who used the software for disclosure and litigation support; (k) IT companies (such as IBM, Oracle, HP and Lucent Technologies) who selected Autonomy software to support development; (l) Consulting and professional services customers (such as IBM Global, KPMG and PricewaterhouseCoopers) who used the software for profiling and data; (m) Energy and utility customers (such as BP and Shell); (n) Aerospace organisations (such as NASA, BAE Systems, Boeing and the US Air Force) which used the software for engineering knowledge sharing; and (o) Healthcare organisations (including the UK NHS, Eli Lilly and Blue Cross/Bleu Shield) which used Autonomy software to promote best practices and help protect patient safety and manage litigation.

^{^[7]} According to Dr Lynch’s evidence in his first witness statement, storage technology was rapidly advancing and data storage prices falling quickly: thus, the cost of storing a GB of data fell from around $1 million in the early 1980s, to around $10,000 in the 1990s, to around $10 in the early 2000s and to around $0.10 in 2010. There was huge demand; but obviously technological efficiency was the only means of securing competitive advantage.

[8] A specialisation emphasised by the Defendants because it was a key area where a product focused on structured data, and in particular, Autonomy’s SPE (see later), would have an application.

^{^[9]} The acquisition of StorHouse from FileTek was impugned by the Claimants: see paragraphs 2030 to 2035 below.

^{^[10]} Dr Lynch illustrated the connection between probabilistic theories and computing, and the use of an algorithm based on the rule that a large number of weak connections produce a better result than a single strong connection, by positing a search for the word “penguin”. In a simple search engine, you might get results for flightless birds in the Antarctic, the book publishers, a character in the Batman comic series, a US ice hockey team or a brand of clothing. However, if unstructured data contains contextual references to the Antarctic, ice, fish, feathers, eggs and black and white, there is a very high likelihood that the text is about penguins (the birds), even if the word “penguin” does not appear in the data.

^{^[11]} An Autonomy product called the Structured Probabilistic Engine (“SPE”). SPE was the subject-matter of a dispute arising principally out of the fact that Autonomy attributed considerable costs to its development and marketing whereas HP claimed that it was simply a hastily developed “repositioning” of IDOL technology containing no new features, which in reality cost next to nothing to develop and market but which was used as a cloak for some of the costs/losses of the hardware sales in Q3 2009. The Claimants relied on this as a basis for undermining the credibility of the Defendants.

^{^[13]} Mr Welham also confirmed that there were no major disagreements to his knowledge.

^{^[14]} Mr Cooke, currently the senior partner of the firm, provided a witness statement which the Claimants did not challenge.

^{^[15]} That is not to say that Dr Lynch should be taken to have pored over and absorbed the detail of the long spreadsheets that Mr Hussain sent to him regularly, which he plainly treated as more Mr Hussain’s way of working than something he had to master: but he did often require, and usually showed signs of having absorbed, a synopsis of such spreadsheets, as I shall come on to illustrate more specifically.

^{^[16]} Ms Orton left Autonomy to go to Invoke Capital with Mr Hussain, Mr Kanter, Dr Menell, Ms Eagan on around 23 May 2012.

^{^[17]} By way of illustration, I found only two references to Autonomy’s board of directors in the Claimants’ written closing of 2558 pages and 5225 footnotes; and five references in Dr Lynch’s written closing of 1692 pages and 4398 footnotes (two describing how it was comprised, and three dealing with major corporate acquisitions described later of (a) a company called Interwoven Inc and (b) a company called MicroLink).

^{^[18]} For example, the Defendants relied on the equivalent paper for H1 2011 as showing Deloitte specifically considering whether particular revenue was correctly categorised as OEM revenue, though the passive language is unclear whether Deloitte were noting what they were told or bringing judgment to bear.

^{^[19]} The planning process started before the end of each quarter.

^{^[20]} This was also summarised in a note to Mr Knights, copied to Mr Welham, from Ms Antonia Anderson on 26 January 2010, although some concern on behalf of Mr Barden about the issue of disclosure is apparent: “Discussed this with Phil and he agreed that based on their argument that IDOL is pervasive and that decisions are made based on info that is not split by brand it does not seem unreasonable that the information is not available in order to provide subdivision of revenue into product type, but he thought the current explanation should go further to explain why this is not available. Something along the lines of: “Although branded differently, in effect all, customers buy the same one product and for financial reporting the group does not disaggregate into different categories by brand or any other method.”

^{^[21]} According to the New York Times, Mr Quattrone had become known a s “the go-to-guy for sellers looking to command sky-high prices".

^{^[22]} On the Claimants’ case the “miss” was the more notable because it was "despite revenue being inflated by 27% through the full array of fraudulent mechanisms".

^{^[23]} After HP's acquisition of Autonomy was announced, Mr Larry Ellison of Oracle said on a conference call that "Autonomy was shopped to us” but that the price had been "absurdly high" . Dr Lynch categorically denied this, even suggesting that any approach was unauthorised, which he must have known was incorrect, since he knew that Mr Quattrone had put Oracle's name on a list of the most likely buyers for Autonomy.

[24] According to Mr Sarin’s witness statement, which was not challenged in this respect, the SCD group included 30 to 40 people split between two sub-teams: one (“Strategy”) focused on corporate strategy, the other (“Corporate Development”) on identifying and carrying out acquisitions to execute that strategy. The head of the SCD Group was Mr Robison (who was HP’s Chief Strategy & Technology Officer). Mr Johnson led the Corporate Development team (of which Mr Sarin was part, as a Senior Director). Mr Girish Nair was head of the Strategy sub-team. Both Mr Johnson and Mr Nair reported directly to Mr Robison.

^{^[25]} In cross-examination, Mr Apotheker described “middleware” as being, in simplified terms, “software technology that sits between the application and the actual data layer and that enables the system to actually work as one”. “Middleware” was another part or aspect of the technology “stack”.

^{^[26]} The terminal growth rate used was 4%, and the weighted average cost of capital 10.5%. The 4% terminal growth rate was a rate selected by Mr Sarin’s team.

^{^[27]} At $34 HP would have ceded more than 40% of the synergies.

[28] Consideration was also given to a company called Software AG (see paragraph 181 below).

^{^[29]} I was told nothing about the reasons for the choice of Mr Lane, nor about Mr Lane’s experience, though I gathered it to be in software.

^{^[30]} Mr Apotheker also (initially hesitantly) suggested in this passage that he “believe[d]” Dr Lynch had said, in explaining “how he was able to deliver these kinds of results”, that Autonomy was a “very focused pure play software company” . At §38 of his witness statement Mr Apotheker said that Dr Lynch had spoken of its ““pure software” model” . When cross-examined, he explained his understanding of this to be that “it’s a company that tries to do as much software as it possibly can, to the exclusion of everything else” and subsequently that Autonomy’s “business model is essentially driven by software and software only”. The Defendants submitted that this is likely to be something suggested to Mr Apotheker, and his repeated statement that this was something he “believed” had been said indicated that he did not in fact recall it, and it is unlikely that anything of this kind was said. I return to this later.

^{^[31]} As Mr Apotheker agreed, Software AG was strong in process management. It was not operating in the unstructured data field, nor doing search or content analytics.

^{^[32]} The references to numbers were to revenues.

[33] Mr Apotheker told the Board of HP at the time that the business case for the Acquisition hinged on synergies. In an email to Mr Lane dated 5 September 2011 (and thus after the Acquisition, abut before his removal as CEO) Mr Apotheker reminded him that “Autonomy makes total sense if one believes that HP can generate the synergies we build into our business plan. The quality of the synergies is high: you will remember that they exclude any drag-on revenues related to additional hardware sales and we only included a very small drag-on effect for services. All the other synergies are driven by leveraging the IDOL platform, combining it with HP IP/R&D, deeper penetration of existing markets and significant and identified upsell/cross sell opportunities…I for one, and so does my team, firmly believe that we can achieve these synergies in the allotted time frame.” The email chain was forwarded to Ms Whitman by Mr Lane on 25 September 2011 (after Mr Apotheker had been removed).

[34] It was the lower figure of $2.2 billion forecast revenues in 2014 which was used as the input for the DCF valuation, which attributed a stand-alone value to Autonomy of $10.3 billion and a value inclusive of synergies of $17.6 billion.

^{^[35]} As I explained earlier, HP did not accept this. Part of their case was that Dr Lynch had planned a sale with the encouragement of Mr Frank Quattrone and his boutique investment bank called Qatalyst Partners, which from late 2010 was “shopping” Autonomy to the “potential acquirer universe” including HP.

^{^[36]} Mr Apotheker did not read the KPMG draft due diligence report on the financial aspects.

^{^[37]} In this passage Mr Apotheker suggested that this was at Autonomy’s request, but as he accepted, HP considered this a good idea. The documents show that HP itself was advised about limiting due diligence because of the interloper risk. Mr Apotheker accepted (in the passages quoted in this paragraph) that HP structured due diligence so as to minimise interloper risk.

[38] Mr Sarin and Mr Gersh told me that it was not unusual that no final report was issued, nor was it unusual for reporting accountants to flag tasks in their statement of work (“SoW”) that they had been unable to complete. However, in this case, the list of outstanding or yet to be completed matters were significant: and see paragraph 270 below. KPMG’s warning that they had not completed due diligence needs to be seen in that context.

^{^[39]} During the period from 25 July to 27 July, Autonomy’s share price had risen from £16.62 to £17.20; by 5 August it had fallen to £15.44.

[40] BarCap’s overall message was plainly positive, highlighting that Autonomy had “demonstrated double digit organic growth even during past downturns” and “[d]elivered 40% + operating margins over the last 3 years - among the highest in the software industry”. BarCap’s expressed view was that the combination of Autonomy and HP was “[h]ighly complementary with minimal overlap…” and would “form the basis of the next generation information platform”.

