Investors' Question Board

What is the expected impact of the convertible bond on 2010 consensus expectations?

The factors to consider are as follows:

Did Autonomy make any sales to MicroLink in Q4'09?

No. There were no Q4'09 revenues from MicroLink.

What kind of contribution should I be modelling for MicroLink in Q1'10 and for the full year?

For Q1'10 analysts seem to be modelling less than $1-2m revenue contribution. This gives a de minimis EPS impact in Q1'10 at c. -0.005 cents, and for the full year we would expect less than $7-8m as a base case contribution to revenues (EPS +1 cent). This takes account of the non acquired parts of the business, the stub effect and the fact that run rate Autonomy resell business is not double counted.

I see on the website that Intangibles and PPE reconcile to opening and closing balance sheets apart from $3m, which can be explained by FX/accruals. However, it doesn't seem to work for the individual asset classes. Take opening intangibles of $401m, subtract amortisation of $18m, add capitalised R&D of $6m – that leaves $389m. But the balance sheet figure is $10m higher than this. The same calculation for PPE gives you $7m going the other way and hence the $3m difference overall. The only thing I can think of is that nearly all of the $11m of CapEx was spent on intangibles, is that correct?

A/ Please see questions dated 3rd February, which point out how the calculation needs to be done, including the factors impacting tangible and intangible assets and stating that there was indeed capex spend on software (for example: for data centre security etc), which was around $8m.

On the 4Q09 organic growth calculation I can get to 18% by adding $8m currency headwind to the $223m, and dividing by last year's $145m plus $70m for IWOV less $19m of discontinued. Is that the right way to look at it? If so, what are the average dollar rates you are using in the two years to get the $8m currency impact for the fourth quarter?

There is no need to adjust for FX movements. There was simply a minor typo on one of the recent press releases where 2008 and 2009 were transposed. However, the actual organic growth calculation has been posted on the question board since the acquisition of Interwoven, and has not changed.

I asked the company at the analyst briefing about the $216m cash collected in 4Q09 and whether it was a net figure. If I add 5% for sales tax it gives me $227m for gross receipts. Is that right? Also, if you take $216m from the quarter's sales of $223m to get $7m, and add that to opening debtors of $219m that gets me to $226m. This is $4m less than the $230m reported which is presumably due to a weaker dollar?

You need to remember to compare like items – net cash collected with net receivables. You are correct that the cash collected figure of $216m in 4Q09 has been adjusted for sales tax. However, the attempt to reconcile opening and closing debtors on the balance sheet misses a number of important factors: adjustment for support and maintenance attached to sales made in the quarter, other support & maintenance invoices from prior year sales, bad debt provisions and FX.

Why were debtors up $78m per the 4Q09 cash flow statement (or 44% on $178m at end 2008) when organic growth in the business was just 16%? Bearing in mind that acquired debtors do not have any impact on working capital in the cash flow statement?

Again you need to remember to compare apples with apples. The balance sheet movement in trade receivables reflects the combined business of Autonomy and Interwoven and as a consequence you need to compare pro forma revenues with the growth in trade receivables. Pro forma revenue growth in 2009 was approximately 55% and this is in-line with trade receivable growth of 63%.

I would like to question what Sushovan Hussain CFO, said at the analyst briefing about EBITDA growth being the biggest influencer of cash conversion. I think this is wrong. It is sales growth that is important because this is what increases debtors and absorbs working capital when you are a growing company. Imagine a company with no sales growth, but that manages to double its EBITDA by cutting costs. EBITDA goes up by 100% but cash conversion is unaffected ie EBITDA growth is irrelevant to cash conversion in this example.

It's important to appreciate the determining factors in high growth, high operating margin businesses are different from traditional businesses. There are a number of factors that can impact the working capital requirement of a growing business, but the principle one is growth in revenues coupled with its effect on EBITDA growth in the period under consideration. The reason we feel your example is inappropriate is that it misses the fact that the only way a company can achieve EBITDA growth on flat revenues is by extensive cost cutting. Consequently, it will still have a negative working capital requirement but one driven by a reduction in payables, rather than an increase in debtors, and thus cash conversion will also likely be below 100%. However the change in CFFO due to the cost cutting is coming from not paying out existing cash, this is a different timing effect to receiving the extra cash from extra sales.

However, restricting ourselves to the case of two companies BOTH GROWING revenues ie similar to the case for Autonomy, let's analyse the determinant of cash conversion...

