A guide to today’s responses to HP’s UK Civil Court action

Today we acknowledge receipt of HP’s civil claim against the former management of Autonomy.

As readers will see from our comments below, we utterly refute the allegations made against us. HP has waged a three-year smear campaign riddled with half-truths and obfuscation. They have intentionally made the claims as complex and convoluted as possible. This is why, in the interests of transparency, we have posted a series of updates to this blog today so that everyone can see the allegations for themselves and fully understand our position.

There are three separate sections, which we set out below, as well as a downloadable PDF version of the letter from our lawyers Clifford Chance.

1. Lawyer’s letter containing preliminary detailed response to HP allegation

2. Our summary of the main allegations being levelled by HP

3. A reaction from Mike Lynch to this latest development in HP’s ongoing campaign against us

4. Downloadable PDF of lawyer’s letter (item 1 above)

The links above will take you directly to each specific post, or you can get to them by scrolling down the page.



In re Hewlett-Packard Company Shareholder Derivative Litigation

Dear Mr. Wolinsky:

As you know, we, along with Steptoe & Johnson, represent Dr. Michael Lynch.  Hewlett-Packard Company’s (“HP“) decision to file in the referenced action (the “California Action“) the Particulars of Claim (the “Particulars“) filed against our client by various HP subsidiaries in the United Kingdom (the “UK Action“), a proceeding to which, quite notably, HP is not a party, is a continuation of HP’s transparent effort to generate one-sided publicity for its specious claims and false statements, avoid disclosure and engagement on the merits, bury HP’s own malfeasance, and insulate its directors and officers from liability.

We write to you now to make plain—as we have made clear to HP on several previous occasions—that the claims in the Particulars are without basis.  While Dr. Lynch looks forward to unmasking the falsity and hypocrisy of these allegations in the courts of England (where, if at all, the matter belongs), in the interest of full disclosure, we request that, once the Particulars are unsealed, you amend your filing in California to include this letter, which sets forth below Dr. Lynch’s preliminary responses to the Particulars.  In this way, the court, parties, and the public can have a fair understanding that the allegations in the Particulars are without basis and emphatically denied and that this matter will be fiercely contested in the courts of England.[1]


HP’s (notably unparticular) Particulars make clear that HP has had very good reason for its two and a half years of stalling, misdirection, and evasion regarding the details of its allegations against our client and other members of Autonomy Corporation plc’s (“Autonomy“) former management.  Simply put, after two and a half years of apparent investigation—no doubt at a cost of many tens of millions of dollars to the shareholders of HP—it is clear on the face of HP’s own filings that its claims are baseless.

HP’s patchwork tale of alleged misconduct rests on a faulty foundation of false facts, unsupported inferences, and a misunderstanding and misapplication of the relevant legal and accounting standards.  Nearly two and a half years after HP announced its write-down, it is clearer than ever before that HP’s claims are merely a tactic to obscure the true source of HP’s and Autonomy’s losses:  the wrongdoing and ineptitude of HP’s own directors and officers.

That HP’s claims are without merit is plain from the preliminary response set forth below.  But several points omitted from HP’s claims bear noting at the outset.  In particular, the evidence shows that at all times:

  • Autonomy was in full compliance with the law and applicable accounting standards;
  • Autonomy employed a professional and experienced finance team that paid careful attention to ensuring that its accounts were properly stated;
  • Autonomy was transparent with its deeply experienced and careful professional auditors team from Deloitte, who audited the precise issues HP appears to contest, and who, to this day, stand by the accounts;
  • Autonomy had a dedicated, active, and experienced Audit Committee who met regularly with Deloitte to review Autonomy’s accounts and relevant policies; and
  • Dr. Lynch, who is not an accountant, had little involvement in sales and in making the corresponding accounting decisions related thereto.

Accordingly, it should not be surprising that there is not one shred of actual evidence establishing any pre-acquisition misconduct by anyone at Autonomy, let alone evidence of fraud.  There are no documents or witnesses—and HP has pointed to none—that demonstrate that any former member of Autonomy’s management acted with anything but honest intentions, good faith, and reliance on the professionals described above.  This is hardly the stuff of a legendary fraud.  In fact, in 2010 the Financial Reporting Review Panel (the “FRRP“) of the Financial Reporting Council, the regulatory body in England responsible for regulating accounts and accountants, reviewed many of the same transactions at issue here and found no basis for a continuing inquiry—as did Autonomy’s Audit Committee and its external auditors at Deloitte.

Despite the lack of any credible evidence of wrongdoing, HP first surfaced its opaque and spurious claims in November 2012.  Around that time, it reportedly brought the allegations to the attention of regulators in the United Kingdom and the United States.  Reflective of the dearth of actual evidence, well more than two years on, the UK’s Serious Fraud Office closed its investigation without charges, and Dr. Lynch was dismissed from the only active lawsuit in the United States that named him as a defendant.

Nor can HP’s far-fetched fraud accusations be reconciled with the fact that almost every senior member of Autonomy’s management eagerly stayed on at HP Autonomy after the acquisition.  Does HP claim that the fraud was so cleverly hidden that Autonomy management could join HP secure in the knowledge that the misconduct would remain undetected, despite knowing that HP would have full access to Autonomy’s books and records?  No.  To the contrary, HP’s claims are purportedly based on the very books and records HP controlled the day the acquisition closed.  And, regarding those books and records, HP conveniently fails to mention that despite the claimed $5 billion fraud, no cash is missing at Autonomy:  Indeed, once all of the supposedly manufactured revenue is removed, Autonomy’s financials show over $450 million cash surplus.  How is this fraud?

Yet another unexplained mystery is why it took HP two and a half years of ever-changing allegations to decide on a story that it is willing to tell in a pleading.  The only constant in that time is the grinding of HP’s tired wagons circling around its own officers and directors—the very people who actually bear responsibility for the destruction of Autonomy’s value (as well as numerous companies before).  Indeed, HP’s campaign of misdirection highlights its fear that its officers and directors will be found liable for the losses, culminating now in a plea to the judge overseeing the shareholder derivative litigation in California to forbid anyone from bringing any Autonomy-related claim against those individuals.

In reality, there is nothing mysterious about any of this.  The simple truth is that HP’s losses were not caused by anyone at Autonomy.  They were caused by HP’s own incompetence and malfeasance.  HP has demonstrated once again its inability to make a potentially transformative acquisition work.  It grossly overestimated the projected synergies to be realized from the deal, bungled the integration of Autonomy into HP, and then hid the whole thing from the market for as long as it possibly could.  That is the real story of fraud here.

Pre-Acquisition:  HP’s Quest for Autonomy’s “Almost Magical” Technology

After boardroom scandals, failed acquisitions, and instability in its executive suite, HP looked to the October 2011 acquisition of Autonomy to reverse its fortunes and transform itself from a low-margin hardware provider to a high-margin enterprise software company.  In particular, HP anticipated that:  (1) Autonomy’s IDOL software, which HP board member (and now CEO) Meg Whitman has called “almost magical,”[2] could be combined with software from HP’s newly acquired structured-data solution, Vertica, to create an industry game-changing technology that could process both structured and unstructured data; (2) HP could leverage its global sales footprint and complementary hardware to drive Autonomy sales around the world; and (3) HP would, overall, use its low-margin legacy assets to drive high-margin Autonomy software sales.

HP was eager to finalize the deal without competition from other potential suitors, or exposing sensitive competitive information to competitors.  HP’s bid reflected that.  It was aggressive in its pricing, but reflected a reasonable value for Autonomy in the hands of a competent acquirer.  However, HP’s officers and directors were well aware that HP was not a competent acquirer.  At the time of the Autonomy merger, HP—and only HP—knew that (a) it was about to abandon its Palm operating platform, which it had acquired only a year earlier for $1.8 billion; (b) it was preparing to announce its departure from the PC market, its core business, which accounted for roughly one-third of its revenue; and (c) it was about to release poor earnings results and lowered guidance, which would certainly lead to a drop in the stock price.  In fact, these announcements—made the same day the Autonomy acquisition was announced—resulted in a one-day drop of 20% in HP’s share price, reducing its market capitalization by $12 billion.

Post-Acquisition:  HP’s Botched Integration and Desperate Search for a Scapegoat

The architects of the Autonomy acquisition were HP’s then-CEO, Léo Apotheker, and then-Chief Strategy and Technology Officer, Shane Robison.  In September 2011, before the Autonomy acquisition was even completed, HP fired Mr. Apotheker and replaced him with Ms. Whitman.  Mr. Robison left HP the following month.  Thus, Autonomy was left to try to integrate into HP without the very people who had conceived of the acquisition and who were uniquely positioned to execute that integration.  There was no handover of the integration plan, and key decisions made pre-acquisition were not implemented.  HP did not even hire a new Chief Strategy and Technology Officer to replace Mr. Robison.  In short, a period of chaos ensued at a critical time for the integration.

The damage was immediate and severe.  Hundreds of key Autonomy employees left in the year following the acquisition.  HP never integrated IDOL and Vertica to create the new product on which the acquisition was largely predicated.  HP’s preexisting staff actively worked against Autonomy by marketing its competitors’ products, due to dysfunctional internal incentive structures.  For example, HP salespeople did not receive “quota credit” or commissions for sales of Autonomy products, as they did for third-party software products, and HP’s Enterprise, Storage, Servers and Networking (“ESSN“) department refused to support Autonomy’s efforts to use ESSN hardware because sales to Autonomy did not count toward ESSN’s sales quota.  HP prevented Autonomy from executing its own marketing strategy and repeatedly excluded Autonomy from participating in HP’s public relations events.  HP did not provide Autonomy with the necessary sales and services support or staff, and HP’s Software department engaged in a power struggle with Autonomy leadership for control over the direction of HP’s software offerings.  HP mishandled the pricing of Autonomy products, sometimes heavily discounting them to incentivize HP’s low-margin hardware sales and at other times marking them up significantly to boost the revenues of lagging HP departments.  Perhaps most telling of the internal discord, HP staff refused to sell or recommend Autonomy products until Autonomy was “certified” for HP hardware, a process that was estimated to take a year.

Predictably, these problems caused a sharp drop in Autonomy’s historical close rates.  Dr. Lynch provided Ms. Whitman with a presentation in May 2012 analyzing the reasons for this decline, including Ms. Whitman’s failure to control the infighting and dysfunction at HP.  A few days later, Dr. Lynch was fired.

Faced with a crisis of its own making, HP chose to write down the value of Autonomy and to make unsubstantiated allegations of accounting irregularities against Autonomy’s former management.  Those statements by HP to the market, beginning on November 20, 2012, and continuing to the present day, were false, and HP knows it.  Since HP first announced the write-down, HP has struggled to explain the numbers underlying its claim.  In that time, HP has busied itself conducting a supposed “investigation,” the result of which was pre-determined from the outset, to support its claims and scapegoat Dr. Lynch, Mr. Hussain, and Autonomy’s former management.  That the allegations against former Autonomy management are unsupported is evident from HP’s ever-shifting rationales for its accusations.  For example, HP claimed in November 2012 to have been unaware that Autonomy sold any hardware.[3]  By September 2014, HP conceded that it had been aware that Autonomy sold hardware but claimed not to have known of the type of hardware that Autonomy sold.[4]  In fact, HP knew about Autonomy’s hardware sales all along.  Additionally, it appears that HP may be using its allegations against Autonomy’s former management and the accompanying re-characterizations of historical revenue to make improper claims for corporate tax refunds.

While we are still reviewing the Particulars, to which we will respond formally in court in due course, it is clear on the face of the Particulars that they are merely the latest evolution of familiar, and flawed, allegations that HP has pressed for the past two and a half years.  What follows is a summary response to some of the core allegations made in the Particulars.  We look forward to receiving and reviewing discovery from HP and are certain that documents solely in HP’s possession at the moment will further reveal the true folly of HP’s allegations.


A.      Autonomy’s Sales and Accounting Systems were Overseen and Approved Internally and Externally by Highly Competent, Independent Professionals

Throughout its life as a public company, Autonomy’s accounts and policies were reviewed by external auditors and judged to be compliant with applicable accounting standards.  Specifically, for the years 2005 through 2010, Deloitte audited Autonomy’s annual statements, which were prepared in accordance with International Financial Reporting Standards (“IFRS“) as adopted by the European Union, the regime under which UK listed companies are required to report their accounts.  Autonomy’s accounting and disclosure complied with IFRS.  Autonomy’s external auditors at Deloitte also reviewed Autonomy’s quarterly and interim financial statements to ensure compliance with the relevant standards and rules.[6]  As such, Deloitte reviewed Autonomy’s financial statements for the first six months of 2011 and reported on July 27, 2011—just before the acquisition by HP—that it had no reason to believe that any aspect of the accounts had not been “prepared, in all material respects, in accordance with” the governing standards.[7]

Deloitte’s judgment was based on all of Autonomy’s relevant financial information.  Deloitte’s practice was to review all sales contracts and invoices for more than $1 million and a sample of contracts for more than $100,000.[8]  In addition, each quarter Autonomy’s management provided a report to the Audit Committee, with a copy to Deloitte, describing (among other things) Autonomy’s financial results, significant events, software and hardware sales, bad debts, and cash collection.