[41] The suggestion that it was Dr Lynch who deployed the term is supported by an email from Dr Lynch to Ms Whitman on 26 October 2011 (just a few weeks after the Acquisition) in which he twice used the term “stepchild” in describing Autonomy’s status in HP, albeit in urging HP to welcome “this talented stepchild” which would “need a little bit of help getting it accepted in the early days…” Much later, in September 2017, Dr Lynch appears to have used the phrase in an interview with the Evening Standard, stating his view that “if Leo Apotheker, HP’s chief executive who was fired shortly after the takeover, had stayed it could have been “an industry-changing deal”, but instead, “we were left there as an unwanted stepchild.”

[42] Ms Lesjak explained in cross-examination that such a pre-meeting took place every quarter to talk through points that had arisen in the quarter. Usually only Ms Lesjak, Mr Jim Murrin (who was the controller of the EY team at that time) and the key audit partners of EY would attend, but others also did sometimes. After first circulating an earlier draft of this judgment, the Claimants reminded me that, contrary to the Defendants’ submission and my assumption, there was no evidence before the court that Ms Lesjak received the presentation before the pre-meeting. However, even if not pre-circulated, it was not disputed that the two-page document was discussed at the pre-meeting.

^{^[43]} HP’s Enterprise Servers, Storage and Networking Division.

^{^[44]} Dr Lynch accepted that Mr Hussain probably joined the call.

^{^[45]} Thus, for example, Ms Meeta Sunderwala (a senior director in HP’s EFR team), reported in an internal HP email dated 28 June 2012 that Ms Lesjak had “told EY that the Autonomy issues were related to short-term challenges with working in the HP environment and not an issue with the overall market or sales funnel.” Ms Sunderwala noted also that “[t]ypically, it is very unusual to impair an acquisition in the first year post close” and that “Based on the comments from Cathie, we were not planning on doing an impairment analysis in Q3…”. Ms Lesjak told me that she did not remember telling EY the above; but Ms Lesjak was, in my assessment, prone not to remember things which tended to embarrass her now or cast doubt on her cultivated image as prudent, assiduous and a safe pair of hands.

[46] Or, as the Claimants preferred to describe it, “challenges with operating Autonomy in the HP environment”.

^{^[47]} Mr Scott gave his evidence in the US criminal proceedings, which was admitted in these proceedings as hearsay, on terms which gave him substantial immunity. I shall have more to say about its reliability later. In a nutshell, it was largely self-serving, and designed to insulate himself from threatened charges of wire fraud in the US (which carry a penalty of up to 20 years’ imprisonment for every count) for his own role and complicity. In his meetings with HP and with the US DoJ in relation to these matters, Mr Scott was recommended to and represented by the law firm also representing HP.

^{^[48]} In cross-examination, he rowed back from this, caveating it as “ the economic substance of the Autonomy business as a software business ” (signifying, as I understood this, that the exercise was to examine what businesses were a complementary part of Autonomy’s software business, rather than to examine their propriety or accounting treatment).

^{^[49]} Ms Lesjak suggested that PwC were also undertaking a “forensic investigation” at this time: but as Mr Nigel Curl of PwC (in Forensic Services) confirmed in an email to Mr Yelland of 12 July 2012, the input that PwC were assisting Mr Yelland with was limited to a review of the methodology. Mr Yelland again confirmed this in cross-examination, though he added that PwC “were themselves doing work on hardware at the time, which is work [he] used to help inform the rebasing exercise”. It is clear that PwC were retained in about May 2012, and both the engagement of “Forensic Services” and the rubric “privileged and confidential" on their exchanges with Mr Yelland as noted in paragraph 372 above suggest some anticipation of proceedings: but PwC did not produce a written report until March 2013 and there is a disputed question as to quite what their role was in any “deep dive review”: see below.

^{^[50]} HP’s Enterprise Financial Reporting group focused on M&A technical accounting.

[52] The Claimants’ case, however, is that the rebasing exercise assisted the preparation of a new ten-year forecast for Autonomy which was in turn used to generate a ten-year DCF model which then informed the impairment analysis of HP’s business units (including Autonomy). The Defendants themselves relied on the fact that “HP…had Mr Yelland’s rebasing conclusions when they considered whether an impairment was necessary in August 2012”.

^{^[53]} Ms Lesjak attempted to suggest that this was not accurate, but accepted that the memo was prepared by senior people within her department and that they must have got this information from HP senior management. But it is consistent with the document recording what Ms Lesjak told EY at the time. In any event Ms Lesjak accepted that HP concluded that there were no impairment indicators at the time.

^{^[54]} Mr Joel Scott, Chief Operating Officer and General Counsel of Autonomy Inc.

[55] As CFO, Ms Lesjak was the very person whose responsibility it was as CFO to recommend the impairment.

^{^[56]} Ms Lesjak did attempt to suggest that the second “bucket” was to do with “ accounting improprieties ”, but could not explain how and was forced to accept that it was limited to factors such as bad debt and one-off costs; they were nothing to do with the allegations relating to hardware, channel sales, hosting revenue or anything of that kind.

[57] The Claimants invited me to note that PwC were involved in interviewing Autonomy personnel, and in particular, interviewed Ms Harris for some three hours on 13 November 2012. The “Memorandum of Interview” recorded that this was “in connection with HP’s internal investigation of Autonomy’s accounting practices before, during, and after HP’s acquisition.”

^{^[58]} HP’s Strategy and Corporate Development group.

^{^[59]} The share price on 1 August 2011 had been $35.20.

^{^[60]} The Palm business had been bought for $1.2bn in 2010.

^{^[61]} Ms Sunderwala noted: “ Add 1.25ppt discount rate to Autonomy and ES (most execution risk)…You can add 0.25ppt to Software, TS and ESSN if looks like we need more downward pressure; otherwise keep as is (not sure if Software can handle it…) ” Further documents also showed a desire to avoid increasing the discount rate for the Software division.

^{^[62]} Ms Lesjak was aware that this was the preference of the tax department.

^{^[63]} Email exchanges over 23 to 24 October 2012 show that this was done overnight. Ms Lesjak did not argue with this. Ms Lesjak attempted to suggest that the $11bn book value of Autonomy was in fact made up of $1.5bn of synergies and $9.5bn “stand-alone value”, but that was incorrect.

^{^[64]} Mr Youngjohns explained in a blog at the time that HP had prepared one of the most extensive communication plans he had ever seen in 30 years. He explained in cross-examination that he had never “ seen anything quite like this in any other role ”.

^{^[65]} I suppose it is possible that short sellers and/or derivative traders may claim to have been indirectly prejudiced; but an irony is that all shareholders of Autonomy enjoyed a premium greater than on HP’s case they should have received; and there is nothing to suggest that any customers or third parties suffered any loss on any transactions with Autonomy on either side of the Atlantic. Shareholders in HP would have their own recourse for loss allegedly resulting to HP, subject always to rules restricting claims for derivative loss.

^{^[66]} According to the Defendants’ analysis, prior to the US criminal proceedings, Mr Egan had had 5 meetings with HP lawyers and 17 meetings with US prosecution lawyers in addition to meetings with his own lawyers. Mr Baiocco met with HP’s lawyers some 3 or 4 times, and with US prosecutors some 6 times. By the time of this trial, he had also given evidence at length to the Grand Jury and in the US criminal proceedings. Mr Baiocco had his own lawyers, paid for by HP. Other witnesses had fewer but more than one meeting each with US prosecutors.

^{^[67]} Financial Services and Markets Act 2000 (Liability of Issuers) Regulations 2010/1192, regulation 3 (Transitional provision).

^{^[68]} Or, under Sch 10A, acquired, continued to hold or sold.

^{^[69]} For an overview of the liability of issuers prior to the enactment of s. 90A, see Professor Paul Davies QC, Davies Review of Issuer Liability: Discussion Paper (HMSO, March 2007) (“the Davies Review”).

^{^[70]} Davies Review of Issuer Liability: Discussion Paper.

^{^[71]} See Davies Review of Issuer Liability - Discussion Paper para 44.

^{^[72]} Explanatory note to Companies Act 2006, s. 1270, paragraph 1637.

^{^[74]} Davies Review of Issuer Liability: Discussion Paper (above).

^{^[75]} Explanatory note to the Companies Act 2006, paragraphs 1637 and 1643.

^{^[76]} Margaret Hodge MP, Hansard , HC, Standing Committee D, Col, 482 (6 July 2006): https://publications.parliament.uk/pa/cm200506/cmstand/d/st060706/am/60706s02.htm

^{^[77]} That is, any reports and statements published in response to a requirement imposed by a provision implementing Article 4, 5 or 6 of the Transparency Directive, and any preliminary announcement of information to be included in such a report or statement: s. 90A(1).

^{^[78]} The relevant paragraph in the RRAPoC is confined to earnings calls in respect of Quarterly Reports published after 1 October 2010.

^{^[79]} Albeit in the context of the Misrepresentation Act 1967, rather than FSMA.

^{^[80]} See also Barley v Muir [2018] EWHC 619 (QB) at [177] per Soole J, contrasted with implied misrepresentation at [178]. See also Primus Telecommunications Plc v MCI WorldCom International Inc [2004] EWCA Civ 957 per Mance LJ at [30] (a case of implied misrepresentation but not in terms confined to that, Mance LJ also left the door open to a different approach in fraud). See also Kyle Bay Ltd v Underwriters Subscribing under Policy No. 01957/08/01 [2007] 1 CLC 164, at [30]-[33], per Neuberger LJ.