Company A is growing sales and EBITDA due to natural operating leverage in the business (ie costs are flat to up):

Cash conversion is affected by timing differences as the traditional cash conversion calculation (CFFO/EBITDA) does not capture the lag for growth. In a perfect company with perfect cash collection, but taking 90 days to be paid, the CFFO in any quarter is generated by profits in the prior quarter.

If for example the profits have doubled between Q3 and Q4, the CFFO in Q4 is generated by Q3 profits, which is half the value, so by definition the cash conversion is 50%, although the company is converting all its profits to cash. In the example you gave if the company is paid in 90 days the cash conversion would be 50% in this specific quarter. So for high EBITDA growth companies (driven by revenue growth) this is the effect that dominates. In fact it is purely mathematical.

Company B is growing sales but not EBITDA. EBITDA is thus similar in both quarters. What we need to consider is that cash conversion effectively becomes approx 1 + NWC/EBITDA as there is no EBITDA growth. The cash conversion would therefore still be below 100% (due to the debtor growth leading to negative NWC) but conversion rate would be higher than in scenario A as the EBITDA values in the two quarters are more similar.

Conclusion: For high operating margin companies generating growth in revenues it is the EBITDA growth that can become the dominant factor in determining the maximum possible cash conversion in any particular quarter, not the sales growth per se. In the case of a company like Autonomy with strong operating margins and high EBITDA growth generated from revenue growth, this effect dominates the others. This is a well understood effect for purer, high growth software companies and is about timing of cash flows.

Please explain the basis for the MicroLink acquisition - is it a low margin business?

Autonomy does not have all the required clearances to bid other than indirectly for some government contracts. Despite our technology being key to certain systems this bidding chain means we have to give up value (in some cases as much as 70%). This acquisition will allow us to supply directly, increasing our value in these deals and simplifying interactions with the end customers allowing improved support and customer service. The low margins come about on the resell of our software by MicroLink and the margin they make on it. This effect is not seen going forward as we will now be selling with full margin.

I have seen an additional research note that comments specifically that Fixed Assets have fallen between 3Q09 and 4Q09 yet CapEx was high in the quarter. How can this be?

The commentary you have seen forgets that the CapEx line in the cash flow statement does not only relate to tangible asset spend such as computer hardware but can also include intangible asset investment in items such as software for security to run in a data centre. This is a straightforward mistake and the calculation is wrong as a result.

We continue to hear questions over why goodwill increased in 4Q09 if there was no acquisition?

We answer this in the question below. But to reiterate, the increase is due to the increase in Deferred Tax Liability of $28.0m with the small remainder being FX adjustment as our functional currency is Sterling.

Please could you explain, or give a quick reconciliation, between the $630 million spent on Interwoven and the original Enterprise Value calculation ($590m) in terms of the banking fees, the additional charge for stock, etc.?

For full details of the reconciliation to the end of 3Q09 please see the question "Can you confirm whether the difference in acquisition costs for Interwoven reported between the second and third quarters relates to restructuring costs?" dated 27th October below.

The additional $1.5m of acquisition related spend in 4Q09 relates to the release of Interwoven self help provisions that were taken by the prior management team of Interwoven before the acquisition.

I notice that CapEx has increased over the last couple of quarters. What interests me is that this quarter it is higher than usual while the Fixed Asset number has reduced slightly from 3Q to 4Q09. Please can you explain what is happening?

Taking 3Q09 aggregate Fixed Assets and Intangibles of $439.1m and adjusting for CapEx of $16.6m (fixed asset and product development costs) and depreciation of -$25.5m gets you to a 4Q09 asset value of $430.2m. The small difference between this and the actual 4Q09 value of $433.3m relates to FX movements and accruals.

Please can you explain the $27m increase in Goodwill in 4Q09?

It has been suggested that the increase occurred due to a small acquisition during the quarter. That is not the case and we reiterate there were no acquisitions at all during the quarter per the 4Q09 press release. In fact, Goodwill increased during 4Q09 due to the recognition of additional deferred tax liabilities following the completion of the preliminary purchase accounting for Interwoven, specifically the finalization of the review of acquired tax losses and purchased intangibles.

The deferred tax liability also increased by a similar amount (4Q09: $85m up from $57m in 3Q09). Lastly, although Autonomy reports in US dollars, the functional books of the company are in sterling so there is a small FX effect being the balance.

We have heard a rumour that you have changed your accounting policies relating to the amortization and depreciation rates of your fixed assets between 2008 and 2009. This is based on a reduction in the D&A rate as a percentage of fixed and intangible assets between 2008 and 2009: The inference being you are overstating profitability. Is this the case?