Deloitte reported directly to Autonomy’s independent Audit Committee, which was appointed by the company’s non-executive directors.  The committee was responsible for monitoring the company’s financial statements and public announcements relating to financial performance.  In 2009, the committee included Barry Ariko, a former Silicon Valley hardware and software executive.  In 2010 and 2011, the committee was chaired by Jonathan Bloomer, a chartered accountant, former senior partner at Arthur Andersen, former CEO of Prudential, former Chairman of the Financial Services Practitioner Panel of the Financial Services Authority (“FSA“), and current member of the UK Takeover Panel.  Each quarter, Deloitte provided the Audit Committee with a detailed report analyzing all significant accounting decisions for the relevant financial period, and the committee met quarterly with Deloitte to review Autonomy’s financial statements and discuss significant accounting risks, including revenue recognition and disclosure.  Each meeting included a “no management present” session with Deloitte.  As CEO, Dr. Lynch was not a member of the Audit Committee, did not attend Audit Committee meetings, and was not responsible for accounting decisions.

Notably, revenue recognition was identified as a key focus in Deloitte’s reports to the Audit Committee and, in each case, the Audit Committee approved Autonomy’s financial statements and Deloitte issued unqualified audit opinions.

Nonetheless, HP appears to contend, in paragraph 133.5 of the Particulars, that the location of Dr. Lynch’s and Mr. Hussain’s desks somehow demonstrates a shared awareness of, and responsibility for, all the transactions at issue and their related accounting.  The reality is that Dr. Lynch and Mr. Hussain relied on Autonomy’s extremely experienced auditors and Audit Committee, professional sales force, technical team, finance team, and attorneys.  This argument seems especially ludicrous given that Dr. Lynch and Mr. Hussain had separate offices, with the exception of one Autonomy location that utilized an open-plan office—far from the den of conspiracy implied by HP.

B.      End-of-Quarter Deals are Common and Legitimate in the Software Industry

 HP makes much of the fact that Autonomy often closed deals at or near the end of a fiscal quarter.  That misplaced emphasis simply illustrates how out of touch HP is with the nature of the software business that it supposedly has been operating for three and a half years now.  As everyone who has even a passing familiarity with the software industry knows, it is common for deals to be done toward the end of or on the last day of the quarter.  The reason is simple:  As in any business, customers prefer to hold out for the best deal for as long as possible, and because software can be delivered instantaneously, these negotiations can continue right up until the closing minutes of a quarter.  HP’s ignorance of this is bemusing, especially in light of its purported laser-like focus on accounting rules.  This practice is so commonly accepted and well known that it is described in accounting treatises.[9]  And, despite HP’s professed ignorance, the truth is it engages in the same practices.

C.      Setting Revenue Targets is a Proper Business Practice

HP also complains (in paragraphs 135.2 and 135.3 of the Particulars) that Autonomy managed its sales pipeline in order to hit its revenue target for each quarter.  Again, this allegation evidences a shocking naiveté about the way the software business—and, indeed, any publicly traded company—actually works.

Autonomy was a FTSE 100 company—one of the largest 100 companies listed on the London Stock Exchange (“LSE“) by market capitalization.  As with most LSE companies, and especially those in the FTSE 100, market analysts tracked Autonomy’s financial performance and came to a “consensus” regarding the company’s anticipated revenues and share price for each quarter, half-year, and year-end, based both on analysts’ independent research and on Autonomy’s financial announcements.  Autonomy, like its peers, monitored these numbers and aimed to hit its projected targets for each quarter, in order to manage market expectations and rationalize growth.  Results that fell far short of expectations obviously could lead the market to conclude that the company was in trouble, while an anomalous quarter that far exceeded projections could lead to unduly inflated expectations for future quarters.

Autonomy’s management of its sales pipeline was proper.  Indeed, if it had not done so, it would have been vulnerable to allegations of management incompetence.  Nor did Autonomy do anything out of the ordinary (much less fraudulent or illegal) to achieve those targets.  Indeed, the company’s annual report and accounts repeatedly described its efforts in this regard.  For example, in the Financial Review section of the Annual Report for the year ended December 31, 2010, the CFO, Mr. Hussain, disclosed:

  • “Close management of sales pipelines on a quarterly basis to improve visibility in results expectations”;
  • “Annual and quarterly target setting to enable results achievement”;
  • “Close monitoring by management of revenue and cost forecasts”;
  • “Adjustment to expenditures in the event of anticipated revenue shortfalls”; and
  • “Close monitoring of reseller sales cycles.”[10]

The same Financial Review also disclosed (in case anyone might have been unaware of the fact) that late in-the-cycle purchasing was common in the software industry.[11]

D.      HP’s Own Experts Reviewed Autonomy’s Accounting and Raised No Concerns


Immediately after the Autonomy acquisition was finalized in October 2011, HP hired Ernst & Young (“E&Y“) to perform an independent review of Deloitte’s audit files on Autonomy.  Those files included Deloitte’s quarterly reports to Autonomy’s Audit Committee, which described, among other things, Autonomy’s strategic sales of hardware and reseller transactions (discussed in more detail below).  HP does not claim, and there is no indication anywhere in the available record, that E&Y identified any concerns or issues regarding Autonomy’s accounts or accounting policies.  Rather, E&Y appeared to be comfortable with Autonomy’s accounting.

At around the same time, HP hired KPMG to review Autonomy’s opening balance sheet and past transactions.  KPMG, too, expressed no concerns about the propriety of Autonomy’s accounts or policies, and appeared to be comfortable with Autonomy’s accounting.  In fact, HP and its outside advisors reviewed Autonomy’s accounts thoroughly enough to determine that it would be acceptable to take a $45 million write-off and promptly did so, yet neither party found anything to suggest that there was any impropriety at Autonomy, let alone a massive fraud.


A.        Background

Autonomy sold hardware to drive software sales.  This perfectly legitimate practice was widely known, handled openly within the company, and carefully monitored and approved by Deloitte and the Audit Committee.  In 2009,[12] there were three significant developments in the market for Autonomy’s Digital Safe software:

(1) Shift to appliances:  Many industry analysts were predicting that traditional sales of software would be replaced by the sale of software pre-loaded on hardware—what became known as an appliance.  At around this time, Google developed a search appliance that directly competed with Autonomy’s software products, and Whit Andrews of Gartner, an influential industry analyst, actually de-rated Autonomy for its weakness in the appliance category.  Autonomy was in a difficult position because it could not purchase hardware cheaply enough from manufacturers to produce a competitively priced appliance product or optimize appliance performance via customized hardware.

(2) Customers combining hardware and software on-site:  Autonomy had been selling hosted Digital Safe archiving to banking customers.  As these clients’ archiving needs exploded in an increasingly digital and regulated world, some customers decided to purchase both the software and the necessary hardware (though not necessarily at exactly the same time) and then put the two together on-site as needed.

(3) Strategic supplier status:  By 2009, a number of large Autonomy customers were reviewing their IT supply chains to streamline the number of suppliers that they would deal with directly.  The trend was to designate the largest vendors as “strategic suppliers,” who would in effect act as general contractors, procuring from smaller vendors whatever products they themselves did not make, and taking for themselves some of the profit on the resale.  This shift jeopardized Autonomy’s access to such customers, because it did not sell them enough software to qualify as a strategic supplier.  Unless Autonomy could find a way to significantly increase its aggregate sales to these clients, it would have to route its future sales through strategic suppliers, and would therefore have to dramatically reduce its margin in order to maintain an attractive price to the end user despite the intermediary’s markup (typically 30 percent).  Or, worse yet, certain larger companies with strategic supplier status—such as EMC or IBM—could have decided to produce directly competing software products, potentially eliminating entire categories of key sales.  To ensure strategic supplier status with such customers and thereby protect its software margins and sales opportunities, Autonomy thus sought to increase its aggregate sales volume by offering hardware.[13]

As a result of these three market shifts, Autonomy sought to develop strategic relationships with hardware suppliers in order to:  (1) obtain hardware in competitive volumes and at competitive prices and, through such marketing relationships, generate sales of that hardware to in turn generate sales of software; (2) develop relationships with hardware vendors that could help Autonomy find a solution to the strategic threat arising from the industry move toward appliances; and (3) develop relationships with hardware suppliers to foster joint marketing activities.  In particular, Autonomy’s management made it a priority to develop closer relationships with key hardware suppliers EMC, Dell, and Hitachi.  Autonomy fully disclosed this initiative, and the reasons behind it, to the Audit Committee and Deloitte.  Autonomy’s internal Strategic Deals Memorandum, to which HP refers in paragraph 142.9.1 of the Particulars, sets out Autonomy’s rationale for selling hardware.

The strategy was successful:  Autonomy entered into strategic hardware agreements in Q3 2009 with Citigroup, Bloomberg, and JPMC.  Autonomy initially partnered with EMC to procure this hardware, but shortly after Autonomy began shipping EMC hardware under this arrangement, EMC’s content group acquired an Autonomy competitor, resulting in the termination of the hardware relationship.  Autonomy continued to maintain a cordial relationship with EMC, however, and in 2010 Autonomy even considered acquiring EMC’s content division for approximately $2.4 billion.  Although the deal fell apart when EMC’s content division repeatedly missed its numbers, the negotiations show that the relationship with EMC was a continuing and potentially beneficial one.[14]

As its hardware relationship with EMC came to an end, Autonomy began to form a closer relationship with Dell, which was in any event better positioned to help Autonomy because Dell, unlike EMC, was able to supply servers as well as storage.  As a result, Autonomy used Dell as its hardware supplier from the end of 2009 until 2011.

B.        HP’s Baseless Allegations

HP has alleged three broad categories of accounting irregularities:

  • Autonomy’s loss-making hardware sales;
  • Autonomy’s accounting for hardware sales; and
  • Autonomy’s disclosure of hardware sales.

None of HP’s allegations hold water.

Indeed, HP seems to find its own arguments about hardware less than persuasive, given how many times it has changed its story since it began its public witch-hunt in 2012.  First, HP claimed in its November 20, 2012, press release, and in an interview with Bloomberg the same day, that it had “no knowledge or visibility” of Autonomy’s hardware sales until May 2012.[15]  But HP has since had to admit that E&Y’s November 2011 review of Deloitte’s work papers “identified Autonomy’s hardware revenue, which E&Y reported to HP,”[16] and that KPMG’s 2011 review likewise identified $41 million in Autonomy hardware sales, which “HP management” discussed with Autonomy at the time. [17]  Indeed, we are confident the evidence will show that not only did HP ask specific questions about Autonomy’s hardware sales, including questions about the type and amount of Dell hardware transactions, but also that Autonomy provided prompt, accurate, and complete responses to those questions. Unable to sustain its initial public claims of ignorance in the face of hard evidence to the contrary, HP then made a new allegation:  that although it was aware of the fact and quantity of Autonomy’s hardware sales, it did not know what type of hardware Autonomy sold.[18]  But HP was soon forced to abandon this argument too; a September 12, 2014, Wall Street Journal article quoted an HP spokeswoman who admitted that in 2011 Autonomy “described [its hardware sales to HP] as either ‘appliance sales’ or ‘strategic sales’ designed to further purchases of Autonomy software.”[19]  Contemporaneous emails confirm this fact.[20]

With its prior false claims about Autonomy’s hardware sales conclusively debunked, HP has retreated to vague and unsupportable generalities, asserting that it never knew that Autonomy sold “pure” hardware (whatever that means).  What is clear, however, is that HP was well aware in 2011 that Autonomy sold hardware for the strategic purpose of driving its core software sales.  HP nonetheless appears to imply (in paragraph 54 of the Particulars) that if hardware and software sales were not included on the same order, then there could be no link between the sales.  This implication is utterly unfounded and flatly wrong.  Strategic hardware sales were undertaken to promote not only present but also future software sales; thus, such sales could plainly be linked even though they took place at different times and therefore appeared on separate purchase orders.  To use an analogy, on HP’s reasoning the sale of a video game system could not possibly be linked to the future sale of a game to be played on that system because the two sales would not be recorded on the same piece of paper.  Such an argument is contrary to plain common sense (a problem that confronts all of HP’s allegations against Autonomy, as will be seen below).

Further evidence that the hardware and software sales were linked—if any further evidence were needed—comes in the form of HP’s own continued resales of Dell hardware to drive Autonomy software sales well into 2012, long after the acquisition was completed and HP’s finance team had taken control of Autonomy’s accounting.  Senior HP accountants and advisors inquired about the Dell hardware resales and knew that they were made at a loss.  Indeed, the available record suggests that these resales were ultimately terminated not by HP but by Dell.