^{^[81]} Invariably for a claim under s. 90A (which, broadly speaking, applies only to information in quarterly, interim or annual reports); but not invariably for a claim under Sch. 10A, given the broader range of “published information” that such a claim may cover.

^{^[82]} Sch 10A §3(2) and 3(3). S. 90A contained a similar concept: see s. 90A(4) (“ person discharging managerial responsibilities within the issuer in relation to a publication ”).

^{^[83]} See the discussion of Derry v Peek and Armstrong v Strain at paragraphs 549 and 550 below, in the context of the claim in deceit.

^{^[84]} On that basis it would not be enough that Deloitte considered further disclosure might be preferable on a particular point.

^{^[85]} And see paragraph 472 below and the decision in Ivey.

^{^[86]} The same argument applies in relation to omissions: Sch 10A §3(3).

^{^[87]} Sch 10A §3(4) and, in similar but not identical terms, S. 90A(5).

^{^[88]} S. 90A(5): “ A loss is not regarded as suffered as a result of the statement or omission in the publication unless the person suffering it acquired the relevant securities– (a) in reliance on the information in the publication… ”; Sch 10A §3(1): “ An issuer of securities to which this Schedule applies is liable to pay compensation to a person who—(a) acquires, continues to hold or disposes of the securities in reliance on published information to which this Schedule applies… ”.

^{^[89]} See also the Court of Appeal’s comments on that conclusion: “ The judge may very well be right in his conclusion that, as a matter of law, no such cause of action exists as a matter of principle. But it is conceivable that in certain exceptional circumstances, for instance where the defendant, by the very making of the deceitful statement or for some other reason, had assumed liability to the claimant, a cause of action could exist .” Chagos Islanders v Attorney General [2004] EWCA Civ 997, at §36.

^{^[90]} Sch 10A §8(3): “ References in this Schedule to the acquisition or disposal of securities include— (a) acquisition or disposal of any interest in securities, or (b) contracting to acquire or dispose of securities or of any interest in securities, except where what is acquired or disposed of (or contracted to be acquired or disposed of) is a depositary receipt, derivative instrument or other financial instrument representing securities. ”

^{^[91]} See RRAPoC §§97-201, under the heading “ Liability of Autonomy to Bidco under Sch 10A FSMA ”.

^{^[92]} As to Bidco’s incorporation, it is pleaded in the RRAPoC that: “ Hewlett-Packard Vision BV (“Bidco”) was incorporated in the Netherlands on 15 August 2011 ”. For the draft due diligence report, and HP’s board meeting, see paragraphs 265 and 294 above.

^{^[93]} Davies Review of Issuer Liability: Discussion Paper (above).

^{^[94]} See J Cartwright, ‘Misrepresentation and Non-Disclosure’ (4 ^th Ed.) at [3-50].

^{^[95]} Sch 10A §3(1); and see to similar effect s. 90A(3).

^{^[96]} The above passage shows that, once inducement has been established as a fact it is irrelevant to ask whether the Claimants would have acted in the same way had the statements made in the published information been true. However, as discussed below, it is relevant to ask the different question, whether the Claimants would have acted differently if accurate information had been published (assuming arguendo that the Claimants succeed in their case that there were inaccuracies in it).

^{^[97]} Edgington v Fitzmaurice (1885) 29 Ch D 459, at 483, followed in BV Nederlandse Industrie van Eiproducten v Rembrandt Enterprises Inc [2019] 1 Lloyds Rep 491, at §§28, 32 and 43 (Longmore LJ).

^{^[98]} Dadourian v Simms [2009] EWCA Civ 169 at §§99 and 101 (Arden LJ).

[99] Lord Mustill put it this way in the Pan Atlantic case (at [551]): the representor

“will have an uphill task in persuading the court that the…misstatement…has made no difference…there is a presumption in favour of causative effect.”

^{^[100]} S.90A(5) is in materially the same terms.

^{^[101]} See Davies Review of Issuer Liability: Final Report para.30 p18.

^{^[102]} Assuming ( arguendo ) that the call in question constitutes published information.

^{^[103]} As explained below, to the extent that analogies with the common law are relevant, loss and causation of loss is a separate element from inducement, and requires ‘but for’ causation.

^{^[104]} Again to the extent common law analogies are relevant, see the discussion of loss in relation to deceit below.

^{^[105]} Contrast the hypothetical example given by Christopher Clarke J in RZB v RBS where “ P buys a house from V. He had been considering several houses. He is minded to buy the one which he eventually buys because of its size, shape and character. Shortly before he makes his ﬁnal decision V’s agent tells him that a particular celebrity has the house next door, a circumstance which he regards as advantageous. … In fact, as it turns out, there is no celebrity next door. ” In that example, if the agent had not chosen to tell a lie, nothing would have been said about whether there was a celebrity next door. There was no obligation on V or his agent to volunteer information about whether the neighbours were celebrities.

^{^[106]} Thus, the Claimants averred as their primary case that the bid would have proceeded, albeit at a lower price. By amendment the Claimants pleaded an alternative case that the bid would not have proceeded (see further below), in which case their loss would be the difference between the price that Bidco paid for Autonomy and Autonomy’s true market value on the date that Bidco’s offer became unconditional.

^{^[107]} Contrast the language used in s. 2(1) of the Misrepresentation Act 1967, discussed at paragraph 570 ff below.

^{^[108]} See also Smith New Court at 279-280

^{^[109]} As argued at paragraphs 17 and 18 above, in the present claim, the FSMA claim against the issuer is being used as a stepping stone to the claim against the PDMRs; but that should not affect the court’s interpretation of s. 90A and Sch 10A.

^{^[110]} See paragraph 437 above - and contrast the position with a prospectus.

^{^[111]} There is another passage in the judgment of Hobhouse LJ, in the section dealing with damages, which states that “ In general, it is irrelevant to inquire what the representee would have done if some different representation had been made to him or what other transactions he might have entered into if he had not entered into the transaction in question ”: p 441 at B-C . However, the words need to be read in context: on the facts the judge had decided that it was a no transaction case; and in such cases it is indeed irrelevant to go on to ask whether other hypothetical deals might have occurred. That has no application in a case where the court concludes that in the counterfactual world the same transaction would have occurred in any event; the question then is whether the claimant can show that it would have occurred on the same or different terms. Furthermore, as Leggatt J noted at §217(1) of Yam Seng , the dictum of Hobhouse LJ cannot have been intended to state a proposition of law.

[114] After circulation of an earlier version of my judgment in draft.

^{^[115]} Cartwright, Misrepresentation, Mistake and Non-Disclosure (4 ^th Ed., 2017), §5-07.

^{^[116]} Derry v Peek : “ Now, as to the evidence. The plaintiff's case is that the defendants made an untrue statement, which they knew to be untrue, and likely to influence persons reading it; therefore they were fraudulent. It is not necessary to consider whether a prima facie case was made out by the plaintiff. We have all the evidence before us, and must judge on the whole. The alleged untrue statement is that, “The company has the right to use steam or mechanical power instead of horses,” and that a saving would be thereby effected. Now, this is certainly untrue, because it is stated as an absolute right, when in truth it was conditional on the approval of the Board of Trade, and the sanction or consent of two local boards; and a conditional right is not the same as an absolute right. It is also certain that the defendants knew what the truth was, and therefore knew that what they said was untrue. But it does not follow that the statement was fraudulently made. There are various kinds of untruth. There is an absolute untruth, an untruth in itself, that no addition or qualification can make true; as, if a man says a thing he saw was black, when it was white, as he remembers and knows. So, as to knowing the truth. A man may know it, and yet it may not be present to his mind at the moment of speaking; or, if the fact is present to his mind, it may not occur to him to be of any use to mention it. …”

^{^[117]} Cited with approval by Rix LJ in AIC Ltd v ITS Testing Services (UK) Ltd (The Kriti Palm) [2007] 1 Lloyds Rep 555.

^{^[118]} Clerk and Lindsell on Torts (22 ^nd Ed, 2017), §18-30, and see §77 in Eco 3 Capital Limited v Ludsin Overseas Limited [2013] EWCA Civ 413 (set out at paragraph 546 above).

^{^[119]} Zurich v Hayward [2017] AC 142, at §34 (Lord Clarke); and Rembrandt Enterprises at §25.

^{^[120]} Males J found in that case that MOG had discharged the burden of shifting the presumption, even though, as he noted, amongst the problems for a defendants is that “…it is not attractive, and will generally be unconvincing, for a fraudster…to assert that its fraud, which it may have gone to some lengths to perpetrate, has actually made no difference…”

^{^[121]} Livingstone v Rawyards Coal Company (1880) 5 App. Cas. 25.

^{^[122]} As to this element, see Raiffeisen Zentralbank v RBS [2010] EWHC 1392 (Comm) at §221: “ The requirement that a representation must be made with the intention that it should be acted on by the other party applies equally under section 2(1) of the Misrepresentation Act as in a fraud case ”.

^{^[123]} Cartwright, Misrepresentation, Mistake and Non-Disclosure (4 ^th Ed., 2017), §7-17.

^{^[125]} Yam Seng at §206: “ The decision in the Royscot Trust case has been much criticised. As has been pointed out by academic writers, the policy considerations which justify a broad measure of damages where fraud has been demonstrated do not apply, or in nothing like the same degree, in cases of mere negligence. Nor does the language of s 2(1) seem to me to compel such a conclusion. It is possible to construe the words ‘and as a result thereof … has suffered loss’ as requiring the claimant to show that he has suffered loss as a reasonably foreseeable result of a misrepresentation having been made to him, and to treat the following words as imposing an additional requirement (that the defendant would be liable to damages had the misrepresentation been made fraudulently) which must also be satisfied. Unless and until the Royscot Trust case is overruled, however, it represents the law; and I must therefore apply it. ”

^{^[126]} because Autonomy had to purchase hardware from third party suppliers, and the purchase price paid by Autonomy exceeded the amount received by Autonomy in respect of the onward re-sale.