There has been no change in accounting policy, no change in the treatment of tangibles or intangibles assets, nor any change in the amortization or depreciation rates applied. Any type of simple ratio analysis looking at depreciation and amortization rates as a percentage of net fixed assets (tangible plus intangible) is useless as it will not consider the mix of assets, the gross asset base, or where we are in the life of the underlying assets.

The amortization rates we use are stated in our 2008 R&A (page 33):

"Other intangible assets excluding goodwill are measured initially at purchase cost and are amortised on a straight-line basis over their estimated useful lives, on the following bases:

Patents and trademarks 3 years

Software licences 3 years

Purchased intangibles 3 - 12 years"

The acquisition of Interwoven in 2009 has meant that the percentage of "purchased intangibles" has increased from 67% to 80% as a proportion of total fixed and intangibles assets, and the effective age of the assets has fallen. Consequently you would expect the amortization rate as a percentage of the net balance to fall - as has been the case.

We have heard a rumor that the reduction in DSOs from 97 days in 3Q09 to 88 days in 4Q09 was due to factoring of receivables. Is this the case?

Definitely not. The reduction was due to extremely strong cash collection in 4Q09. In fact this was $216m versus 3Q09 revenues of $193m.

In 3Q09 working capital movements in the cashflow statement included a $44m positive payable swing. Could you please explain this movement and explain how it was impacted by the additional product spend that appeared much lower at $24m?

There has been some commentary that has questioned whether because the additional product spend in 3Q09 was disclosed as $24m, then the difference of $20m (between $44m in payables movement in the cash flow and the $24m itemized non-ongoing spending in the P&L) was somehow benefiting the P&L and not recognized as OPEX within the quarter thus over-stating operating margins in 3Q09. This is wrong with margins in 3Q09 reflecting all P&L costs, some of which in fact had not yet been invoiced. In accordance with IFRS, the cashflow movement in the working capital category must reflect all the P&L movement in the same quarter.

As previously disclosed around $24m new product related launch spend was spent late in the quarter. Additionally, there was also good working capital management of payables in Q3. However, as stated in the conference call, the positive benefit will reverse in Q4 as the invoices for the 3Q spending are received and paid in 4Q09.

I have noticed from the last few press releases that there is a growing focus on media and online retail type use cases. Is this a sign of shifting emphasis for your business, or perhaps renewed focus on a Meaning Based approach in this segment of the market?

It is still too early to tell whether there is any major trend emerging as yet, but we do believe that Meaning Based Marketing (MBM) will be a major growth area for Autonomy in a cyclical upturn, in addition to sustained demand for regulatory and infrastructure technology. The immediate and obvious return on investment these MBM solutions generate makes it a very simple sales pitch as retailers look to capture increasing consumer demand. In addition, we suspect that many companies will not look to reverse the reductions in employee headcount they made during the recession, but will instead seize the opportunity to put technology in place to automate processes and maximise efficiency.

I have seen two pieces of commentary on the 3Q09 results that I do not understand: The first is that the 3Q09 new product launch spending of around $20m, that was expensed in the P&L but was still a payable, may now have been cancelled in Q4. I do not understand why or how this would be done? Secondly, in a separate comment it was speculated that the extra cost in 3Q09 was not new product related spend but actually related to provisions taken against deteriorating receivables thus implying low quality revenue. It was then suggested that there is a risk of future write-offs of this implied low quality revenue?

Both premises are factually inaccurate. There has been no cancellation of payables in 4Q09, and there was no such increase in the provisions for doubtful accounts in Q3 (as there has been no deterioration in receivables in 2H09). As a consequence, these statements do not make sense.

I used to think of Autonomy as a search vendor but with your acquisitions of ZANTAZ and Interwoven you seem to be moving away from pure Search. How should I think about your core business?

To understand Autonomy you have to think back to the late 1960s when computers first started to be used for business. Back then computers were far too basic to understand the rich forms of information that human beings deal in. So the solution was to take human friendly information and distil it into a much simpler form - the database. So if a person lives at "3 Acacia Avenue" that information is entered into the particular row and column of the database that is for the first line of the address. The information is structured so that its position tells the computer what it is for. And the computer could then identify, for example, that column 3 was the amount of inventory in the warehouse, and when that number went close to zero it could automatically issue a purchase order to replenish the warehouse. We had automated a business process and replaced a human being who would have had to do it manually.