HP’s numerical estimates of the revenue generated by what it calls “pure” hardware revenue are likewise far off the mark.  Despite repeated requests, HP has failed to provide the documentation behind these figures, but from the available information HP appears to include in the category of so-called “pure” hardware revenue (a) profitable sales of hardware, (b) hardware sold with software, (c) hardware sold for combination with Autonomy software on the customer site as part of appliance programs, and (d) hardware sold under agreements where customers explicitly linked hardware and software purchases from Autonomy for calculations of overall discount targets.  HP does not explain, of course, how such transactions can possibly be characterized as loss-making sales unconnected to profitable software sales (if, indeed, that is its intended meaning of “pure” hardware sales) and thus even on its own definition HP overestimates the percentages of revenue involved.[21]

1.      Autonomy Properly Accounted for Its Hardware Sales

Autonomy appropriately accounted for the costs associated with its strategic hardware sales.  Contrary to HP’s assertion (in paragraphs 68.3 and 68.4 of the Particulars) that Autonomy improperly accounted for some of those costs as sales and marketing expenses rather than costs of goods sold (“COGS“), IFRS provides no accounting standard governing cost accounting for discounted strategic sales, or the extent to which such costs should be allocated between COGS and sales and marketing costs.  Rather, this is an area reserved to the directors’ judgment, subject to the scrutiny of internal and external auditors.  Autonomy’s allocation of these costs was at all times reviewed by its Audit Committee and Deloitte, who uniformly approved Autonomy’s accounts.

Under IAS 8.10, the objective of the cost allocation should be to reflect the underlying economic substance of the transaction.  Accordingly, the appropriate test is whether and to what extent Autonomy’s strategic hardware sales were made for marketing purposes—i.e., to promote future software sales.  Consistent with this principle, Autonomy estimated the COGS component of its hardware costs by looking to the standard cost of such hardware (as provided, for example, by EMC), and then categorized the remainder of its actual costs—i.e., the amount above the standard hardware cost, which could only be attributable to Autonomy’s strategic purpose in undertaking the sales—as a sales and marketing expense.  This straightforward and common-sense approach follows the normal accounting and business methods for establishing the value of separate parts of a transaction and, again, was fully overseen by Deloitte, who satisfied itself that this accounting complied with IFRS.

HP’s allegations to the contrary are based on selective quotations taken out of context and mischaracterized.  For example, paragraph 142.8.1 of the Particulars alleges that Mr. Hussain misrepresented to the Audit Committee in Q3 2009 that Autonomy and EMC were engaged in a “highly targeted joint marketing program” and that Autonomy had “spent US$20 million on shared marketing costs with EMC” when, in reality,  neither the joint marketing program nor the shared marketing expenditure existed.  The full quotation from the note reveals, however, that Mr. Hussain was accurately describing movement in quarterly operating costs: “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).”[22]

Similarly, in paragraph 142.8.3 of the Particulars, HP provides a misleading selective quotation in support of its claim that Mr. Hussain misrepresented to the Audit Committee in Q2 and Q3 2010 that the costs of hardware had been charged to COGS when, in fact, a portion of those costs had been charged to sales and marketing expenses.  Mr. Hussain’s actual statement, however, was critically different from HP’s portrayal:  “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 [sic] 50% of the cost would be used for marketing purposes.”[23]  This sentence conveyed that costs equal to the amount of the sales were allocated to COGS and the loss on the sales was allocated to the sales and marketing line.  Deloitte made this clear in its report:

Management has taken all of the costs associated with the Dell hardware sales to cost of sales with the exception of the loss of approximately $3.8 million which has been allocated to sales and marketing expenses. . . . We have reviewed management’s analysis of the linkage between the loss making strategic hardware sales and subsequent profit making software sales and accept the decision of management to allocate the loss of $3.8 million to sales and marketing.[24]

Finally, Autonomy’s marketing-cost allocation did not mislead the market by distorting its gross margins.[25]  The prosaic reality is that the gross-profit metric fluctuated in any event, and any impact of the hardware accounting was insignificant and consistent with the usual variations in this margin, which were highlighted in various analyst calls.[26]  In this context, the argument that a variation of a few percentage points in gross margin would have materially affected investors’ views of the company is patently false.  Additionally, HP was aware of Autonomy’s hardware accounting immediately following the acquisition, if not sooner, yet made no indication that the accounting was problematic until over one year later.

In short, Autonomy’s hardware resales were precisely targeted, properly accounted for, fully explained to its internal and external auditors, and did not distort either the market’s assessment of Autonomy’s value or HP’s valuation of the company.  Contrary to HP’s baseless accusation in paragraph 142.4 of the Particulars, Autonomy was not acting as a generic hardware reseller such as Morse.  Autonomy’s hardware sales, typically made at a loss and always designed to benefit the software business, were plainly dissimilar to the business of a generic reseller of hardware.

2.      Autonomy Fully Complied with Its Disclosure Obligations

As HP implicitly concedes in paragraph 30.1 of the Particulars, Autonomy acted in full compliance with IFRS in classifying the income generated by its hardware sales as revenue.  There is no proper basis for deducting the hardware transactions from Autonomy’s revenue figures.  The hardware transactions took place, hardware was delivered, and revenue and cash were generated in return.

IAS 18 requires disclosure of each significant category of revenue, with significance to be determined as a matter of judgment.  Here, Autonomy’s highly qualified accounting executives determined that its hardware sales did not constitute a significant category of revenue for the purposes of IAS 18, and the Audit Committee and Deloitte concurred in the reasonableness of that judgment.  Indeed, that judgment was not only reasonable but also patently correct.  The discounted hardware sales were merely ancillary to Autonomy’s software sales, and the overall effect was a high-margin software business.

Under IFRS 8, the revenue stream from any particular component of sales should be separately disclosed where two criteria are met:

(a)         The component qualifies as an “operating segment” under IFRS 8.5 to 8.10; and

(b)        The revenue stream from that component meets the quantitative threshold established by IFRS 8.13.

An operating segment, as defined by IFRS 8.5–8.10, is a corporate component:

(a)     That engages in business activities, and may thereby earn revenues and incur expenses;

(b)     That has operating results that are regularly reviewed by the chief corporate decision maker; and

(c)     For which discrete financial information is available.

A corporate activity whose revenues “are only incidental” to the corporation’s business is not an operating segment.

Autonomy properly applied these principles.  Its hardware sales did not qualify as an operating segment because:

(a)     Autonomy did not operate a separate hardware sales division;

(b)     Hardware sales were not subject to separate review by Dr. Lynch; and

(c)     Autonomy’s hardware sales were incidental and made primarily to drive the core software business.

As a result, Autonomy had no obligation to separately disclose its hardware sales under this guidance.  Autonomy’s single-segment disclosure was reviewed by its Audit Committee and by Deloitte, who concurred with management’s decision to disclose Autonomy’s revenues as a single segment and properly found that Autonomy’s hardware sales disclosure was consistent with its accounts.

Nor did Autonomy mislead the market by stating that it was a “pure software” company, despite HP’s incessant allegations to the contrary.  Indeed, that argument is not only false but also utterly disingenuous:  HP is (and was) well aware that the term “pure software” did not signify that Autonomy did not sell hardware, nor did the market ever interpret it to mean any such thing.  Autonomy used the term merely to distinguish itself from other software companies that derive a significant portion of their revenue from the provision of professional services, including consultancy.

Autonomy’s business model focused on developing and licensing IDOL software.  The company was not designed to be a professional services provider; rather, its model depended on the existence of a community of expert systems integrators and consultants who could customize and implement the relevant IDOL software for each customer’s particular uses.  This meaning was clear not only from the very statements of which HP complains, which drew explicit distinctions between Autonomy’s “pure software” approach and the services component of similar companies, but also from the many public sources of information disclosing that Autonomy sold hardware.[27]

Indeed, it is self-evident that software needs hardware to work and, as a result, it is quite common for enterprise software companies to sell some hardware as an essential adjunct to the core business of promoting and procuring software sales.  Thus, it simply is not plausible that a prospective buyer that knew anything at all about the software industry could have thought that the words “pure software” meant that Autonomy sold no hardware at all.  And, perhaps most tellingly, investment bank analysts’ notes, which HP actually reviewed during its pre-acquisition valuation of Autonomy, made clear that the market understood the term “pure software” to relate to a professional services-lite operation, not an absence of hardware.

In conclusion, Autonomy did nothing wrong in selling hardware, appropriately accounted for its hardware sales in accordance with IFRS, and sufficiently disclosed its hardware sales to its auditors and to the market.  Further, the evidence will show that HP was aware of Autonomy’s hardware sales prior to the acquisition and that HP continued to resell Dell hardware post-acquisition.  In these circumstances, HP’s allegations of fraud are transparently meritless.


A.        Reseller Sales are Commonplace and Commercially Sound

Unsurprisingly, as group CEO, Dr. Lynch had little involvement in negotiating reseller deals or determining the related accounting.  As a major player in the industry, however, Dr. Lynch was generally aware of the benefits of reseller deals and the importance of developing a reseller ecosystem.

In common with most software companies, including HP, Autonomy sought to develop and strengthen its reseller ecosystem, and thus often sold through reseller partners.  Both Autonomy and the resellers derived commercial benefits from this arrangement.  For Autonomy, these sales helped expand the network of providers that understood, marketed, and serviced Autonomy products.  Through repeat business, resellers developed a deep understanding of Autonomy products, which enabled them to provide services and support to Autonomy’s customers (lower-margin tasks that Autonomy had little interest in performing itself).  Additionally, certain resellers held GSA pricing schedules and/or government and security clearances beyond those held by Autonomy’s own employees. In partnering with these resellers, Autonomy gained access to markets it had not already penetrated.  Under IFRS, a reseller was itself Autonomy’s customer for accounting purposes, and thus selling to the reseller amounted to a valid and enforceable sale.

Contrary to paragraph 74.3.6 of the Particulars, resellers could derive numerous commercial benefits from these sales.  Such deals gave resellers the opportunity to connect with a new customer, gain a new reference logo, provide services, lock out competitors, and simply make sales and margin.  And even where there was an intended end user of the product, the reseller could always negotiate with and sell to a different end user if it thought it could get more favorable terms.

As one means to foster these relationships, Autonomy would occasionally pay a marketing assistance fee (“MAF“) to compensate a reseller for its assistance in a sale or to make up for its lost margin where an end user ultimately elected to purchase directly from Autonomy.  Autonomy took such steps voluntarily and in its discretion after a sale was completed.  In each instance, such steps advanced Autonomy’s valid commercial interests in the success of its partners, the sale of its products to interested customers, and the related development of an ecosystem around its products.  Deloitte reviewed and approved Autonomy’s use of MAFs, and openly discussed them in the Reports to the Audit Committee.

HP’s feigned surprise at finding that Autonomy engaged in reseller deals is especially disingenuous in light of the fact that HP operates a number of its own reseller compensation programs—ranging from opaque “contra” transactions to programs in which an HP reseller is entitled to rebates and referral fees for sales by HP where the partner somehow influenced the sales process.  HP’s Reseller PartnerOne EMEA Program Guide, dated November 2012, explains that the PartnerOne program compensates reseller partners for, among other things, referrals:  “A referral is when a customer buys directly from HP and the partner has been recognized for their contribution to the closing of that deal via a referral fee.”  The guide goes on to list a few examples of such compensable contributions:  “Referral fees are discretionary payments made to non-reselling partners, where the partner had demonstrable influence on a deal made directly between HP Software and the customer.  Demonstrable influence can refer to: identification of sales opportunity, recommendation of the HP solution, involvement in the sales cycle, proof of concept, solution architecture, or associated return on investment.[28]

Notwithstanding the frequency of these relationships in the software industry and the well-understood benefits of developing a reseller ecosystem, HP suggests that Autonomy was somehow unique in developing this ecosystem and attempted to conceal its relationships with resellers. Contrary to paragraph 136.5.1 of the Particulars, the fact that Autonomy partnered with resellers was transparent and any decline in Autonomy’s use of resellers after the acquisition was primarily due to HP’s desire to funnel business to its own services arm, HP Enterprise Services.

B.        HP Fundamentally Misunderstands the Relevant Accounting Principles

HP’s allegations regarding reseller transactions are based on its fundamental misunderstanding of IAS 18, Autonomy’s accounting policy, and the nature of the reseller transactions themselves, all of which were legitimate and appropriately accounted for under IFRS.  Further, it should be noted that HP singles out only a handful of individual transactions, making the baseless assertion that the resellers were not on risk for those transactions (despite the voluminous documentary evidence making clear that they were), and then desperately tries to extrapolate from these deals to the 23,000 reseller transactions that Autonomy conducted in the relevant period.

HP’s first, and most basic, error is its failure to comprehend that for accounting purposes it is the reseller, not any intended end user, that is Autonomy’s customer.  Cathie Lesjak, HP’s CFO, misunderstood this as long ago as November 2012, and HP apparently has never regained its wits on the issue.[29]  The naming of an intended end user in the sales documentation with a reseller, although helpful to Autonomy in its management of the business, is not and never was a requirement under IFRS.

HP also misunderstands the concept of “risk” under IAS 18.  HP’s allegation that resellers were not “on risk” when Autonomy recognized revenue on its sales to them is unsupported by the evidence and, indeed, flies in the face of the resellers’ repeated and well-documented acknowledgments—in response to inquiries by Deloitte—that they were indeed on risk.  Nor was there any pattern of returns or cancellations sufficient to cast doubt on the propriety of revenue recognition upon a sale to any reseller.  Deloitte actively reviewed reseller transactions, was alive to the issue of risk transference, and, in each case, approved Autonomy’s accounts.