^{^[127]} Gross margin is the margin on each incremental sale of a product; or, across a company, the average gross margin across all sales.

^{^[128]} The meaning which the Defendants insisted they intended was critically different from the meaning the Claimants insisted the phrase conveyed to the reasonable reader. This was the subject of extended dispute, both as to the facts and the law: this is summarised in paragraphs 1609 to 1632 below.

^{^[129]} It was not clear to me when it was first introduced in Autonomy’s published information and other presentations. However, it is to my mind of some relevance to the dispute as to its intended meaning that it was in use by early 2008 and, for example, appeared in the ‘Strategy Review’ in Autonomy’s report for Q1 2008 [L/32/9] and thus before the hardware reselling programme commenced.

^{^[130]} See also the following passage from the judgment of Holroyd J in Ewer v Ambrose (1825) 3 B. & C. 746, at 750 (cited by Brooke LJ in McPhilemy v Times Newspapers Ltd (No 2) [2000] 1 W.L.R. 1732, at 1739F): “ if a witness proves a case against the party calling him, the latter may show the truth by other witnesses. But it is undoubtedly true, that if a party calls a witness to prove a fact, he cannot, when he finds the witness proves the contrary, give general evidence to show that the witness was not to be believed on his oath, but he may show by other evidence that he is mistaken as to the fact which he is called to prove .”

^{^[131]} Though the Defendant latterly relied on an email exchange between Dr Lynch and Mr Egan on 2 September 2009, which was subject headed “key customer strategy” , as evidence of the nature and purpose of the hardware reselling strategy said to have been discussed and agreed at the Loudham Hall meeting in July 2009. The Claimants contended that the email exchange was pretextual: see paragraphs 811 to 822 below.

^{^[132]} Dr Lynch estimated that only about one percent of overall sales of hardware were to prospective rather than existing software customers.

^{^[133]} The Claimants accepted that Autonomy had always sold some hardware, and of course it had always had to purchase large amounts of hardware for its data centres and hosted and archiving businesses.

^{^[134]} The benefits cited by Mr Ariko, in the email to which Dr Lynch was replying, were “ incremental revenue and margin contributions ” .

^{^[135]} After a reconciliation, by August 2010, according to Dr Lynch.

^{^[136]} EMC’s content management and archiving division.

^{^[137]} He also emphasised , for example, that though he was “involved in making this type of hardware sale when Autonomy first began this practice in Q2 2009…[he] was involved only occasionally thereafter.”

^{^[138]} Mr Egan elaborated on this by reference to a large software licensing and data hosting transaction with Bank of America in which he was involved in Q1 2011: and I have quoted further from the same passage in his witness statement in examining that example at greater length in paragraph 1140 et seq below.

^{^[139]} According to Mr Sullivan’s evidence in the US criminal trial, this included what was known as the HULK, “a big giant storage system that was supposed to store stuff…very, very cheap” which EMC had just produced and Autonomy was thinking of trying out.

^{^[140]} Mr Hussain’s email to Mr Sullivan, 30 June 2009 stated: “ I am very happy with the negotiations that you guys have carried out. My understanding is that you have managed to strike a very aggressive pricing on the existing hardware refresh ($3m) having taken formal pricing from Hitachi (the only alternative) and also pre bought digital safe storage capacity at the same attractive rates for $6m (all including 3 year maintenance). This hardware will save significant costs as our digital safe business is increasing at a very rapid rate. ”

^{^[141]} For example, in an email of 25 June 2009 to Dr Menell Mr Wang stated: “ we had a breakthrough with EMC today in having them agree to sell us essentially hard drives without their fancy software which is irrelevant for Digital Safe... Given that we will be using EMC only for storage, we also convinced them to removed [sic] the IU server that was bundled in the unit. so what we're essentially left with a cabinet with 360 hard drives, the price of which should be substantially better than the standard ATMOS unit ”.

^{^[142]} On the basis that increasing Autonomy ’ s volumes of hardware further gave it serious negotiating power which could lower the price of its purchases. Dr Lynch gave unchallenged evidence that this led to a saving of $10m for Autonomy on hardware purchases - a saving equivalent to around half of the $21m of losses incurred (on the Claimants ’ case) on hardware transactions during the Relevant Period.

^{^[143]} It is to be noted also that there was nothing in the agreement that obliged EMC to use the money it received from Autonomy in any particular way. The potential significance of this is developed further later in the context of how the Defendants sought to justify the allocation of a substantial portion of the hardware costs as sales and marketing expenses.

[144] This was a point Dr Lynch was especially quick to emphasise when cross-examined on the EMC arrangements.

[145] As Mr Chamberlain noted in an email dated 25 September 2009: “We are not taking on any performance rights in respect of hardware, software, maintenance or services that remains with EMC. Bloomberg owe us the full amount and the only thing we have to do is deliver the equipment”.

^{^[146]} It appears from an email dated 17 October 2009 from Mr Robert Mark to Mr Hussain that Autonomy was pursuing two software deals with Citi in Q4 2009, one for the migration of Citi data from Iron Mountain to Zantaz and the other an EAS deal. The same email also suggests that a 23% hardware discount was to (in effect) be compensated by a higher maintenance fee.

^{^[147]} Which Mr Egan depicted as not only initiating a “strategic sale relationship into the Federal space where we have decided to invest heavily” but also (and illustrating the dual purpose) “to sell with SGI and work to add HW to our book of business and add SW to SGI in a lucrative quarter.”

^{^[148]} In his evidence in these proceedings, and in the US criminal proceedings, Mr Egan accepted that he had, and he accused others of having, concocted pretextual documents in the context of (a) reciprocal transactions to give apparent credence to Autonomy’s stated need for the product it was purchasing; and (b) a transaction with Capax to provide EDD services (as to which see paragraph 1963(2) of this judgment) to give the appearance that the services were actually being provided by Capax (though it is the Claimants’ case that they were not).

^{^[150]} This became clear by mid-July 2009 when, after approaching EMC, Mr Sullivan reported to Dr Lynch and Mr Hussain by email dated 15 July 2009 that he had already got verbal commitments for up to $20 million of hardware and could “probably get twice that if you want” (see paragraph 782 above also).

^{^[151]} I should clarify that although tempted to do so, I have not read that report. At a virtual hearing in February 2021, submissions were made to me on behalf of the Claimants that I should do so; that was opposed by the Defendants. I was not asked to state a decision and did not do so. However, I determined privately that, notwithstanding the eminence of its authors and its status more generally, I should not take account of what was and is opinion evidence, even if akin to expert evidence, which I should not accept collaterally nor weigh with or against the expert evidence admitted and tested in these proceedings.

^{^[152]} They may also have lost their way in their response to Mr Hogenson’s letter and its aftermath.

^{^[153]} However, a gloss on that is important, especially since I understood the Defendants occasionally to take the inference as extending to an expectation that all Autonomy ’ s correspondence would also be open for review. Such an extended inference would not be justified. There were various instances, of which the clearest was extended correspondence between Mr Sullivan and EMC (see paragraph 773 et seq below), in which it was clear that Deloitte had not seen or sought to see emails. The impression I have gathered is that, in terms of documentation, Deloitte confined itself to accounting records and neither they nor Autonomy expected Deloitte to be shown or seek out Autonomy ’ s internal or third party emails unless Autonomy chose to copy them.

^{^[154]} This was in a sequence of questions and answers, the first part of which is quoted in paragraph 839 above. In cross-examination in these proceedings he confirmed the truth of what he had said.

^{^[155]} The FY 2010 Report to the Audit Committee stated “ Under IFRS 8, additional entity-wide disclosures are prescribed that are required even when an entity has only one reportable segment. These include information about each product and service or groups of products and services ... As part of our audit of the notes to the financial statements, we shall review the disclosure made in relation to operating segments and we shall ensure that this meets the requirements of IFRS 8. ”

^{^[156]} Mr Welham confirmed that this process was carried out every quarter.

^{^[157]} Mr Welham confirmed that the tickmarks were prepared by Deloitte, and that the whole spreadsheet, although based on a document provided by Autonomy, was a Deloitte working paper.

^{^[158]} The description is as follows: “IDOL Product is normally delivered as licensed software paid for up-front with an ongoing support and maintenance stream. This model is becoming less significant with the rise of cloud computing. In 2010, IDOL Product revenue totalled $251 million”; Deloitte has ticked off the number $251 million.

^{^[159]} See e.g. Deloitte’s Report to the Audit Committee on the 2009 Audit: “These hardware sales did not include any IDOL software component.

^{^[160]} In cross-examination, Mr Welham confirmed that he remembered this conversation, and that to his knowledge, there never were any major disagreements with management regarding accounting policies or conclusions.

^{^[161]} In the course of his evidence to the FRC on 4 September 2013.

^{^[162]} (Largely comprised of email exchanges between the individuals at Autonomy principally involved and especially, in addition to the Defendants, Mr Kanter, Mr Chamberlain and Mr Sullivan).

^{^[163]} Of which the only detailed description was in Dr Lynch’s own witness statement: see paragraph 720 above.

^{^[164]} See the explanation of this in paragraph 65 above.

^{^[165]} Dr Lynch told me in cross-examination that he did not think that a fixed amount had been fixed at the Loudham meeting: “As I recall the conversation, it was more people offering how much they could do and what might be possible”. That answer itself points to the focus being on revenue maximisation rather than on promoting the software business, though of course the Defendants contended that what was good for one would be good for the other.