Computers are now a lot more sophisticated and they have caught up with the way human friendly information works, to the point where they can now understand meaning. So the computer can watch emails being sent within a bank and identify those that mean there is a compliance problem. Or listen to calls in the contact centre and identify a number of calls all about the same product issue, perhaps an exploding battery. The important thing to realise is that it is the processing of information that yields the value. Simply finding it for a human being to do the work is of limited value. So while it happens that if you can enable a computer to understand meaning you can make the world's best search, that is not the ultimate goal for Autonomy. The company's entire strategy is about accelerating the adoption of technology to automate the processing of human friendly information. Whether in acquiring a brand that is prevalent in the regulatory technology space and then injecting Meaning Based Computing technology into it, or in Web Content Management (WCM) and empowering computers to understand customer interactions. The fundamental idea is the same: computers should map to our human world and solve our problems, rather than the other way around.

Autonomy appears to have a huge opportunity but how do you measure your addressable market and think about its size?

Behind every piece of enterprise software currently in existence there is a database. It might be software to administer a hospital or a CRM system all have a database at the heart and then a tailored interface to make different operations possible depending on the use case. Analysts estimate that approximately 20% of enterprise information is of the structured type that is held in a database. The remaining 80% is in human friendly forms such as email, telephone conversations and video, which cannot easily be put into a database. IDOL is the equivalent of the database for unstructured information. It sits behind every enterprise application and allows human friendly information to be processed.

That database market is effectively based on a royalty model, with an effective royalty rate of between 8-10% resulting in annual turnover of approximately $20bn in 2008. Even if Autonomy's end market were only the size of the structured market it is still a massive opportunity that we have only begun to tap. In reality, there is much more unstructured information in the world than structured, by some estimates four times as much, and Gartner's head of worldwide research, Peter Sondergaard recently estimated that this area will one day generate revenues of multiple trillions of dollars.

I was interested to see the recent announcement about Autonomy's new "Collection to the Cloud" solution. Where does this sit within the three phases of development in the regulatory market that you've discussed previously?

This is part of Autonomy's legal hold solution set, and it fits squarely into the second phase of development we describe below, where existing rules are reinterpreted in the light of the recent turmoil, and the updating of the legal system in light of technology changes. The technology uses IDOL's conceptual capabilities to identify and collect material pieces of electronic information from laptops, desktops and all the different types of enterprise information repository out there. The information that's collected is then stored in Autonomy's cloud-based archive in our secure data centres. It dramatically reduces the cost and manual effort involved in the preservation and collection exercise that a large corporation might have to go through as part of legal proceedings. One of the big advantages of this system is that it also enables organisations to comply with data privacy laws by ensuring that personal data that's not relevant to the legal case is not collected. We felt this was the right time to launch it because we've recently seen increasing interest from procurement in solutions that will address that second phase of development in the market.

How positive was the initial response from customers to the recently launched range of Meaning Based Marketing solutions?

These solutions are aimed at large online retailers to help them optimise their websites. The technology allows them to increase conversion rates by segmenting and targeting customers more accurately, through an understanding of the meaning behind customer interactions and online catalogues. We had not anticipated a significant response given that in the current environment of weak IT spending, for projects without a regulatory driver the available budget is quite restricted. In contrast to our expectations we actually signed deals with major blue-chip retailers for modest systems on the order of $50-100K. This is a very encouraging sign because it shows that even with commercial IT budgets stretched so thinly, there is still significant demand for the new technology. We expect that in the event of an economic upturn, these customers will return to extend systems and deploy on a much larger scale, which would generate deals of ten times the magnitude. The reason for the strong showing at this stage is probably the fact that even a small system immediately yields a quantifiable return on investment. The systems that are now live and running the new technology have typically experienced an increase of 5-10% sales.

Can you comment on the new SPE product, its core differentiation, and how it has been received in the market?

Autonomy is applying the same advanced technologies it has used to revolutionize the unstructured market to transform the $20bn annual database market, by turning legacy RDBMSs into next-generation probabilistic inference engines that can understand "shades of gray." Corporations are dependent on RDBMS technology to power literally every type of business software application. While databases perform operations, the software doesn't understand the meaning of the data itself and is limited to viewing the world as black or white. With the use of advanced probabilistic methodologies, Autonomy delivers new levels of database processing intelligence, enabling the software to suggest results, even when an exact match doesn't exist in the database. This technology represents a radical shift in the intelligence businesses can gain from information, by delivering the ability to understand the "meaning" in the data in the billions of corporate databases in use today.