Consistent with IFRS and IAS 18, Autonomy’s revenue-recognition policy provided for recognition of revenue on sales of IDOL product to resellers when the software licenses at issue had been “delivered in the current period, no right of return policy exist[ed], collection [wa]s probable and the fee [wa]s fixed and determinable.”  Autonomy’s policy was approved by Deloitte and the Audit Committee, disclosed in Autonomy’s annual report, and led to the appropriate recognition of revenue under the five-element IAS 18 test:

(1)       Autonomy transferred the significant risks and rewards of software ownership to each reseller upon executing a purchase order and delivering the software.

(2)       Autonomy did not retain continuing managerial involvement to the degree usually associated with ownership or effective control over the goods sold.  The reseller had ultimate control over any resale.

(3)       Autonomy’s sales to resellers were fixed with regard to price at the time of sale and were not contingent upon future events.  As a result, Autonomy was able to reliably measure revenue at the time of each sale.

(4)       Autonomy recognized revenue on a sale to a reseller only where it was probable—i.e., more likely than not—that Autonomy would be paid by the reseller.  Autonomy assessed collectability according to a number of criteria, including the reseller’s creditworthiness based on third-party credit reports, payment history, and the strength of the reseller’s balance sheet.

(5)       Autonomy maintained records of all sales and, as a result, reliably measured costs.

1.      Transfer of Risks and Rewards

Autonomy transferred the significant risks and rewards of software ownership to each reseller when the contract was executed and the software was made available for delivery.  Autonomy transacted with resellers pursuant to non-recourse agreements, under which neither the fixed price for a license nor the obligation to pay was contingent on a reseller’s eventual resale of the license to an end user.  As a result, each reseller assumed the risk of not reselling the software to an end user.  As a further precaution, it was Deloitte’s policy to obtain written revenue confirmation directly from each reseller before approving Autonomy’s recognition of revenue for deals over $1 million.  Those letters confirmed that the resellers were on risk and that no side agreements, written or oral, were in place.[30]

2.      Continuing Managerial Involvement

Autonomy did not retain continuing managerial involvement to the degree usually associated with ownership or effective control over the goods sold.  Any post-sale involvement by Autonomy did not amount to the type consistent with ownership or control that IAS 18.14 contemplates.  Contrary to HP’s insinuations, IFRS does not prohibit contact with an end user after a sale to a reseller; rather, it is not only permitted but also commonplace for software companies to maintain contact with an end user following a sale to a reseller.  For example, a reseller may rely on the software company to act as its negotiating agent; a transaction with a reseller may be only one part of a larger contemplated deal; or a software company may have a long-standing relationship with an end user that is important to maintain.

Contrary to HP’s statement in paragraph 74.3.3 of the Particulars, IFRS does not require a reseller to add value in connection with a specific sale to an end user in order for the revenue on the manufacturer’s sale to that reseller to be recognizable.  Under IAS 18.14(b), in order to recognize revenue on a sale, a seller may retain neither “continuing managerial involvement to the degree usually associated with ownership nor effective control over the goods sold.”  HP appears to assert that resellers lacked active participation in end-user negotiations following a purchase from Autonomy and that this undermines Autonomy’s satisfaction of this element of the test.  It does not.  Simply stated, there is no requirement under IFRS that a reseller initiate or take the lead in negotiations on an onward sale to an end user, or otherwise add value to the particular product sold, before revenue can be recognized on the transaction between the seller and the reseller.

The relevant consideration under IAS 18.14(b) is not whether the reseller participates in negotiations with the ultimate end user, but whether the logistics of the resale are within the seller’s ultimate control and whether the seller is responsible for the management of the goods after the sale.  Here, Autonomy lacked, and the resellers possessed, control over both the decision to whom and whether the reseller sold the software, and management of the software following the reseller’s purchase.  Both features demonstrate that Autonomy relinquished managerial involvement and effective control to the resellers that were its customers.

3.      Fixed Price Not Contingent on Future Events

On infrequent occasions, a reseller’s contemplated onward resale to an end user fell through, either because no end user purchased the software at all, because Autonomy entered into a direct transaction with the anticipated end user, or because the reseller decided to sell to another reseller instead of the originally contemplated end user.  Those subsequent events did not, however, undermine the appropriateness of Autonomy’s recognition of revenue at the time of sale to the reseller.  Accounting judgments are not made with hindsight.  Simply put, a change in circumstances that occurs after a sale, which was not contemplated at the time of the transaction, does not undermine the initial recognition decision.  Additionally, Deloitte carefully tracked reseller sales across quarters and clearly understood the fate of these deals.

While, at times, resellers used the failure of an onward resale to try to excuse nonpayment to Autonomy or to negotiate extended payment terms, these unremarkable commercial tactics did not affect a reseller’s obligation to pay or the propriety of Autonomy’s recognition of revenue at the time of the sale to the reseller.  Deloitte was aware of these issues and tracked and considered reseller payment histories when assessing the propriety of revenue recognition.  For example, the Report to the Audit Committee on the Q1 2010 Review stated:

Management alerted us to the fact that two deals sold to Microtech in Q4 2009 have been credited in this quarter and resold directly to the two end users.  This was as a result of the end users wanting to transact directly with Autonomy.  This reduced the profit in the period by approximately $4 million.  As there is no significant history of deals being reversed in this way, management has recognised the revenue at the point of sale to the reseller.  Management has confirmed that these were isolated incidents which are not expected to be repeated in future periods.[31]

On the limited occasions when a reseller deal was cancelled or debt forgiven, Autonomy granted the relevant concession as a business decision and not because this was expected at the outset.  Moreover, such rare events in the audited period were known to Deloitte, who ensured that the related accounting treatment was appropriate.

4.      Probability of Payment and Records of Cash Received

Cash for reseller deals was nearly always received, and the cancellation rate on contracts was low.  In its purported restatement of the Autonomy accounts, HP appears to indicate that a significant number of these sales should never have been booked at all, despite the fact that the bulk of the cash was actually received.  On the currently available information, out of revenues in the period of around $2.3 billion, HP has removed $140 million of revenue from reseller deals.  However, HP itself concedes that cash of $102 million was received against this revenue.  HP makes no explanation regarding where this now-surplus cash came from.

C.      Specific Deals

While each deal was different and each reseller and end user unique, the following discussion  addresses a few of the many glaring deficiencies in HP’s allegations, which highlight HP’s incorrect application of IFRS.[32]

1.      Capax (End User Kraft)

Capax was a key part of the Autonomy ecosystem, and HP still considers it a major HP Autonomy partner to this day.[33] Deloitte was comfortable with Autonomy’s revenue recognition on sales to Capax, as discussed below.  Moreover, in 2011, the FRRP considered Capax’s ability to stand by its obligations to Autonomy and concluded as follows, in a letter dated August 22, 2011:  “The Panel notes that [Autonomy] started to do business with Capax in early 2009 and that Capax has had an excellent payment record since then. . . .  The Panel notes that [Capax's accounts] show Capax to be solvent at 31 December 2008 and to have traded profitably in the year then ended.”[34]

In Q3 2009, Autonomy executed a $4 million deal with Capax for end user Kraft.  It is plain that Autonomy’s accounting treatment of this sale was in accordance with IFRS and approved by Deloitte, as demonstrated by the Report to the Audit Committee on the Q3 2009 Review:

Kraft Food Inc (“Kraft”)

This is a $4 million licence deal for a suite of Autonomy software including Zantaz Digital Safe, Aungate and Introspect.  Support and maintenance has been charged at $0.2 million which is consistent with fair value on a deal of this size.  It should be noted that this deal has been signed through the reseller, Capax Discovery LLP (“Capax”).  As Capax are up-to-date with their payment terms with Autonomy, and all other revenue recognition criteria have been met, management has concluded it is appropriate to recognise revenue.[35]

In Q4 2009, for reasons unforeseen at the time of the reseller deal, Autonomy closed a direct deal with Kraft.  Thereafter, Autonomy, in an exercise of its discretion and for sound commercial reasons, cancelled the reseller deal and offered Capax compensation for its earlier assistance.  HP’s allegations about the Kraft transaction were investigated and dismissed by the FRRP, which in 2011 considered Kraft’s decision to seek a direct agreement with Autonomy and concluded that “Kraft may have felt more comfortable dealing with the key supplier which was also a global company able to provide 24 hour support worldwide.”[36]  The FRRP also understood the reasons why Autonomy paid a MAF to Capax on this deal.[37]  The FRRP dismissed the allegations concerning the Kraft transaction and declined to pursue the matter further.  Having been reviewed and approved by Deloitte, the Audit Committee, and the FRRP, it is clear that the accounting for the deal was proper, despite HP’s desperate attempts to claim otherwise.

2.      MicroTech (Intended End User the Vatican)

HP alleges in paragraph 78 of the Particulars that Autonomy’s sale of software to MicroTech, with the Vatican as the intended end user, lacked commercial purpose.  HP is wrong.  From approximately 2008/2009 to 2012, Autonomy and the Vatican negotiated a deal to digitize the Vatican Library’s collection of 80,000 manuscripts.  Autonomy and the Vatican worked together over many months to develop a full test bed system in Vatican City.  This system was the subject of television interviews, and the Vatican issued press releases about the project.  In November 2009, Mr. Hussain presented the deal to Autonomy’s board for approval.  At the time, Autonomy was to receive €75 million over ten years from the deal, which would be not only lucrative but also high-profile and likely to draw market interest to Autonomy’s cutting-edge solutions.

On March 31, 2010, as negotiations continued, Autonomy sold MicroTech $11.55 million of software to be resold to the Vatican as part of the larger Autonomy/Vatican deal.  Autonomy wanted to involve a partner with sufficient Autonomy expertise to write application code.  Other partners would in time be required in Italy to help with installation.  The auditors were familiar with the MicroTech/Vatican deal, which was discussed and examined during the Q1 2010 review, and confirmed that the accounting was proper.

Negotiations with the Vatican continued into 2012, but HP ultimately decided not to pursue the deal, apparently ceding at least part of it to EMC.  The Financial Times reported on January 23, 2015, that the digitization of the Vatican Library was about to take place with NTT DATA as the service provider.  In short, it is clear that the Vatican deal was a real opportunity and not, as HP insinuates, a pretext for Autonomy to recognize phantom revenue.  Indeed, there is every reason to believe that Autonomy would ultimately have secured the deal for itself had HP not abandoned the negotiations.

The MicroTech deal satisfied all of the revenue-recognition tests under IFRS.  MicroTech was on risk for the balance of its debt, regardless of whether a deal closed with the Vatican.  Autonomy did not retain effective control over the goods sold.  And, indeed, MicroTech made payments in Q4 2010 and Q2 2011 to reduce the Vatican balance that it owed to Autonomy, paying off all but $2.3 million of its $11.55 million balance on the deal prior to the acquisition. HP then wrote off the remaining $2.3 million despite the fact that negotiations with the Vatican were still ongoing at the time.

3.      FileTek (Intended End User the VA)

Contrary to HP’s claim in paragraph 91.1 of the Particulars, FileTek had been reselling Autonomy technology for over a year before the Department of Veterans Affairs (“VA“) transaction of which HP complains.  Likewise, Autonomy had been working with the VA for some time.  When this particular deal was first considered, evidence was mounting that the VA would be issuing a request for information (“RFI“) with a value of tens of millions of dollars, and that Autonomy was likely to be the chosen technology (since, as mentioned above, the VA was an existing Autonomy customer).

In such situations, there are multiple ways in which a manufacturer or reseller could seek to participate in a sale.  First, the reseller could respond to the expected RFI.  Second, the reseller could supply the likely bidder or bidders for the RFI.  Third, the reseller could sell a stopgap solution directly to the customer, if issuance of the RFI was delayed.

It is unsurprising that Autonomy and FileTek decided to partner on the VA deal, given that FileTek was involved in archiving structured data and was keen to get into the much higher growth area of unstructured data archiving offered by Autonomy.  Additionally, Autonomy was a logical partner given the likelihood that Autonomy would be involved in the VA deal, however it was structured.  The VA purchase order from FileTek reflects this, stating that the software could be licensed “either directly by [FileTek] or through an agreed to prime contractor to the United States Veterans Administration Authority (the ‘VA’) or to a Alternate Licensee.”  FileTek could therefore sell the software to anyone, not just the VA.  For example, FileTek could sell the software to a bidder on the VA deal, or as a stopgap to the VA, or to any other party or reseller not related to the VA.  It thus appears to have made commercial sense for FileTek to enter into the deal, and FileTek paid Autonomy fully for the deal.


As a matter of common sense, where partners had invested in creating an Autonomy-related business by, for example, building their capacity and knowledge with respect to a particular Autonomy product, it was in Autonomy’s commercial interest to support them in that endeavor by assisting them with implementation, funneling business in their direction, or using them as vendors when Autonomy had a particular need.  Where Autonomy made purchases from its customers, it was appropriate under IFRS to account for those purchases in the same way as any other purchase so long as the exchange involved “dissimilar goods,” had “commercial substance,” and the “fair value” of the goods could be measured reliably.

Nonetheless, HP alleges that Autonomy engaged in so-called “reciprocal” purchases, and seeks to improperly remove large swaths of revenue generated by legitimate transactions by “netting” them against distinct sales of dissimilar goods to the same customers.  This approach is not only incorrect as a factual matter but also betrays, once again, a fundamental misapprehension of the relevant accounting rules.