^{^[166]} Hence his reference to 4 or 5 Porsches instead of the one Dr Lynch had jokingly offered him: see paragraph 782 above.

^{^[167]} Dr Lynch accepted in cross-examination that Mr Hussain kept him abreast of the position in relation to these hardware deals.

^{^[168]} Though Dr Lynch understood (as did Mr Sullivan) that for legal purposes Autonomy did have the right to delivery.

^{^[169]} Dr Lynch told me in cross-examination that this figure was “the forecastable revenue that can be obtained, not the theoretical maximum” by which I took him to mean that not all the EMC deals would proceed. Even if that is so, the excess over the figures in contemplation at the Loudham Hall meeting is very considerable and in my view such as to invite reassessment of the alleged strategy.

^{^[171]} Informatica was a large NASDAQ quoted company, Coremetrics a small one; Autonomy was considering acquiring both.

^{^[172]} Dr Lynch sought to dismiss the argument that hardware sales were used as a “ stop-gap” on the basis that such sales were themselves built into and part of the monthly overall targets, and had to be achieved accordingly. That would be an answer if hardware revenues did not exceed the target set for them: but it soon was evident and known to Dr Lynch that it would be possible to obtain significantly more revenue from sales of EMC hardware than had been contemplated at Loudham Hall or budgeted for thereafter.

^{^[173]} Mr Brent Hogenson (then Autonomy’s CFO for the Americas) and Mr Michael Mooney (then Senior Vice Preside nt, Field Sales Operations at Autonomy).

^{^[175]} A chronological summary of Autonomy’s hardware reselling programme with Dell follows in section (6) starting at page 383 below.

^{^[176]} The Claimants submitted that this must be a reference to software provided by Dell on the basis that Autonomy would not need Dell’s permission to sell its own software: but it is clear from e.g. clause 8 that it included third-party software and I do not agree that it did not include Autonomy software.

^{^[178]} See also Yan 1 §28, which was unchallenged in cross-examination: “ Following Autonomy’s acquisition of Zantaz, the Product Development team was encouraged, and at times, pressurized, into trying to package Digital Safe into a self-contained appliance for sale to on-premise customers. This led to the development of what was known amongst the software developers as “Safe in a Box” (subsequently marketed as “Arcpliance”), which was essentially a computer box with the capacity to archive a limited amount of data ”.

^{^[179]} Only one VAR transaction was concluded (VT15); and that did not involve any of Autonomy’s usual VARs.

^{^[180]} The reference in Mr Hussain’s email to “bav” was to the Vatican Library.

^{^[181]} I address this at greater length later: see paragraph 1477 et seq below.

^{^[182]} It may be recalled (and see paragraph 939 above) that SHI International was a company that purchased computer hardware (typically from Dell) and resold it to BofA.

^{^[183]} That is to say, adjusted for conversion of loan notes, as Mr Hussain had alluded to in his email of 16 January 2011 to Dr Lynch (see paragraph 1065 above) when comparing Q1 2011 and Q1 2010.

^{^[184]} The Prisa deal was a VAR transaction entered into with DiscoverTech (as to which see paragraphs 591 to 707 of the Schedule of Impugned VAR Transactions) for which the licence fee payable was $3.6 million.

^{^[185]} Or ‘spiff’. Dr Lynch also made the point that Mr Sullivan was also rewarded for software sales.

^{^[186]} Dr Lynch did not elaborate what disputes he had in mind; but it seems to me that they could have included issues as to what time lag between one purchase and the other should be permitted, how pre-purchases of software in advance of hardware sales might be treated, or whether any account should be taken of customer retention by virtue of the other advantages to existing customers of dealing with a single supplier for all IT needs.

^{^[187]} Ms Eagan was one of those (with Mr Kanter and Dr Menell, and also Ms Orton) who joined Invoke Capital in May 2012 after leaving Autonomy. She was co-CEO (with Ms Gustafsson) of Darktrace prior to its IPO with particular responsibility for the marketing side of the business. She is now the Chief Strategy and AI Officer at Darktrace. She has always been based in the United States.

^{^[188]} Autonomy had resold, at a loss, more than $30 million of Dell hardware to SHI (as reseller) who, in turn, on-sold to Bank of America.

^{^[189]} A reference to the Interwoven acquisition (as also is the reference to “additional headcount” ).

^{^[190]} As will be seen, the way the sources of revenue were presented was changed in Q1 2010: broadly speaking, from the start of 2010, “ Product including hosted and OEM ” was reported as three separate categories: (i) IDOL Product, (ii) IDOL Cloud and (iii) IDOL OEM. The Claimants emphasised that whatever the breakdown, it always added up to 100% of Autonomy’s total reported revenues.

^{^[191]} Autonomy acquired Interwoven Inc earlier in 2009 for approximately $800 million.

^{^[192]} A different format for “Supplemental Metrics” was adopted for the Q1 2010 Quarterly Report and thereafter, dividing Autonomy’s business into IDOL Product; IDOL Cloud; Service revenues; Deferred revenue release (primarily maintenance); and OEM derived revenues.

^{^[193]} The Claimants alleged that what they presented as subterfuge was apparent in the Q&A scripts for the Earnings Call and in what was said at the Q1 2010 Earnings Call. Their case was that not only the market was misled: so too were Deloitte and the Audit Committee. Even with that explanation it triggered inquiry, as I elaborate on when dealing with the Earnings Call for that quarter.

^{^[194]} Mr Miles pointed out in his oral closing submissions that there is no evidence that any of the Claimants relied on either of the Investor Bulletins referred to by the Claimants, and they are not the basis of any claim (nor is any mention made of any such Bulletin in the RRAPoC). However, the Claimants relied on them as part of the evidence demonstrating how any hardware sales were depicted as “appliance” or “Arcpliance” sales, the fact of “pure hardware” was enveloped in that fog, and revenue from “pure hardware” sales of which recognition had been deferred tucked away as “Inventory” (and its source and nature was subsequently falsely described).

^{^[195]} The Claimants submitted that given the extensive inventory-related discussions on the Q1 and Q2 2010 earnings calls, a reader of the Bulletin would have understood this all to be appliance-related.

^{^[196]} The “ Core IDOL reported revenues ” ($144.3 million) appearing in the table at the top of page 3 of the Report was - as stated in footnote 1 to the table - the sum of the IDOL Product ($54.4 million), IDOL Cloud ($52.7 million) and IDOL OEM ($37.2 million) categories.

^{^[197]} I was not referred to any Investor Relations Bulletin for this period.

^{^[198]} The notion of a “standard reseller margin” of 55% was contradicted by Autonomy’s references only days before to an allocation based on a 25% and then a 35% margin.

^{^[199]} Shortly after receiving that email, Mr Chamberlain asked Mr Hussain whether he could forward on Deloitte’s email to Mr Sullivan. The Claimants drew my attention to Mr Hussain’s response, in which he stated that Mr Sullivan could be emailed in relation to the first point only, namely obtaining evidence from EMC quantifying the hardware amount. As for the “other points” (i.e. Deloitte’s requests relating to the asserted marketing programme and appliance development initiative), Mr Hussain refused Mr Chamberlain’s request, describing these requests as “more management stuff”.

The Claimants contended that Mr Hussain’s refusal must be seen against the backdrop of (a) Mr Sullivan having already made clear to Mr Hussain and Mr Chamberlain that none of these matters had been discussed with EMC and that EMC had not committed to anything, and (b) Deloitte having expressly suggested that when providing the clearer explanation of the rationale for the strategy they required, Autonomy would need to involve “those who negotiated the deal”, i.e. Mr Sullivan. They submitted that the reality is that Mr Hussain did not want Mr Sullivan to see how he (Mr Hussain) had described the arrangements with EMC to Deloitte, because Mr Sullivan would know that the description was untrue. I accept this.

^{^[200]} Mr Knights’ annotations on the draft he prepared seem to confirm an earlier discussion with Dr Lynch and Mr Hussain: see paragraph 1314 below.

^{^[201]} Clarion (a mid-range product) and DMX (a high-range product often called Symmetric) were different EMC product lines. According to an earlier email from Mr Mussulli to Mr Sullivan, the breakdown as regards the three end-purchasers in the Q3 2009 transactions was: JPMC = 100% DMX; Bloomberg = 100% Clarion; Citi = 35% Clarion, 65% DMX.

^{^[202]} Mr Goodfellow was, in my view, a partisan witness and sometimes a rather too strident advocate of the Claimants’ case.

^{^[203]} The only documentary evidence actually produced was an email dated 11 September 2009 from Mr Egan to Dr Lynch, copied to Mr Hussain, with subject heading “JPMC NPower thing”, recommending $35k level sponsorship of a JPMC evening, and stating that it would be “an evening honouring Joe [T]ucci [CEO of EMC] and all the EMC deal people who are doing this quarter deal will be there. So will all the JPMC upper management of IT…” Dr Lynch’s response was “ok”. In fact, the event was not a joint marketing event with EMC, but an event organised by JPMC to which Autonomy agreed to make a charitable ‘sponsorship’ donation in a relatively small sum which Mr Egan (in the same email) said would be “deductible at some level” .

^{^[204]} “This looks like a great program and we are excited to participate in it”.

^{^[205]} which also had the effect of minimising the importance of justifying the various suggested categories of alleged allocation of the ‘premium’.

^{^[206]} According to Mr Welham’s witness statement, the PSR for H1 2009 was Ms Joanna Hacking, but Ms Bennett was the PSR for the remainder of 2009. The PSR in Q1 2010 was Ms Lisanne Fitzgerald; during the period Q2 2010 to Q2 2011 it was Mr Garrie Lumb.