Contrary to HP’s claims in paragraph of the Particulars, the theory that Autonomy made purchases to provide customers with cash to pay debts on previous deals fails for four key reasons:

(a)     Fair value on each purchase was proved at the time, and in some cases Autonomy sold the acquired item on at a profit;

(b)     The resellers often used the cash paid by Autonomy to provide services, and this is well documented;

(c)     To net revenue on such transactions would misstate the accounts; and

(d)     HP cannot explain why Autonomy would have spent cash on products of no value to it when it would have been much simpler and far more cost-effective to simply introduce the reseller to a new end-user customer for its Autonomy software, which would have been appropriate under IFRS.

The purchases at issue had commercial substance and fair value.  IAS 18.10 provides that “[t]he amount of revenue arising on a transaction is usually determined by agreement between the entity and the buyer or user of the asset.  It is measured at the fair value of the consideration received or receivable taking into account the amount of any trade discounts and volume rebates allowed by the entity.”  Consistent with this rule, when Autonomy entered “exchange transactions,” the related revenue recorded was “measured at the fair value of the goods or services received.”

Autonomy provided its auditors, and Deloitte carefully evaluated, evidence supporting the “fair value” of the goods that Autonomy purchased from its customers.  Autonomy normally obtained such evidence by gathering quotes from parties selling competing products.  Deloitte considered this information, in conjunction with signed purchase orders, before confirming that fair value had been appropriately established.  There can be no question that Deloitte was careful and thorough in its evaluation of these transactions.  For example, in assessing Autonomy’s purchase of FileTek software, StorHouse, in Q4 2009, Deloitte went so far as to bring in its own IT people in order to understand the commercial rationale for the purchase.

Further, the instances cited by HP include sale and purchase transactions with the same customer where cash was received and paid.  As stated in paragraphs 10 and 11 of IAS 18, “[t]he amount of revenue arising on a transaction is . . . . measured at the fair value of the consideration received or receivable . . . . [and in] most cases . . . is the amount of cash or cash equivalents received or receivable.”  In HP’s recent filings in the UK Action, for each of the allegedly reciprocal purchases that involved cash consideration, both the revenue and costs recorded by Autonomy reflected the fair value of the relevant products sold and purchased.  Examples of these are addressed below in more detail.

HP’s claim that buy and sell transactions with the same counterparty should be netted against each other is incorrect; under paragraph 12 of IAS 18, dissimilar goods cannot be netted and exchange transactions involving similar goods also are not netted; rather, under paragraph 12 the latter are treated as generating no revenue at all, and thus there is nothing to be netted.

A.      Specific Deals

1.      The MicroTech ATIC

Throughout 2009 and 2010, Autonomy made strides to grow its federal business.  With this objective in mind, it set ambitious growth targets for 2011.  In order to secure federal contracts, however, Autonomy needed a space that complied with federal clearance requirements in which to demonstrate its capabilities to federal customers.  As a possible solution, Autonomy considered partnering with a federally cleared entity.

In November 2010, MicroTech proposed building a federally cleared facility—the Advanced Technology Innovation Center (“ATIC“)—in which to showcase Autonomy solutions to federal customers.  MicroTech proposed to both market existing solutions and help Autonomy develop and test new solutions.  To do so, MicroTech needed to hire a full-time federally cleared staff, lease a sufficiently large office space, and acquire the necessary technology[38]—and much of this cost would be incurred up front.  In light of the substantial investment and construction involved, it was always understood that the facility would not be available until sometime in 2011.

The parties entered into negotiations on the deal on the basis of a comprehensive document.  On December 30, 2010, after negotiating price and license terms, including a discount of almost 17 percent for early payment, Autonomy purchased a three-year license for the ATIC.  Autonomy paid MicroTech $9.6 million for the license the following day.

Mindful of the fact that MicroTech was a repeat Autonomy partner, Deloitte reviewed the transaction, confirmed that Autonomy paid fair value, and approved the accounting.  Specifically, the Report to the Audit Committee on the Q4 2010 Review stated:

In December 2010 a licence was purchased by Autonomy from Microtech for $9.6 million and capitalised on the balance sheet as an intangible asset.  This relates to a 3 year license to use a fully US federal government certified demonstration facility which will remain the property of Microtech throughout the licence term.  The licence will include six dedicated US federal cleared employees of Microtech who each have many years of experience in dealing with US federal government sales.  We understand that these individuals have significant business connections with US federal organisation[s].

Given that Microtech is a regular customer of Autonomy, we have reviewed this transaction to ensure that it makes commercial sense and that there is no indication that this should be considered a barter transaction. . . . We held discussions with Sushovan Hussain and Dr Pete Menell who explained that at present Autonomy struggle[s] to make significant sales into the US Federal government agencies because they do not have the appropriate level of certification.[39]

In June 2011, the ATIC was launched with much marketing fanfare and was made available for Autonomy’s use.  Contrary to HP’s claim in paragraph 78.7 of the Particulars, the ATIC was hardly a van (and the fact that HP is making such an assertion after two and a half years of “investigation” is simply astonishing).  Rather, the facility included a MicroData Center; Emerging Technologies Center; Test, Evaluation, and Integration Lab; and Mobile data center (which could be used to offer demonstrations to customers off site).  These components were specified in a thirty-page document.  Because the ATIC was launched in June 2011, however, Autonomy had few opportunities to utilize the facility before being acquired by HP.  Following the Autonomy acquisition, HP decided not to use the ATIC after all, and wrote the asset off without informing Autonomy management.  HP then introduced a new method of managing federal customers.

2.      FileTek StorHouse

HP claims, in yet another misguided attempt to find some justification for the write-down, that Autonomy’s purchases of FileTek’s StorHouse software in Q4 2009 and Q2 2010 were reciprocal transactions and valueless to Autonomy.  Even further, HP claims that StorHouse was never successfully integrated with Autonomy’s software, used by the company, or sold on to a third party—all in the face of significant documentary evidence to the contrary (which would have been revealed by even a cursory review of the purchase orders to which HP had full access following the acquisition).

By way of background, Autonomy had established itself as an industry leader in unstructured information, but it had long been a strategic goal to combine that capability with structured data power.  In fact, the prospect of creating a unique combined solution was a key motivating factor in HP’s subsequent acquisition of Autonomy.  As part of that goal, Autonomy purchased FileTek’s StorHouse software in Q4 2009 and Q2 2010.

Deloitte undertook a comprehensive review of the purchase, with the help of its own IT specialist.  Contrary to the statements in paragraphs 88.4 and 146.1 of the Particulars, Deloitte reported the following in the Report to the Audit Committee on the Q4 2009 Review:

We have reviewed the accounting for sale to and the purchase from Filetek.  In addition to our standard procedures when auditing revenues, we involved a member of our IT specialists to ensure that the software purchased made commercial sense and was not in any way linked to the sale of Autonomy product to Filetek.  We have sighted management’s work on confirming the fair value of this purchase and concur with the accounting treatment applied to the purchase and the sale to Filetek.[40]

As part of the Q4 2010 Review, Deloitte also confirmed that Autonomy paid fair value for the purchase and that the accounting treatment was appropriate.[41]

Deloitte’s statements regarding the demonstration in January 2010 are in direct contradiction to HP’s assertions in paragraph 88.4 of the Particulars that Autonomy did not download the software until April 2010.  In fact, shortly after the purchase, engineers in both the United Kingdom and the United States undertook serious and substantial efforts to integrate the FileTek software, in consultation with FileTek’s own team.  Autonomy ultimately considered four options for integration and chose the most technically complex and ambitious of these options, which, according to Autonomy engineers, provided the best offering to the customer and used both systems to their full potential.  It did so by permitting Autonomy’s software to directly access data from a structured database.

By June 2010, Autonomy engineers confirmed that the FileTek software had been successfully integrated into its Digital Safe platform, with Autonomy presenting an integrated offering to Kraft.  Autonomy’s successful integration of FileTek is reflected in:  (1) the Report to the Audit Committee on the Q2 2010 Review, which specifically notes Autonomy’s success with Kraft; and (2) subsequent Autonomy contracts, which list FileTek software as part of the Digital Safe 9.0 platform.  In July 2010, Autonomy conducted an additional technical demonstration for Deloitte.  That demonstration is noted in the Report to the Audit Committee.[42]

Over the next year, Autonomy made a number of significant sales of Digital Safe with integrated FileTek technology, and expanded its use of FileTek software with its archiving product LiveVault and in connection with newly acquired Iron Mountain data centers in the United States and the United Kingdom.  Autonomy’s integration of FileTek software into its Iron Mountain data centers, and its use of the software, is well documented.  For example, in August 2011, Autonomy circulated a list of its offices and sites that were using FileTek software.

In conclusion, Autonomy’s purchases of FileTek software were consistent with its strategy for entering the structured-data market, acquired at fair value, successfully integrated into Autonomy software, and sold to third parties.  The transactions were reviewed by the Audit Committee and Deloitte, and also by the FRRP, and consistently found to be proper.

3.      VMS

At paragraph 85.3 of the Particulars, HP entirely mischaracterizes the context in which Autonomy acquired data feeds from Video Monitoring Services (“VMS“). Prior to the VMS transaction, Moreover Technologies Inc. (“Moreover“), then a young company, used Autonomy technology, and in consideration provided Autonomy with news data feeds at no charge, which Autonomy in turn used to demonstrate its own technology.  At around the time of the VMS deal, however, Moreover informed Autonomy that it had begun to use an Autonomy competitor’s technology and would no longer supply Autonomy with free data feeds.  Thus, Autonomy sought an alternative source for such data feeds, considered quotes from several suppliers, and ultimately executed a deal with VMS.

It appears that HP’s real reason for proposing that these sales be reversed is not that they were part of “reciprocal” transactions, but that HP ultimately wrote off the related receivable balances because VMS went bankrupt in the second half of 2011.  Where no cash is ultimately received on a transaction because a reseller’s financial position has deteriorated significantly, this can give rise to a bad-debt provision in accordance with IAS 18 but does not affect the appropriateness of the original revenue recognition by the seller.  If “an uncertainty arises about the collectability of an amount already included in revenue,” IAS 18 (paragraph 18) requires that “the uncollectible amount or the amount in respect of which recovery has ceased to be probable is recognised as an expense, rather than as an adjustment of the amount of revenue originally recognised.”  With respect to both of the VMS sales at issue, the accounting treatment that HP proposes is clearly inconsistent with IAS 18 and, therefore, wrong.

Q2 2009 (sale of software)

HP wrongly alleges that Autonomy’s sale of software licenses to VMS for $9 million on June 30, 2009, and Autonomy’s purchase of data licenses from VMS for $13 million on the same day, were “reciprocal” transactions “that as a matter of substance . . . should be analysed as a single transaction and accounted for on a net basis.”  As explained above, this claim is based on a misapplication of IAS 18.  HP appears to disregard paragraph 12 of IAS 18, which states that “when goods or services are exchanged or swapped for goods or services which are of a similar nature and value, the exchange is not regarded as a transaction which generates revenue.” (emphasis added).  VMS used Autonomy software products to power its core information-monitoring services, and Autonomy bought data feeds from VMS to power its portals and social media analytics products and demonstrations (since Autonomy’s existing provider was planning to switch off the data feeds).  In short, the respective software bought and sold was clearly not of a “similar nature.”

Further, contrary to HP’s characterization in paragraph 85.3 of the Particulars, the purchase was not limited to data feeds but also included a license to resell VMS’s normally chargeable service to Autonomy customers at no further charge.  The cost of Autonomy’s purchase from VMS was comparable to quotes from other suppliers.

These transactions and the related accounting treatments were reviewed by both the Audit Committee and Deloitte.  The auditors commented:

This is a $9 million deal to supply VMS with a perpetual licence for a suite of Autonomy software products including TeamSite, LiveSite, Qfiniti and IDOL.  Also in the quarter, Autonomy has separately purchased $13 million of software and associated services from VMS.  Given that there is clear commercial rationale for the separate transactions, separate contractual arrangements and evidence that both transactions have been made at fair value, management has confirmed and concluded that there are no links between the contracts that would impact the accounting. . . . [W]e examined the terms and conditions of the contract to ensure that no circumstances exist which might impact the recognition of revenue.[43]

Thus, contrary to paragraph 146.2 of the Particulars, Deloitte was plainly aware of both transactions and acquiesced in Autonomy’s accounting treatment.

HP mistakenly claims in paragraph 85.1 of the Particulars that a business plan for the purchase of the license to resell the VMS feeds to Autonomy customers was backdated.  Not only is this claim utterly irrelevant, but it also ignores the obvious explanation that any incongruity in the date could have been an innocent oversight.  Although HP has not given sufficient specifics, our preliminary analysis shows that the earlier date on the document could simply be a print header not seen on screen and inserted automatically when material is copied from an older document and pasted into a new one.  What is clear is that there is absolutely no evidence that anyone at Autonomy was attempting to pass off the business plan as having been created earlier, and HP certainly has provided none.

Q4 2010 (sale of software and hardware)

HP appears to assert that Autonomy’s sales of software and hardware to VMS in Q4 2010, for $4.8 million and $6.0 million respectively, were also part of “reciprocal” transactions and thus should be accounted for on a net basis.  Both of these sales and the related accounting treatments were reviewed by the Audit Committee and Deloitte.