^{^[207]} According to Mr Welham’s witness statement, in Q1 2010, the EQAR was Mr Chris Brough, in H1 2010 and Q3 2010, Mr Chris Robertson, and thereafter until Q2 2011 Mr Jonathan Dodsworth.

^{^[208]} According to Mr Welham’s witness statement, Mr Brough was the IRP for the Q1 to Q3 2010 reviews, followed by Mr Stuart Barnett until Q2 2011.

^{^[209]} Mr Kanter produced a mock-up of a table of revenue sources which did not in any way differentiate software revenue from hardware sales as such. Instead, and in contrast to a proposed differentiation from “ Interwoven-related revenues” , it seemed to envisage hardware revenues being lumped into “ Core Autonomy IDOL revenues ” and included in the calculation a figure to be given for “ IDOL Organic Growth” .

^{^[210]} Which he explained as follows:

“In principle if these are to be included in the press release Deloitte have a responsibility to ensure that they are not inconsistent with our understanding of the numbers.

We do however need to ensure that the information is consistent with the approach applied in putting together the financial statements and does not invalidate the segmental or revenue analysis arguments that have previously been put forward.”

^{^[211]} His attendance is not recorded in the Minutes for the Audit Committee meeting to consider the Q3 results held on 16 October 2009. The three members of the Committee were present. Messrs Hussain, Kanter (Secretary), and Chamberlain of Autonomy, and Messrs Knights, Knight, Welham and Ferguson of Deloitte were recorded as having been in attendance.

^{^[212]} Dr Lynch told me in cross-examination that he did not know this.

^{^[213]} Dr Lynch denied this, relying on the fact that he was in California (marooned there by the Icelandic volcano eruption which restricted international flights), and suggesting that “ I don ’ t think these would be my changes. The words may be my changes but there ’ s certainly something else going on here as well” . However, it is clear, in my view, that Dr Lynch did make the changes. The changes were attached to an email from Dr Lynch to Mr Hussain in response to a request for comments, the message in that email being “ some suggested edits you may want to check for accuracy”.

^{^[214]} The discomfort within Deloitte was expressed by the PSR on the Autonomy account for Q1 2010, Ms Lisanne Fitzgerald, whose comments on 18 April 2010 on the draft Q1 2010 report to the Audit Committee were summarised to Mr Chamberlain by Mr Welham; she stated:

“I feel pretty uncomfortable about this - it seems that the magnitude of these hardware sales will grow and have been occurring over a period of months - how long can they really be deemed to be marketing and not cost of sales. It strikes me that this is a way of them preserving gross margin which I am not sure is right…”.

^{^[215]} Mr Welham was not challenged on his witness statement in this regard. The point was of some importance to the segmental analysis; and in Mr Chamberlain’s report sent to Deloitte on 18 January 2011 it was emphasised, entirely falsely, that “No information is provided to MRL on hardware revenues”.

^{^[216]} Mr Bloomer confirmed that (a) he and his committee had relied on what they were told by Mr Hussain and Deloitte (b) they understood the linkage to show that hardware sales for a particular customer had driven subsequent software sales to that customer (though he subsequently said that “…my understanding at the time was that there were a number of major customers who bought both hardware and software and that selling the hardware was an aid to selling the software….But there was no - in my mind at the time - there was no sort of chicken and egg debate about which came first. There was a series of big customers who bought both hardware and software.”)

^{^[217]} Shortly before written closing submissions were submitted, Dr Lynch uploaded an email that Mr Egan appears to have sent to Mr Hussain on 3 November 2010 concerning a Morgan Stanley deal. In response to Mr Hussain’s query (“Software and hardware together?”), Mr Egan responded: “Separate but proposed in concert as it helps Morgan get arms around discounted hw as part of larger relations” . In fact, Autonomy’s own linkage analysis does not suggest that Morgan Stanley entered into a software deal in Q4 2010 that was ‘linked’ to any hardware deal (see column U, row 12). In any event, as is plain from the way in which the hardware sales were generated and concluded, Morgan Stanley’s ability to get its “arms around discounted hw” was not in any way dependent upon or linked to its software purchases from Autonomy.

^{^[218]} The issue of segmental analysis was an abiding and important one. It may be recalled that Autonomy management’s view was that the group had just one operating segment, being the sale of IDOL software. The importance of that is that if a company has more than one operating segment, each segment must be differentiated in the accounts. The importance of this is clear: segmental accounting would have destroyed the achievement of the purposes of the hardware reselling strategy. By the same token, however, it has seemed to me that it was Deloitte’s approval of the view that Autonomy had a single operating segment which caused Mr Bloomer and the Audit Committee not to query the hardware sales further and to accept the allocation of costs from what was in reality a different business to the software business, without any differentiation or warning note.

^{^[219]} Mr Hussain sent Dr Lynch a draft for comment at 13:11 on 16 April 2011: that draft did not include the underlined words. Mr Hussain sent Dr Lynch a further draft half an hour later, including the underlined words; the covering email stated “ Please use this draft for review - updated with frrp and shareholder letters comments ”. This suggests that Dr Lynch had not yet reviewed the earlier draft, and also that the changes between the two drafts were not proposed by him.

^{^[220]} Dr Lynch was aware that a spreadsheet was prepared each quarter that showed the software bought by customers who bought hardware. However, he did not see the spreadsheets at the time.

^{^[222]} About 100 pages were devoted to it in the Claimants’ written closing submissions. Dr Lynch dealt with it in about 30 pages.

^{^[223]} In summary, on the Q3 2009 earnings call, combined with the accompanying presentation, Dr Lynch attributed the SPE marketing spend figures to five steps:

(1) Advertising in the trades;

(2) Discussions with analysts;

(3) A Quick Start initiative;

(4) A Beta programme-which was already underway and released to several hundred Beta customers, and

(5) Specialist Seminars.

^{^[224]} For accuracy: of the $11.7 million, $7.3 million was said to be referable to SPE.

^{^[225]} 75% was permitted by HMRC as qualifying ‘technical’ work for tax relief purposes; but as Ms Gustafsson agreed the further implicit suggestion that 75% was an accurate estimate of the time spent on SPE was obviously wrong. The notion that all the SEs had spent all their technical time on SPE was absurd: it would have crippled the Autonomy group, as Mr Lucini pointed out.

^{^[226]} Mr Hussain’s submissions barely addressed this aspect of the matter; but I have taken him to support and adopt Dr Lynch’s submissions.

^{^[227]} Mr Lucini was cross-examined in relation to this: he clarified that all he was seeking to say at that interview in 2012, and certainly as he saw the position now, was that he would not have known without finding out from Ms Harris.

^{^[228]} The cross-examiner did not clarify in every instance whether the individual was working on SPE.

^{^[229]} Often referred to in the documents as “CLSP” or “CL+SP”.

^{^[230]} The SPE page on Autonomy’s website included a link for registration.

^{^[231]} The transcript of the Earnings Call refers to the poor line quality.

^{^[232]} I address the scripts in greater detail when considering the issue as to “guilty knowledge”.

^{^[233]} For comprehensiveness it may be noted, as Mr Holgate did note, that the Q1 2010 Annual Report did (fleetingly) refer to hardware (at p51) in “ Costs of revenues: Costs of license revenues includes the cost of royalties due to third party licenses, costs of product media, product duplication, hardware and manuals”. But I do not think it is realistic to read that as a disclosure of volume hardware reselling.

^{^[234]} Mr Apotheker gave the examples of TIBCO and Software AG when cross-examined in these proceedings.

^{^[235]} In reply to a pre-action letter of claim.

^{^[236]} The description is as follows: “IDOL Product is normally delivered as licensed software paid for up-front with an ongoing support and maintenance stream. This model is becoming less significant with the rise of cloud computing. In 2010, IDOL Product revenue totalled $251 million”; Deloitte has ticked off the number $251 million.

^{^[237]} See e.g. Deloitte’s Report to the Audit Committee on the 2009 Audit: “These hardware sales did not include any IDOL software component”.

^{^[238]} Further, the Defendants submitted that Mr Briest’s analysis is plainly not “published information” of Autonomy; and nor is it suggested that his report is a representation made by Dr Lynch.

^{^[239]} Originally, the Claimants contended that Autonomy also had a duty of disclosure in its Quarterly Reports; but in its written closing submissions it was clarified that, in light of Mr Holgate’s evidence that “the case for disclosure of hardware sales in quarterly reports is not as strong as in interim reports” , they were no longer pursuing that contention, although Mr Holgate’s view was that it would have been “good practice” for Autonomy to have disclosed its hardware sales in these reports.

^{^[240]} which is not the same as commercially sensible or advisable.

^{^[241]} In the RRAPoC, the Claimants also relied on various provisions of the Companies Act 2006, the Combined Code on Corporate Governance and the Disclosure and Transparency Rules, but as explained below, it appears that these points are no longer pursued.

^{^[242]} It should be noted that the FRRP’s comments were in relation to a company’s general disclosure requirements. They were not directed at the hardware issue, which was not on their radar at the dates of those letters.

^{^[243]} See Opinion of Martin Moore QC, ‘The True and Fair Requirement Revisited’ dated 21 April 2008 and published by the FRC on its website.

^{^[244]} The IASB noted that “the empirical evidence identified in the academic review shows that the number of reported segments has increased and the number of single-segment entities has decreased.”

^{^[245]} Mr MacGregor accepted that if the hardware sales were a separate sale of a different product to the sale of Autonomy ’ s core IDOL product, then they would have to be disclosed unless hardware sales could reasonably be considered by management to be ‘ incidental ’ to the sales of the core product, with their principal purpose being to ‘ drive’ or ‘ facilitate’ software sales. Mr MacGregor appeared to accept, therefore, that if Dr Lynch ’ s case that Autonomy entered into the hardware sales to drive software sales is not accepted, the hardware sales should have been disclosed.

^{^[246]} Against this, I have reason to recall that the Court of Appeal has stated recently that “…in commercial cases, there will be a wide spectrum of probabilities as to the occurrence of reprehensible conduct”, and once a propensity for dishonesty in and around the same matter has been demonstrated “ it is faintly absurd to elevate the principle that it is inherently improbable that a party would do something dishonest into a relevant benchmark for the determination of the issues”: see Bank St Petersburg PJSC & Anor v Arkhangelsky & Anor [2020] EWCA Civ 408 at [47].

^{^[247]} The same argument applies in relation to omissions: Sch 10A §3(3).

^{^[248]} These were direct sales by Autonomy to UBS (VT28 and VT34) after “ dummy” VAR sales to Capax Discovery addressed in the chapter of this judgment on the impugned VAR transactions. Looking ahead, it will be seen that I have concluded that “ VT28 and VT34 were exemplars of the pattern. The sale to the VARs were illusory and effected no change in control or transfer of risk; no revenue should have been recognised in respect of them.”

^{^[249]} Dated 18 August 2011 a sking whether he should proceed with the Dell hardware sales.

^{^[250]} I should note that in HP’s written closing submissions it was asserted that “ There is no evidence of the EY memorandum [of a review of Deloitte’s working papers] ever having been sent to anyone at HP”. That may be so: but the email referred to shows that the only “surprise” (hardware sales in the amount of $100 million pa) was reported to HP.

^{^[251]} Mr Welham did not recollect the detail of the discussion, but said it is likely that he did discuss these matters with E&Y.

^{^[252]} Jamie Boggs, was described by Mr Gersh as his colleague and “a director [of KPMG] who was heavily involved in both the pre- and post-acquisition engagements for HP related to the Autonomy acquisition”.

^{^[253]} Senior Director, Accounting Policies and M&A Reporting in HP’s EFR team.

^{^[254]} Sales to VARs were described as “Autonomy’s primary revenue channel” in its Annual Reports for both 2009 and 2010. It is of some note that the resellers listed in such Annual Reports were big name companies such as Accenture, IBM Global Services, Cap Gemini, Wipro and HP. MicroLink LLC (“MicroLink”) was mentioned in the context of its purchase by Autonomy in the first half of 2010 but there was no other mention of any of the ‘friendly’ VARs with whom Autonomy conducted the impugned VAR sales.

[255] The Defendants have invited me to re-consider Ms Harris’s evidence in this regard. She described what was done as an “exercise, pre-acquisition, of trying to work out how HP would want to deal with provisions and bad debts under HP’s own policies” and insisted that there “was nothing improper about any of this”. She explained also that after the acquisition she and her team “learned that HP was more conservative than we had predicted, so we had to make more provisions.” She emphasised that a note recording an interview she gave PwC in November 2012 (which she did not see until August/September 2019) had been “spun” and she sought to shrug off as “muddled at this point” a description in that note of her having said, as regards the pre- and post-acquisition write-offs, that she “would have said something earlier if I had been allowed to talk about it or correct it, but Hussain prevented me” and that she “eventually presented little lists of these practices to come clean”. Her evidence at trial on which the Defendants relied was that “It was a housekeeping exercise. There was nothing to “come clean” about and I was not prevented from discussing what we had done with HP.” I shall return to assess this evidence later: see paragraphs [2153] et seq.

^{^[256]} All but two of the 30 were transactions with ‘friendly’ VARs. Only two (VT15 and VT26) were with other VARs (Realise Limited and Tikit respectively).

^{^[257]} Snook v London and West Riding Investments [1967] 2 QB 786. A ‘sham’ in a legal sense is a contractual arrangement dressed up and misdescribed as having one legal effect and in reality having another legal effect; as Diplock LJ (as he then was) put it at p802C-E: “…it is, I think, necessary to consider what, if any, legal concept is involved in the use of this popular and pejorative word. I apprehend that, if it has any meaning in law, it means acts done or documents executed by the parties to the “sham” which are intended by them to give to third parties or to the court the appearance of creating between the parties legal rights and obligations different from the actual legal rights and obligations (if any) which the parties intend to create….for acts or documents to be a “sham”…all the parties thereto must have a common intention that the acts or documents are not to create the legal rights and obligations which they give the appearance of creating.”

^{^[258]} Or were reckless as to whether the relevant presentation was untrue or misleading.

^{^[259]} As explained below, the difficulties in its application were recognised and IAS 18.14 has since been superseded by IFRS 15.

^{^[260]} Usually by FTP or on Autonomy’s Automater system, which was the platform through which Autonomy made software available for download to customers who had licensed it.

^{^[261]} In the case of DiscoverTech, each reseller transaction was governed by an individual VAR agreement. In the case of the other ‘friendly’ VARs there was a master agreement governing all their respective reseller transactions.

^{^[262]} If the criteria are satisfied, recognition is mandatory.

^{^[263]} In Dr Lynch’s closings and in oral submissions this was less absolute: it was contended that the terms of the contract are highly important in determining the substance, contrasting a simple sale of goods contract from a lease finance where in substance the lessee owns that which is leased.

^{^[264]} Mr Holgate accepted that he did not have any prior experience in relation to the reseller industry for information technology in the US.

^{^[265]} Mr Goodfellow’s evidence was that Autonomy delivered the storage cells directly to Citi and, furthermore, delivered a number of them in advance of the direct agreement being signed.

^{^[266]} The Claimants’ case being that “the only money ever paid by MicroLink in respect of the 11 [impugned VAR] transactions was money that Autonomy had channelled to MicroLink for that purpose.” These alleged ‘reciprocal’ VAR transactions included the purchase by Autonomy of MicroLink’s ‘Search Analysis Tool’ (“SAT”) and AIS software for the sum of $9.3 million, for which the Claimants alleged Autonomy had no need.

[267] The Defendants have referred me to evidence that, prior to being acquired by Autonomy (on 4 January 2010), (a) MicroLink was a large reseller of Autonomy products and indeed was its “Global Partner of the Year” in 2007; (b) MicroLink’s Q3 2009 purchase (VT1) was an addendum to an earlier software licence agreement it had entered into with Autonomy in March 2007; and (c) MicroLink’s extensive work with both Autonomy and Microsoft products and customers made it well placed to develop its own software tools to enhance Microsoft’s Sharepoint customers’ use of Autonomy products using its AIS (Autonomy Integration Suite for SharePoint) product. That seems to me, however, to make more remarkable the fact that in the impugned VAR transactions MicroLink played a ‘placeholder’ role.

^{^[268]} By 2013 Capax Discovery was making some $15 million per year in revenue from contracts with HP and it wanted to be (and at about that time in fact became) a preferred HP partner.

^{^[269]} Mr Szukalski now works for Darktrace, and reports to Ms Nicole Eagan, who had previously (and whilst Mr Szukalski was there) headed Autonomy ’ s marketing department: he started there at the end of August or early September 2014.

^{^[270]} MicroLink provided a similar, in fact more stark, example of a VAR not recording its ‘indebtedness’ in its accounts as a liability; according to evidence in the US criminal trial, its larger deals with Autonomy were “just not in the books at all”.

^{^[271]} There is no evidence of any express representation of authority, perhaps because such a representation would not have fitted with the essentially non-binding nature of the assurances as a matter of law, and his repeated emphasis that nothing he said would or could modify the contractual terms set out in the relevant VAR agreement.

^{^[272]} The only apparent exceptions in evidence were, on analysis, not exceptions at all. There were one or two sales to which Autonomy was not party; but that was only because Autonomy, having negotiated the sale, directed that a VAR (but not the VAR to which it had initially sold the software) be inserted. Thus, in VT20 Capax Discovery/DKO the direct sale was negotiated between Autonomy and DKO, without any involvement by Capax Discovery; at the last minute, Autonomy arranged for MicroTech (not Capax Discovery) to step in so that the agreement took the form of a sale from MicroTech (rather than Autonomy) to DKO. VT1 may provide a similar example: RRAPoC Schedule 3 VT1 states that in Q4 2009 MicroLink sold the software it had acquired for the intended end-user US Federal Government to a second VAR called Computer Security Solutions. But that too seems to have followed negotiations exclusively by Autonomy with the end-user US Federal Government and to have been at the direction of Autonomy. I have found no other ‘exceptions’ in the evidence.

[273] The Defendants made the point that the impugned VAR deals were selected by the Claimants as showing this characteristic and were not a representative sample of the universe of VAR deals entered into by Autonomy. That is so: but that does not address the point that the selection included the large end of quarter transactions entered into to generate revenue to make good shortfalls from other sources.

[274] Though the Defendants made the same point as I have recorded in footnote 274 to paragraph 2116 above.

[275] See also my discussion of VT4 in the Schedule of Impugned VAR transactions attached to this judgment.

[276] As noted by the Claimants in their written closing submissions, she fell back on saying that the provisions in issue had been made in Q3 2011, and included in the $45.6 million of bad debt expenses recorded for the quarter. As she admitted, where provision is made in the quarter that does result in a charge to the P&L.

[277] The exact quantification of the amounts written off or the subject of credit notes in relation to VAR transactions was disputed. The Claimants gave a figure of $130 million for the total amount of the revenue “clean up” but accepted that only a proportion related to impugned VAR transactions. The Defendants appeared to accept that write-offs relating to impugned VAR transactions exceeded $15 million, and credit notes issued in respect of VARs impugned by reference to a ‘side agreement’ exceeded $18 million (in addition to credit notes of some $23.2 million in respect of transactions where Autonomy had cut out the VAR, dealt directly with the end user and was compensating the VAR).