In paragraph 86.2 of the Particulars, HP takes out of context a selective quote from an internal email concerning bonus eligibility in which an employee suggested that he could have obtained a higher price in the Q4 2010 VMS deal by more effectively selling the benefits to the customer. 

4.      Vidient

HP incorrectly states in paragraph 137.6 of the Particulars that Autonomy purchased a license and other rights for Vidient that Autonomy did not need or use.  To the contrary, the evidence will show that Vidient was available for resale by Autonomy and was bid into multiple deals, as well as being used in a project for a cleared government customer.

In sum, Autonomy’s purchases from its customers were entirely proper.  Following the acquisition these purchase volumes declined because HP sought to internally source Autonomy’s technology needs wherever possible.


HP’s allegations concerning hosting transactions are baseless.  First, as HP concedes, Autonomy’s accounting treatment is permissible so long as the licenses had independent value and the fair value of the licenses was correct—both of which are in fact true.  Second, HP’s senior management and accountants were well aware of the structure of these hosted deals and saw no issue with them.  Indeed, the evidence will show that senior members of HP management confirmed that Autonomy’s pre-acquisition accounting policy for hybrid hosted deals was correct.

As part of the due diligence, HP reviewed a number of the major hosting contracts that it has now removed from revenue.  Moreover, it appears that HP continued to structure hosted deals in this manner even after Autonomy’s senior management left HP.  The market and analysts, too, were well aware of Autonomy’s license plus hosted model—contrary to HP’s allegations that Autonomy’s published information on hosted transactions was misleading.

HP’s allegations regarding hosting are a transparent attempt to scapegoat Autonomy for the losses caused by HP’s officers and directors.  This is especially obvious given that HP’s revenue adjustments in paragraph 106 of the Particulars are based on the flawed assumption that in an era when storage costs were falling by orders of magnitude, customers would have renewed deals at the same prices from years earlier, the difference being some notional loss.  HP’s purported justification for removing these deals is almost comical—if a customer bought 1TB of storage five years ago for $100,000, but can purchase the same amount of storage today for $50 from you and for $49 from your competitor, is there any chance that the customer will be willing to pay $100,000 for this storage from anyone?

            A.        Autonomy Properly Accounted for Its Hosting Transactions

Under IFRS there are three general considerations when accounting for multi-element transactions such as hosting arrangements.  First, under IAS 18, a multi-element transaction may be separated into components so long as the separation reflects the substance of the arrangement (i.e., when those elements are, or can be, sold separately in the market).  Second, IAS 18.9 and International Financial Reporting Interpretations Committee (“IFRIC“) 13.BC14 require that the revenue for each element of such a transaction be allocated on a fair-value basis, which can be done in multiple ways.  Third, once separate components have been identified and the fair value of each is determined, the revenues associated with a particular component can be recognized when all requirements of IAS 18 are met for that particular component.

The licenses did have independent value; they could be used by customers and, as is well documented, they indeed were used by customers.  Additionally, the fair value of the licenses was correctly measured by application of the residual method, and revenue from the license sales was recognized in accordance with IAS 18.

Further, the evidence will show that Autonomy’s accounting treatment for these transactions was reviewed and approved by its Audit Committee, and that Deloitte reviewed these transactions and the fair-value analysis carried out by Autonomy’s finance team and ensured that all related accounting complied with IFRS.

1.      The Licenses had Independent Value; They Could be and were Used by Customers

After acquiring Zantaz, Autonomy embedded its own IDOL software into Zantaz’s hosted products to create a new product which increased functionality.  At the time, storage costs had dropped substantially as data storage had become increasingly less expensive.  As a result of these two developments, and in an effort to accommodate customers who might be interested in independently maintaining their own data, Autonomy began to offer customers, including former Zantaz customers, an opportunity to either purchase the new IDOL hosting software alone or the complete data archiving service (which included both the software and hosting component).

Under this model, customers could license the software product from Autonomy with the option to have that software and the associated data hosted by Autonomy, themselves, or another party at a location of their choosing.  The license plus hosted model was an attractive option for many customers who wanted protection from potentially sharp rate increases when the Autonomy hosting arrangement came to an end, or the option to independently maintain their own data.  Customers were concerned that, given the very large amounts of data involved and the legal and regulatory requirements, it was practically very difficult for them to switch out of Autonomy software.  As a result, when the fixed-term hosted agreement came up for negotiation, without ownership of the software, they would be in a weak negotiating position.  Ownership of the Autonomy software gave the customer leverage in negotiating a renewal contract with Autonomy as the customer had the possibility of hosting its data elsewhere.

In cases where customers did opt for a license of the software, the separation of components of hosted license sales correctly reflected the substance of the arrangement, as Autonomy sold the hosting and licenses separately.  Autonomy recognized the associated revenue as follows:  License revenue was recognized at the point the contract commenced and the software was delivered; after-sales support revenue was recognized ratably over the period of the license contract; and archiving revenue was recognized ratably over the contracted period of archiving.

HP’s suggestion that hosting customers could not use the licensed software independently is just plain wrong.  Customers could freely transfer their software to their own data centers, should they choose to move their data.  Various partners in fact shipped and used the software independently.  The evidence will show sales of the on-premise Digital Safe license to some of Autonomy’s largest customers and that Deloitte noted such sales.  And, contrary to HP’s somewhat irrelevant claim in paragraph 110 of the Particulars that Digital Safe had no value if only an Autonomy service team could install it on a customer site, there were a number of service companies who could and did install stand-alone Digital Safe at customer sites.   Similarly, Autonomy’s eDiscovery was available for on-premise purchase (i.e., a license) and was indeed sold in that form.

Contrary to HP’s purported analysis, under those circumstances it was entirely inappropriate to set the value of the license to zero, as the license had value in its own right and the customer was entitled to keep it at the end of the hosting period.

2.      The Fair Value of the License was Correctly Measured Using the Residual Method

Autonomy management appropriately determined fair value by using the residual approach of IAS 18.IE11, whereby the value of one element is the difference (or residue) between the total value of the sale and the fair value of another separately identifiable component.  Here, the value of the hosting element was calculated with reference to the standard rate for data storage.  Autonomy’s finance department analyzed the fair value of the hosting element on a reasonable basis, with the fair value of the license calculated based on the residual value of the complete contract.

Specifically, for each deal a cost analysis was performed to capture the costs of the hosting portion.  The analysis did not take into account the effect of the future precipitous fall in storage cost that was usual in the industry, and thus this model’s estimate as to the value of the license was conservative.  Although Autonomy was able to perform this type of work at a lower cost than HP—because of its advanced sector-leading technology, lower overhead costs and less bureaucracy, and reengineered Zantaz system (which had sixteen times more storage per server)—Autonomy’s hosting contracts did not use bargain-basement pricing, as evidenced by the fact that Autonomy was, on occasion, underbid by competitors.

Contrary to paragraphs 102.1, 102.2, and 107 of the Particulars, which assert that these deals should have been renewed at their original prices from three to five years earlier, prices were lower for the following reasons:

(i)              There has been a precipitous fall in storage costs (as every PC purchaser, and HP as a PC manufacturer, knows perfectly well); and

(ii)            The fact that Autonomy reengineered the Zantaz system to have sixteen times more storage per server, thus further reducing cost.

3.      These Transactions were Accounted for in Accordance with Autonomy’s Stated Accounting Policies

Autonomy recognized revenue on each element of these hosting agreements in accordance with IAS 18, as Autonomy deferred revenue associated with any undelivered product or ongoing obligation.  Further, Autonomy’s accounting treatment of its hosting transactions was reviewed and approved by its Audit Committee and by Deloitte, who ensured that all related accounting complied with IFRS.  Autonomy fully disclosed its accounting policies for hosted sales, regardless of how they occurred, and accounted for such sales in compliance with those policies.

HP is incorrect in arguing that Autonomy’s accounting treatment of the revenue associated with the Digital Safe and eDiscovery licenses for hosting arrangements was inconsistent with statements made in its published information regarding the nature of the revenue streams derived from hosting arrangements.  The hosting transactions included a hosted element which produced a revenue stream that amounted to a long-term annuity stream.

The accounting policies note to the 2010 Annual Report states:

Revenues from software license agreements are recognised where there is persuasive evidence of an agreement with a customer (contract and/or binding purchase order), delivery of the software has taken place, collectability is probable and the fee has been contractually agreed and is not subject to adjustment or refund (i.e. is fixed and determinable). . . . If significant post-delivery obligations exist or if a sale is subject to customer acceptance, revenues are deferred until no significant obligations remain or acceptance has occurred.

* * *

Product revenues from the hosted business are separated, using the residual method, into capture and archiving.  Revenues for capture are recognised in the period in which they are delivered.  Revenues for archiving are recognised over the period that the customers have access to the group’s software and proprietary storage technology.  Product revenues from the hosted business relate to the execution of production operations on computers which the company runs in its data centres.  Revenues are generated from the use of our software and computers and from us maintaining the data.  Customers commit for periods between one month and up to three years.  Revenues are generated in two different ways:

  • Paid up front, usually one to three years, in which case the group has an obligation to process the data using the group’s software and archive for the contracted length of time using the group’s proprietary storage technology.  Revenues are allocated between capture and archiving using the residual method, with the archiving element deferred and recogni[s]ed rateably [sic] over the contracted period of archiving.
  • Paid on a pay-as-you-go basis, in which case the charge is based on the volumes of data ingested and stored each month.  Revenues for pay-as-you-go customers are recogni[s]ed on a monthly basis as the product is made available to customers.  There is no deferred revenue in relation to these customers.[44]

HP could, therefore, have been under no misapprehension as to how Autonomy accounted for its hosted contracts.  Indeed, the five transactions highlighted in the Particulars are referred to in the Reports to the Audit Committee and were considered by Deloitte.  The evidence will also show that HP reviewed a number of contracts for Autonomy’s hosted deals during the due-diligence process and was therefore aware of the structure and terms of the deals.

B.      HP Continued to Structure Hosted Deals in this Manner After Autonomy’s Senior Management Left HP

HP Finance itself was made aware of, and approved, Autonomy’s hosting-related accounting.  Autonomy continued to structure hosting transactions post-acquisition in the same manner as pre-acquisition, and the appropriate accounting for such deals was considered and discussed by HP Finance and with other key HP personnel, who took no issue with Autonomy’s pre-acquisition hosting accounting.

HP’s 2011 Form 10-K disclosed HP’s accounting policy for hosting sales as follows:

In accordance with the specific guidance for recognizing software revenue, where applicable, HP recognizes revenue from perpetual software licenses at the inception of the license term assuming all revenue recognition criteria have been met.  Term-based software license revenue is recognized on a subscription basis over the term of the license entitlement.  HP uses the residual method to allocate revenue to software licenses at the inception of the license term when VSOE of fair value for all undelivered elements exists, such as post-contract support, and all other revenue recognition criteria have been satisfied.  Revenue generated from maintenance and unspecified upgrades or updates on a when-and-if-available basis is recognized over the period such items are delivered.  HP recognizes revenue for software hosting or software-as-a-service (SaaS) arrangements as the service is delivered, generally on a straight-line basis, over the contractual period of performance.  In software hosting arrangements where software licenses are sold, the associated software revenue is recognized according to whether perpetual licenses or term licenses are sold, subject to the above guidance.  In SaaS arrangements where software licenses are not sold, the entire arrangement is recognized on a subscription basis over the term of the arrangement.[45]

What this shows is that HP accounted for revenue from licenses sold in software hosting arrangements in the same way that Autonomy did.  While HP Finance ultimately departed from and changed Autonomy’s recognition policy, the evidence will show that it did so for business, not accounting, reasons.


HP is wrong in its assertion that members of Autonomy’s senior management falsely classified revenue as OEM revenue.  While it bears repeating that Dr. Lynch was not involved in the classification of revenue and acted in good faith in relying on others to determine the correct classification, Dr. Lynch always understood—and believes today—that the classification was appropriate.  The allegations set out in HP’s Particulars reflect a failure to appreciate the distinction between an OEM sale and OEM-derived revenue.  Once this difference is understood, it becomes clear that Autonomy’s classification of OEM revenue was proper.

While there is no official OEM revenue accounting definition, Autonomy properly disclosed the basis for its OEM revenue reporting in its accounts, and the market was aware that Autonomy’s OEM revenue figures captured all sales that “derived” from OEM sales.

HP appears to believe that OEM-derived revenue is (and can only be) revenue resulting from a software company (and only a software company) including IDOL in its own product and then selling on that product and paying Autonomy a royalty.  This is not correct.  An OEM sale is one in which a company licenses technology—in this case, Autonomy’s—in order to embed that technology in its own products or services.  Given the nature of Autonomy’s core product, IDOL, described by HP’s Ms. Whitman as “almost magical,”[46] other companies frequently sought to embed Autonomy’s technology in their own products through a variety of contractual arrangements.  OEM customers would typically license a small subset of the many functions in IDOL.