^{^[278]} Mr Rabinowitz made the point in his reply that in only three of the 37 impugned VAR transactions did the VAR pay the licence fee by the due date, and in each of those three cases (VT5, VT23 and VT37) Autonomy had put the VAR in funds.

^{^[279]} Mr Szukalski told me in cross-examination that he had never wanted to do the deal because reselling was not part of FileTek’s business at all, and he confirmed that FileTek had never resold Autonomy software. He had wanted Autonomy to use another company owned by the same person as FileTek called Centennial; but Autonomy had insisted on FileTek.

^{^[280]} Dr Lynch’s written closing submissions stated that the evidence in respect of the FileTek transaction for end-user USDVA “of all the reseller transactions...in fact takes the Claimants’ case the furthest”. That was a fairly transparent ‘Aunt Sally’.

^{^[281]} “At risk deals” were (in Mr Egan’s terminology) sales to the VAR where the VAR had not yet agreed a sale with the end-user. Mr Egan explained in his witness statement that prior to 2008/9 Autonomy had frequently sold its software through resellers, but none was an “at risk deal” : Autonomy had always required the VAR “to first obtain a purchase commitment from its customer, the end-user. Then, and only then, would we accept an order from the VAR for the relevant software.” The Defendants’ case was that Mr Egan had confused the chronology and that Autonomy had routinely undertaken “at risk deals” once it had changed from US GAAP to IFRS accounting which permitted revenue recognition before “sell-through”.

^{^[282]} Deloitte came to learn in the first half of 2010 of a small number of transactions that had been concluded between Autonomy and three VARs, which were later replaced by a direct deal between Autonomy and the end-users. The transactions where Deloitte knew that this had occurred were VT3 (Capax Discovery/Kraft), VT7 (MicroTech/Manulife), VT8 (MicroTech/Morgan Stanley), VT4 (Capax Discovery/Eli Lilly) and VT12 (DiscoverTech/PMI).

[283] It is fair to point out, as the Defendants have reminded me was the case, that as at 7 August 2010 there was no VAR yet involved in this deal: and see my discussion of VT18 in the Schedule of Impugned VAR transactions. The point, however, is that made in the last sentence of paragraph [2195].

^{^[284]} The “whistleblower” who had been Autonomy Inc’s Chief Operating Officer and “came forward” in May 2012.

^{^[285]} Further undermining any argument that the relationship between the contracting parties was exclusively defined by the contract.

[286] (which was not referred to in any closing submissions but which was drawn to my attention by the Defendants in commenting on an earlier draft of this judgment).

^{^[287]} The transcript repeatedly includes the word “umm” in the above passage. This has been deleted in the quotation because it hinders the plain meaning of what was said.

^{^[288]} Which started with a pun “hope Nice is nice”.

[289] It appears from an email dated 22 July 2010 from Mr Kanter to Mr Webb QC’s PA (Ms Sandra Daley), which had not been referred to at trial but to which the Defendants referred me as part of their comments on an earlier draft of this judgment, that what was described as the “Brent correspondence.piz” was sent to Ms Daley at the request of Mr Webb in two “zipped files of all correspondence” apparently attached to the email. The attached zipped files were not in evidence and it is unclear what they comprised. Mr Kanter might have been able to tell me: but he declined to appear and his witness statement was withdrawn. Mr Webb’s evidence in his witness statement was that he “did not deal substantively with the issues raised by Mr Hogenson nor the response to these issues by the Audit Committee and auditors.”

^{^[290]} This email was the first time that Mr Hogenson had taken his concerns outside the four walls of the company and its advisers.

^{^[291]} Then comprised of Messrs McMonigall, Perle and Ariko (of whom none was an accountant). Mr Bloomer had not yet become its chairman.

^{^[292]} That email also explained that the FSA only regulated financial services firms and so this was a matter outside the FSA’s jurisdiction, but that “in the interests of open communication we will proactively work with them should this be of interest”.

^{^[293]} Mr Welham ’ s evidence in his witness statement was “ The impression conveyed to us by Autonomy ’ s management was that Mr Hogenson had raised the accounting question in order to distract attention from his own responsibility for the payroll fraud.”

^{^[294]} By email dated 14 February 2011, Mr Welham asked Mr Hussain, Mr Chamberlain and Mr Kanter, “ Can we see the letter from 30 July 2010 ”. On 16 February 2011, Mr Welham emailed Mr Kanter again asking for “ a copy of the letter sent to the FSA in July 2010 ”.

^{^[295]} The Financial Reporting Review Panel had responsibility for oversight of public and large private company accounts and directors’ reports, and periodic reports by issuers of listed securities.

^{^[296]} In Microlink/Ameriprise (VT1), Capax Discovery/TXU (VT2), Capax Discovery/Kraft (VT3), Capax Discovery/Eli Lilly (VT4), MicroTech/ManuLife (VT7), MicroTech/Morgan Stanley (VT8), Capax Discovery/FSA (VT10), DiscoverTech/Citi (VT11), DiscoverTech/PMI (VT12), Capax Discovery/Amgen (VT16), Capax Discovery/Merrill Lynch (VT21), DiscoverTech /BofA (VT23) and DiscoverTech/BofA (VT24).

^{^[297]} There were two cautionary notes in Ms Sharp’s response: (a) “not all of the Panel’s questions have been answered” and (b) “there are a few dangerous points” ; and she added “We need to close down the enquiry now before it gets more serious”. But she was not a witness; and the parties did not make anything of these.

	[Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback]
	England and Wales High Court (Chancery Division) Decisions
You are here: BAILII >> Databases >> England and Wales High Court (Chancery Division) Decisions >> ACL Netherlands BV & Ors v Lynch & Anor [2022] EWHC 1178 (Ch) (17 May 2022) URL: http://www.bailii.org/ew/cases/EWHC/Ch/2022/1178.html Cite as: [2022] EWHC 1178 (Ch)

England and Wales High Court (Chancery Division) Decisions

External lawyers

Annual and quarterly reports: Deloitte ’ s review work

Deloitte’s review of annual reports

(i) Deloitte’s review of quarterly reports

(ii) Deloitte’s review of revenue

(iii) Press releases and earnings calls

(iv) The relevance of Deloitte’s review and approval

Revenue Reporting

IFRS 8: Single operating segment

Due diligence timetable

Technical due diligence

Top 40 customer contracts

Price negotiations in August

Incorporation of Bidco

HP ’s final approval: boardroom spats and second thoughts

18 August meeting

The announcement of the bid and HP’s loss of nerve

(1) First, “ whether there is any way to get out of the Autonomy deal ”.

(2) Secondly, what size of buyback HP could launch to try to recreate lost value for shareholders.

August 2012 - the decision at that time not to impair the carrying value of Autonomy

(a) Overview and background to the provisions of FSMA

(b) Conditions for liability under s.90A and Sch 10A FSMA

(c) Published information

(d) Untrue or misleading statement or omission

(e) Guilty knowledge

(f) Reliance

(g) Loss in the context of FSMA claims

Representation made to a claimant by a defendant

Defendant’s state of mind

Reliance / inducement

Loss

Misrepresentation Act claims

“ HDS will be about $2.1m – final orders coming in this week – I need to calc the final number based on final margin agreement with Morgan. Morgan is not willing to place the additional $4m order this qtr.

- should have another $500k to $1m from Dell/SHI. discussing some other orders as well – we could provide extra incentive to accelerate this.

- also working on a deal to sell DB a dedicated safe. this could be about $300k to $400k and could be delivered this qtr. There is no rep involved as Dan Manners came directly to us (Roger, Rob and I). I am in Pleasanton and we are trying to cut a deal with him ASAP.

other longer shots for Q1

- Will get orders from UBS in March for up to $1m, but unlikely to ship. Calls later today on this should provide more info.

- I am meeting Bank of NY in NYC on Thursday re: potential dell orders.

- We will probably get an order from Citi through a reseller called insight in March. Around $1m. Shipping will likely make this a q2 deal. ”

“ We have had to make further changes to your numbers.

Capax (FSA) – back in

Hardware - have had to recognize more HW

The revised hardware sheet is attached again with a column identifying the changes. We have manually updated the consol packs and TBs and processed the journals so you should tie in to these revised numbers.”

“No - you do not give this info to mike S - why would he need it? just get what he has sold on a daily basis - create a spreadsheet - I do not want any inventory at the end of this q”.

“ The report poses the question:

‘ What is the go-forward plan for product rationalization, both within the Autonomy portfolio and between the Autonomy and HP portfolios? ’

“ blind drop-ship from Dell directly to Target. So there should be nothing that says Autonomy on any packing slip ”.

Pre-acquisition knowledge and due diligence

Post-acquisition knowledge and discussion of hardware

(2) The pattern as regards the principal ‘friendly’ VARs in the Relevant Period is also striking and would surely have measured with Dr Lynch:

(3) The impugned VAR transactions accounted for 88% by value of Autonomy ’ s VAR transactions above $1 million in the Relevant Period generally.

(2) Knew of and authorised the further steps taken by Mr Hussain in Q3 2011 to make the VARs whole prior to HP’s takeover of Autonomy by means of very considerable write-offs and credit notes to forgive or set off their indebtedness.

(1) The first is that Dr Lynch accepted in cross-examination that he knew that the correct question to ask was whether Autonomy had transferred to the buyer the significant risks and rewards of ownership of the goods.

(4) Dr Lynch accepted in cross-examination that he knew that there was “some sort of managerial control test to do with having control of the goods” and “that the customer had to have control of the goods”.