Following these sales, Autonomy would have the opportunity to “up-sell” additional functionality either to the original OEM purchaser (i.e., the entity that embedded a subset of IDOL functions into its own product) or directly to an entity that had purchased the product from an Autonomy OEM customer.  Autonomy considered and accounted for any follow-on license sale—whether to the original OEM customer or to an OEM customer’s customer—as OEM-derived revenue, because the follow-on sale occurred as a result of (i.e., was derived from) the initial OEM sale.  Such models were discussed extensively in analyst reports from the earliest days of the company, including analyst reports from the same firms used by HP for its valuation calculations.

As defined by Autonomy in its public reports, OEM-derived revenue included revenue made possible by Autonomy’s OEM program.  Thus, Autonomy did not have to transact with an OEM customer itself to generate revenue classified as OEM-derived revenue.  For example, if a customer bought software from Autonomy that linked to a product that already had Autonomy’s technology embedded in it, Autonomy could have recognized revenue from that sale as OEM-derived revenue.  Likewise, if Autonomy sold a connector that connected to a product that had Autonomy software embedded within it, this too could have been recognized as OEM-derived revenue.

This policy, which was reviewed by Deloitte, was entirely appropriate.  First, when Autonomy described the source of its OEM revenue, it appropriately disclosed that OEM revenue was generated in a variety of ways, including through license arrangements.  Second, when Autonomy reported OEM revenue figures, it properly disclosed that such revenues captured all sales that “derived” from OEM sales.  Third, Autonomy disclosed to the market that its classification of OEM-derived revenue was a matter of accounting judgment.  These disclosures ensured that the market understood the basis for Autonomy’s OEM revenue reporting.

VII.     Iron Mountain

When Autonomy acquired the digital archiving assets and technology of Iron Mountain, the terms of the deal provided for Iron Mountain to continue archiving certain data, for which it needed licenses to the relevant software.  Iron Mountain was therefore required to buy back a license for this purpose.  Contrary to paragraph 115.1 of the Particulars, Autonomy valued the OEM component of this transaction in cooperation with Deloitte and as described by Deloitte to the Audit Committee, as clearly demonstrated by the Report to the Audit Committee on the Q2 2011 Review:

Management has also performed a fair value analysis on a $1.5 million IDOL licence sale made by Autonomy to Iron Mountain at the time of the acquisition, which was deemed to be at less than fair market value.  In accordance with the paragraph B50 of IFRS 3 (2008) management has determined that this transaction is linked to the business combination.  Therefore, in accordance with paragraph 38 of IFRS 3 (2008), the asset transferred must be fair valued and included as part of the total consideration paid to Iron Mountain.  This has resulted in a $5.5 million increase to the fair value of the consideration paid (and therefore goodwill) and total post acquisition revenue of $7.0 million recognised.  Management has determined fair value with reference to seven similar sized licence deals.  An average licence value was calculated for sales where IDOL search was the core product offering.  This generated a value of $10.6 million, but did include one significant outlier.  Excluding this outlier the revised average was $7.4 million.  Management has therefore determined that an appropriate estimation of fair value is $7.0 million.

We concur that management’s treatment is in line with the guidance in IFRS 3 (2008) and that fair value has been determined on a reasonable basis, using a comparative group of similar sized deals to similar sized organisations.[47]

Once again, HP’s allegations to the contrary are without merit.


In 2010, an employee based on the West Coast of the United States raised issues concerning the treatment of reseller and so-called reciprocal transactions.  These issues were investigated and referred to the Audit Committee and Deloitte.  Contrary to paragraph 147.2 of the Particulars, upon being contacted by this employee, Dr. Lynch followed good corporate procedure and promptly referred the matter to the head of the Audit Committee and the non-executive Chairman of the Board.  Dr. Lynch’s involvement in the investigation was in response to the Audit Committee’s requests for information.  The issues raised were the same or very similar to those that HP is now repeating.  Deloitte appointed a team, which Deloitte itself referred to as “independent,” to review the issues.  After review, Deloitte concluded that the employee’s concerns were based on partial information and a misunderstanding of the relevant accounting rules, and confirmed that all of the transactions about which the employee had expressed concern had received appropriate accounting treatment.

In its Report to the Audit Committee on the Q2 2010 Review, Deloitte stated:

On the basis of management’s detailed review of the matters raised . . . they have concluded that there is no new material information provided that would have affected their key judgments taken at the time of preparing and presenting the 31 December 2009 financial statements.  Management have therefore concluded that these matters raised have no material impact on the Group’s 2009 financial statements.  Based on our discussions with management, and the review work we have carried out, we concur with this view.[48]

Additionally, the employee referred his issues to the FSA in London, which in turn referred the issues to the FRRP, whose remit includes setting the standards framework within which auditors, actuaries, and accountants operate in the United Kingdom, while monitoring the implementation of these standards and promoting best practice.  The FRRP investigated the very issues and allegations that HP now advances concerning resellers and so-called reciprocal transactions and, in August 2011, declined to pursue any of the matters further.[49]  HP conveniently fails to mention this.

The evidence will show that the employee did not understand IFRS and that his conclusions were based on partial information.  Contrary to paragraph 147.6 of the Particulars, the employee was terminated by the company’s independent chairman.

In sum, the issues raised by HP have already been analyzed, considered, and rejected not only by Autonomy’s auditors and an independent internal committee but also by the responsible regulatory body, which is uniquely positioned to understand and investigate these issues.


All of the matters of which HP complains were transparently disclosed to Deloitte, as recorded in the quarterly reports to the Audit Committee.  Further, many of these matters were then reviewed by the FRRP, the regulatory body directly responsible for monitoring the implementation of proper accounting standards and practices in the United Kingdom.  None of these entities found any problem with Autonomy’s accounts.

Nonetheless, HP purports to be the victim of a $5 billion fraud (while making no attempt to explain the basis of this extraordinary loss and while failing to reconcile its position with the fact that none of the cash is missing).  In reviewing HP’s latest filings in the UK Action, it is clear that for numerous deals:

  • Revenue has been reversed on the basis of differing accounting judgment, different accounting policy, and incorrect application of hindsight.  In fact, whole swaths of revenue have been removed without any individual analysis of the deals. HP fails to explain how any of these deals were fraudulent because it cannot.  For example, HP has asserted that Autonomy’s resellers were not on risk but has produced no direct evidence to support that assertion, much less to rebut the wealth of written evidence confirming that the resellers were, in fact, on risk.
  • HP asserts that the reciprocal deals were somehow of no value, but again provides no actual evidence in support of this assertion, when in many cases the purchases were in full use and often sold on at a profit to Autonomy customers.
  • HP asserts that Autonomy’s purchases from its customers were not made at fair value, but provides no evidence to counter the audit evidence considered by Deloitte at the time.
  • HP asserts that it did not know about the hardware sales, but its own submissions to the US courts show that it did.

For all these reasons, the evidence will show that HP’s allegations are baseless and that its filing of the Particulars in the California Action was an improper attempt to generate publicity for its specious claims, avoid disclosure and engagement on the merits, and persuade the Court to approve its collusive settlement.

Very truly yours,

Christopher J. Morvillo
Clifford Chance US LLP

Reid H. Weingarten
Michelle L. Levin
Steptoe & Johnson LLP

Attorneys for Dr. Michael Lynch

[1]      HP’s concerted effort to deny Dr. Lynch and his counsel access to the documents underlying the most basic allegations in the Particulars makes it impossible for Dr. Lynch to fully respond at this time.  As a result, Dr. Lynch reserves the right to amend or expand his responses contained herein in connection with his formal response to the Particulars in the UK Action, which will be provided once the full record of relevant documents is produced and Dr. Lynch has had an adequate opportunity to review this information.  Additionally, much of the information contained in this response—including the analysis of the applicable accounting principles—is the result of investigation that has occurred subsequent to HP’s public disclosure of its allegations.  Accordingly, nothing in this letter constitutes an admission or acknowledgement that Dr. Lynch was aware of the particular fact or accounting principle prior to HP’s announcement of its allegations in November 2012.

[2]      Katherine Rushton, HP Boss Meg Whitman Admits Autonomy Row Hit Morale, The Telegraph (Apr. 10, 2013), http://www.telegraph.co.uk/finance/newsbysector/mediatechnologyandtelecoms/electronics/998427

[3]      Quentin Hardy & Michael J. de la Merced, Hewlett’s Loss: A Folly Unfolds, by the Numbers, N.Y. Times (Nov. 20, 2012), http://dealbook.nytimes.com/2012/11/20/h-p-takes-big-hit-on-accounting-improprieties-at-autonomy

[4]      See Murad Ahmed, Lynch Fires Back at HP’s ‘Excessive’ Forecasts on Autonomy Deal, Fin. Times (Sept.11, 2014), http://www.ft.com/cms/s/0/671d5764-39c8-11e4-8aa2-00144feabdc0.html#axzz3Z7HD6vFl.

[5]      Contrary to paragraphs 18.1, 18.2, and 19 of the Particulars, Dr. Lynch did not serve as president of Autonomy Inc. during the period, nor was he a shadow director or “de facto director” of any of the various subsidiaries named by HP.  The argument that the group CEO of a FTSE 100 company should be deemed to be a shadow director or “de facto director” of the group’s subsidiaries by virtue of his CEO group-level role is not credible, and even less so when some of those subsidiaries are based 6,000 miles away with an eight-hour time zone difference and have their own CEO, COO, CTO, CMO, and GC.

[6]      In the United Kingdom, interim financial reporting is governed by International Accounting Standard (“IAS“) 34 and particular listing rules, and quarterly reports (i.e., those for Q1 and Q3) are governed solely by particular listing rules.  As a result, Autonomy’s quarterly reports were governed by the rules associated with the London Stock Exchange.

[7]      Autonomy Corporation plc Announces Interim Results for Six Months Ended June 30, 2011 at 20.

[8]      See, e.g., Report to the Audit Committee on the Q4 2010 Review at 5.

[9]      For example, the American Institute of Certified Public Accountants Audit Guide states: “The pressure to meet quarterly or annual earnings expectations creates a strong incentive for entities to complete transactions by the end of the reporting period. . . .  Customers can take advantage of this desire to meet revenue expectations by forcing software vendors to lower prices or provide more liberal sales terms in contracts negotiated near the end of a reporting period.  For these reasons, it is common for software vendors to have a significant number of sales near the end of a reporting period.”  American Institute of Certified Public Accountants Audit Guide: Auditing Review in Certain Industries [AAF-REV] 2011, at 2.14.

[10]     Autonomy Corporation plc, Annual Report and Accounts for the Year Ended 31 December 2010, at 19.

[11]     Id.

[12]     Contrary to paragraph 53.1.3 of the Particulars, Autonomy did sell some hardware in 2008 (and indeed earlier), and the market was aware of this.

[13]     There is reason to believe that HP adopted a similar strategy, selling software at a discount in order to drive its primary hardware sales.

[14]    The fact that sales of EMC hardware could be beneficial not only to Autonomy and its customers, but also to individual members of Autonomy staff if they reached their performance targets, apparently strikes HP as offensive.  See paragraphs 135.5–135.6 of the Particulars (discussing a proposed bonus to incentivize Mr. Sullivan in his negotiations with EMC).  However, the details of an individual bonus—not to speak of an individual email or its compatibility (or lack thereof) with HP’s sense of humor (or lack thereof)—certainly do not suffice to establish a years long systematic fraud on the order of billions of dollars.

[15]     Aaron Ricadela & Amy Thompson, HP Plunges on $8.8 Billion Charge From Autonomy Writedown, Bloomberg (Nov. 20, 2012), http://www.bloomberg.com/news/articles/2012-11-20/hewlett-packard-profit-forecast-8-8-billion-charge; Press Release, HP Issues Statement Regarding Autonomy Impairment Charge (Nov. 20, 2012), http://www8.hp.com/us/en/hp-news/press-release.html?id=1334263.

[16]     DRC Report at 40 (emphasis added); see also DRC Report at 57 (“E&Y noted that Autonomy had sold material quantities of hardware . . . [and] reported its observation to HP management.” (emphasis added)).

[17]     DRC Report at 39 (emphasis added).

[18]       Murad Ahmed, Lynch Fires Back at HP’s ‘Excessive’ Forecasts on Autonomy Deal, Fin. Times (Sept.11, 2014), http://www.ft.com/cms/s/0/671d5764-39c8-11e4-8aa2-00144feabdc0.html#axzz3Z7HD6vFl.

[19]       Lisa Fleisher, Former Autonomy Execs Turning to Unusual Strategy in Fight with Hewlett-Packard, Wall St. J. (Sept. 12, 2014), http://blogs.wsj.com/digits/2014/09/12/former-autonomy-execs-turning-to-unusual-strategy-in-fight-with-hewlett-packard/. (emphasis added).

[20]     Further, contrary to paragraph 142.9.2 of the Particulars—which states that there was no strategic partnership between Autonomy and Dell, and no development by Dell of an appliance consisting of Dell hardware and Autonomy software—the evidence will show that Autonomy and Dell worked together to develop an appliance, an appliance was marketed and sold, and Dell proved to be a helpful partner.  Discussions in November 2009 included technical meetings on the appliance subject, demonstrations to senior executives at Dell, and OEM arrangements regarding Dell products.

[21]     Even if HP’s novel meaning of “pure” is taken as true (without admission and purely for the sake of argument), excluding some obviously inapplicable examples alone affects the alleged numbers considerably.  By our estimates, and contrary to the statements in paragraph 61.1 of the Particulars, in 2009 the amount of loss-making so-called “pure hardware” would have been less than six percent of revenues, and in 2010 the amount of loss-making so-called “pure hardware” would have been less than nine percent of revenues.

[22]     Management’s Note to the Audit Committee on the Q3 2009 Review.

[23]     Management’s Note to the Audit Committee on the Q3 2010 Review.

[24]     Report to the Audit Committee on the Q2 2010 Review at 6 (emphasis added).

[25]     Contrary to paragraph 61.4 of the Particulars, inclusion of hardware in the gross-margin calculation could be viewed in various ways depending on the relevant period.  For example, its small effect is to deflate, not inflate, apparent growth in the period H1 2010 to H1 2011.  Similarly, the inclusion of the hardware revenue reduced the amount of profit.

[26]      For example, during the April 21, 2010, analyst call, management stated: “[W]e’d expect to see [gross margins] continue to fluctuate by plus or minus 2 percentage points, as many of you are used to seeing over the years.” And during the February 1, 2011, analyst call, management noted that “there are natural fluctuations in the gross margin rate quarter on quarter” and added: “One thing I would be very careful about, for those of you that have watched Autonomy for many years, you will be used to the fact that there is, on a quarterly basis, pretty damn random plus or minus 2% on the gross margin.”

[27]    As Autonomy’s 2010 Annual Report explained:  “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.”  Autonomy Corporation plc, Annual Report and Accounts for the Year Ended 31 December 2010, at 13.  And, as further noted in the same document:  “Autonomy operates a rare ‘pure software’ model under which our goal is that most implementation work is carried out by approved partners.” Id. at 16.

[28]     HP Software PartnerOne EMEA Partner Program Guide 2013: Terms and Conditions at 20, 27.

[29]     See Arik Hesseldahl, What Exactly Happened at Autonomy?, Wall St. J. (Nov. 20, 2012), http://allthingsd.com/20121120/what-exactly-happened-at-autonomy/ (quoting Ms. Lesjak as saying, “The VAR sales were reported as licenses, and they weren’t, in some sense, real sales, because there was no end user.”).

[30]     For example, the Report to the Audit Committee on the Q2 2010 Review stated, at page 5: “We highlight to the Audit Committee that there is a large licence deal on which we reported in Q1 2010 with Microtech where the end user is the Vatican ($11.5 million).  . . .  As part of our Q2 2010 procedures we received confirmation from Microtech of the amount outstanding and the absence of any side agreements or ongoing performance criteria.”

[31]     Report to the Audit Committee on the Q1 2010 Review at 2–3.

[32]     As a general matter, and contrary to paragraph 144.1 of the Particulars, it should be noted that although Autonomy referred internally to reseller deals by the names of the intended end users, as a way of differentiating among multiple deals with the same reseller, the Reports to the Audit Committee make clear that the Audit Committee and Deloitte were well aware that the customers in these deals were resellers, not end users.

[33]     Contrary to paragraph 77.9 of the Particulars, Capax Discovery has continued to support Nearpoint, and on November 4, 2014, announced a new Nearpoint roadmap.

[34]     Letter from the FRRP to Sushovan Hussain ¶ 5 (Aug. 22, 2011).

[35]     Report to the Audit Committee on the Q3 2009 Review at 3.

[36]     Letter from the FRRP to Sushovan Hussain ¶ 7 (Aug. 22, 2011).

[37]     Id. ¶ 9.

[38]     MicroTech leased office space in Virginia and, as the proposal made clear, equipped the facility with advanced technology, including extensive multiprocessor server farm, storage, and other cloud related systems; software infrastructure; multiple displays; a security system; virtual enterprise management software; and other advanced components.

[39]     Report to the Audit Committee on the Q4 2010 Review at 18.

[40]     Report to the Audit Committee on the Q4 2009 Review at 4.

[41]     Specifically, Deloitte stated:

The most significant revenues recognized in the quarter were with . . . Filetek Inc ($7.5 million) . . . . Autonomy has separately purchased $10.4 million of software and associated services from Filetek Inc (“Filetek”) in the quarter.  Given that there is a clear commercial rationale for the separate transactions, separate contractual arrangements and evidence that both transactions have been made at fair value, management has confirmed and concluded that there are no links between the contracts that would impact that accounting.  The cost of the software purchased by Autonomy has been capitalised on the balance sheet as an intangible asset and is to be amortised to the income statement over its useful economic life of five years.

Id. at 3.

[42]     Report to the Audit Committee on the Q2 2010 Review at 10.

[43]     Report to the Audit Committee on the Q2 2009 Review at 3–4.

[44]     Autonomy Corporation plc, Annual Report and Accounts for the Year Ended 31 December 2010, at 51.

[45]     Hewlett Packard Co., Annual Report (Form 10-K) at 81 (Oct. 31, 2011) (emphasis added).

[46]     In the words of HP’s CEO, Meg Whitman, in an interview published on April 10, 2013, IDOL software is “almost magical technology.  What it allows customers to do [is] to understand all the unstructured data, the application to legal and compliance—it is terrific technology . . . .”  Katherine Rushton, HP Boss Meg Whitman Admits Autonomy Row Hit Morale, The Telegraph (Apr. 10, 2013), http://www.telegraph.co.uk/finance/news

[47]     Report to the Audit Committee on the Q2 2011 Review at 3.

[48]     Report to the Audit Committee on the Q2 2010 Review at 18.

[49]     Letter from the FRRP to Sushovan Hussain ¶¶ 10, 21 (Aug. 22, 2011).



HP’s allegations have changed significantly since its announcement of the write-down on 20th of November 2012. As HP’s initial false interpretation of the relevant accounting standards has been replaced by a more sober understanding of the facts, it has attempted to retrofit a new set of allegations to justify its massive write-down. Gone are the original accusations that hardware was booked as software and costs were not counted, as well as the idea that the reseller deals were not real (which has been replaced with mere timing allegations). The hosted and OEM allegations likewise cannot support any credible claim of loss. And the tiny number of deals at issue could not, under any circumstances, justify a $5 billion write-down.

In short, even the latest carefully crafted allegations—drafted with the benefit of a nearly three-year investigation—make no case for fraud. Rather, they evidence, at best, a difference of opinion over accounting policies (as exemplified in the rebasing exercise reported by the FT)—a disagreement in which two Big Four firms, Deloitte (Autonomy’s auditor) and Ernst & Young (HP’s firm) agree that there is nothing wrong with the pre-acquisition financials.

The allegations break into four categories:

1)      There should have been more disclosure around hardware sales and gross margin should have been accounted for differently

Autonomy made all required disclosures. Hardware was not a separate business unit and there was no requirement that it be separately disclosed. The gross margin judgement was reasonable and was cleared by Deloitte. And, in any event, the differences in gross margin alleged by HP, which result in at most a variance of 4%, could not have materially affected the valuation of the company.

Many sources made clear that Autonomy sold hardware, as HP well knew. HP continued to make what it calls ‘pure’ hardware sales after the acquisition, and HP’s accountants approved the practice. HP’s claims about what it knew and how it valued Autonomy are contradicted by the evidence and by common sense. This is especially true given that HP’s own valuation of Autonomy considered hardware sales.

Note: we have moved away from the original allegations of hardware booked as software and that hardware revenues should not be counted as revenue. Now that it has emerged that HP knew about hardware, the argument has shifted to the sales being of the wrong type of hardware.

2)      Some Resellers were not on risk and thus the deals should not have been counted until the cash was received (which it was)

HP has detailed its allegations as to a grand total of just 6 reseller deals with an aggregate value of approximately $40m (out of the $5 billion write-down). And even with regard to these 6, HP provides no actual evidence to support its claim that the resellers were not on risk. Indeed, the documentary evidence explicitly shows that the resellers were on risk and that almost all the cash was received on these deals. Nor can HP’s claims of acceleration be credited, as they rely on a failure to understand that under IFRS, the reseller—not the reseller’s intended end user purchaser—is Autonomy’s customer.

Note: we have moved away from the original allegation of the issue being no sell through or problems with the end user. Now it is solely an issue of timing of revenue recognition as Autonomy received cash on almost all of the deals.

3)      Autonomy bought items from customers that had little value or use to Autonomy

HP claims that Autonomy channelled funds to 5 of its reseller customers by buying goods from those customers that Autonomy did not value or use. This allegation pertains to about $50 million in transactions, which, as the schedules show, encompass essentially every product Autonomy ever bought from a customer. In other words, HP has concluded that Autonomy never had a valid reason to make a purchase from any of its customers—despite the fact that HP makes such purchases from its customers as a matter of course. In reality, all items that Autonomy purchased from customers were checked by Deloitte for fair value, and most of the items are mis-described by HP and are well documented as having been used or sold on at a profit.

4)      Hosted deals were structured improperly

At the time of the acquisition, HP and the market were well aware of how Autonomy structured its hosted deals: many analysts covered the topic and it was discussed in Autonomy’s annual reports. Deloitte approved the structure of such deals. HP reviewed several hosting contracts pre-acquisition, and HP’s own accountants agreed with the method and used it themselves at HP. All of the cash was received on these deals, which were mainly with blue-chip clients.

Notably, HP does not claim that the hosted deals were not real; rather, it argues that Autonomy should have renewed all of these deals as service contracts only, rather than including a license component, on the theory that the former would have brought a higher price. This allegation neglects the fact that storage costs had declined sharply in recent years, making it impossible to maintain historical prices on hosting-service contracts. In any event, Autonomy’s pricing was a commercial decision well within management’s reasonable discretion, and HP saw some of the contracts in due diligence and thus cannot seriously contend that it was unaware of Autonomy’s pricing policies.

5)      OEM-derived sales were overstated

HP purposefully confuses OEM and OEM-derived revenues by, for example, only counting revenue from software companies as recognizable OEM-derived revenue. The definition Autonomy used for OEM revenue is included in the annual report and is stated as an estimate. HP has no credible basis for its allegation that Autonomy’s OEM-related accounting violated IFRS or was somehow fraudulent.

Key points:

HP has proffered no explanation as to how its allegations lead to a $5 billion write-down, especially as the allegations have no impact on the cash flow, which was the basis of HP’s valuation of Autonomy.

The comments to the schedules make clear the real reason deals were removed: accounting policy differences, not fraud. Whole classes of deals were removed, as in the rebasing exercise, with no individual analysis. Deals were removed for reasons of hindsight, for example where HP failed to conclude an implementation and had to repay customers—but hindsight is not the applicable accounting standard, as HP well knows.

HP is silent on the fact that all these matters were reviewed and agreed with by Deloitte.

There are many factual inaccuracies, such as statements regarding Autonomy’s use of purchased items, which can easily be rebutted.


Autonomy’s business was real and customers paid for it. If all of the revenue challenged by HP is removed, the result is more than $400 million in surplus cash over the period. If not from legitimate business, then where did this cash come from?

Autonomy was valued by HP on cash flow, and the cash flow has not changed under HP’s allegations.

Almost all of the challenged items were well documented and reviewed by Deloitte. However, the allegations hardly mention Deloitte, much less explain why Deloitte’s conclusions are being challenged. Deloitte is not being sued and stands by its work.



After three years of Meg Whitman’s stonewalling, is this it?

HP’s claim is finally laid bare for what it is – a desperate search for a scapegoat for its own errors and incompetence.

The contents of the claim are a simple re-hash of previous leaks and insinuations that add up to one long disagreement over accounting treatments, and have nothing to do with fraud.

Looking at the detail, out of 23,000 deals and $2.2 billion in revenue that took place at Autonomy during the period, Whitman makes assertions concerning only six specific reseller transactions, which add up to under $40million in revenue, and for which almost all the cash was, in fact, received. Whitman provides scant supporting evidence even for these few assertions. She raises questions over fully-paid hosted deals, the form of which HP’s own accountants approved, and in a structure HP itself also used. She now admits that hardware sales were correctly recorded as revenue and that HP was fully aware of this when they valued the company at $11 billion. Whitman also questions purchases from customers, despite the evidence showing that Autonomy bought from them at the correct price and for good commercial reason.

There are two critical points she conveniently ignores:

  1. These matters were all reviewed by Deloitte, and Autonomy’s audit committee, and involved hundreds of other people besides. Moreover there is no money missing. Yet Whitman claims this is the secret work of two people, working alone, that no-one could have possibly discovered: this is simply not possible.
  2. HP paid $11 billion for Autonomy. Meg Whitman’s claim blames HP’s loss on overpaying for Autonomy based on public information it says was false. The stock market used the very same information to value Autonomy at the time at $6 billion. HP valued Autonomy internally at $17 billion, based on its own estimates of how it could combine Autonomy with HP to create something new. After it bought the company, that plan was abandoned and HP ran Autonomy into the ground. That is the single and simple cause of HP’s loss.

Meg Whitman has been playing a delaying game, promising a smoking gun that has never materialised, hoping to confuse people by misstating IFRS rules, smearing individuals and avoid being called to account. Now that she is forced into being specific, that waiting game is coming to an end and we look forward to hearing her answers in court.



Click here to download a searchable PDF