A self-learning solution, IDOL SPE can automatically make connections in the data that would otherwise need to be interpreted by human beings. The solution analyzes interactions, usage, and multiple datasets to spot patterns in structured data and make non-obvious predictions. Unlike business intelligence solutions that require building and maintaining complex cubes that operate outside the database, IDOL SPE brings probabilistic inference into existing databases.

For instance, an airline can leverage IDOL SPE to improve the online customer experience and increase sales. A customer searching for a flight from New York to San Francisco at a date and time when all flights are sold out will automatically be offered an alternative airport such as Oakland or San Jose, rather than returning "no results" to the customer's search. IDOL SPE infers the result from patterns in the data and its usage without the need for scripts or geographic information.

Analyst feedback has been extremely positive:

We expect this exciting new market opportunity to be a slow burn revenue stream as market understanding catches up to what is now possible, but a fundamental one all the same, that should one day be as significant as our core IDOL business.

Who are Autonomy's main competitors now?

To answer this question it is helpful to think more about the market. At the low-end of the market you have keyword search engines which are off-the-shelf and can be obtained for free in many cases. Autonomy does not operate in that market at all. Then there is a mid-market, where you find department level search, with products that can connect to perhaps 10 of the repositories that are typically found within the enterprise, but there is no security or ability to deliver legally warranted results. Again this is a market in which Autonomy does not play. Having said that, growth in that market does feed our market as people often deploy a simple keyword search engine and then realise that it is not adequate.

Autonomy operates in the high-end market, where you need a technology that is proven to connect to all of the 400 different types of information sources in the enterprise and not lose any results, offer mapped security so that people cannot find documents they are not supposed to see, and offer all of the 500 functions that people might need on a single platform. In that market there really is no competition for Autonomy other than perhaps getting a warehouse full of human beings to try and do the same thing. Often you will see marketing hype around other products but it is very easy to dismiss these with what is called a Proof of Concept (POC) where you run the two solutions in parallel.

Autonomy also now has around 70 standardised customers for whom any further purchases relating to unstructured information are explicitly not competitive. This is a situation similar to perhaps ten years ago when the CIO would realise that nine departments had bought Oracle and the tenth had bought SAP. They very quickly realised they would have to standardise on one or the other. Autonomy's standardisation customers have reached the same decision about unstructured information and concluded that any purchases related to unstructured information across the entire business should be Autonomy.

Can you tell me more about Autonomy's management team?

Autonomy has an incredibly experienced and long-serving management team. Dr Michael Lynch, OBE, Chief Executive Officer (CEO) founded the company in 1996. Sushovan Hussain, Board Director and Chief Financial Officer joined Autonomy in 2001. Dr. Peter Menell, Chief Technology Officer, joined Autonomy in 1999 Andrew M. Kanter, Chief Operating Officer, joined Autonomy in 2000. Autonomy also has a duplicate management team in the US, led by Stouffer Egan, Chief Executive Officer, Autonomy Inc. Nicole Eagan, Chief Marketing Officer, and Mike Sullivan, CEO, Autonomy ZANTAZ.

What has been the impact of downsizing the Interwoven business on its DSOs?

They have moved in-line with the core Autonomy business. There are three reasons for this:

1. new revenues are now on Autonomy contract terms not IWOV legacy terms - importantly any large, joint customers would also have been moved onto Autonomy Master Framework Agreements which are multi-year and consequently tend to have longer terms;
2. Lower services contribution means effective DSO increases;
3. Difference in accounting principles between US GAAP and IFRS means that structural DSO is higher, as no netting can be applied to trade receivables and deferred revenues associated with renewed maintenance.

Why should DSOs be any different on services revenues as compared to licences?

This depends on the contract type. For Autonomy services contribute less than 5% of revenues and will always be part of a purchase order with software and maintenance. Hence payment terms are the same. You have to remember services are often tied to the completion of services contract whereas licence is delivered on day one.

However, IWOV pre acquisition had stand-alone professional services engagements where it might use third party resources and therefore set credit terms of circa 30 days.

What types of revenue streams do you have?

We sell standard software and 95% of Group revenues are product related:

1. Upfront Licence revenues (normally perpetual and recognised on delivery in accordance with IFRS which in terms of revenue recognition is consistent with US GAAP SOP 97-2. SOP 97-2 contains more detailed specific guidance on software revenue recognition which fits within the framework of IFRS)
2. Maintenance Revenues (usually derived as 15% of "1" above, recognised rateably and renewed annually)
3. Hosted/SaaS Revenues (pay as you go or subscription and recognised rateably)
4. OEM revenues (split as development and ongoing)
5. Other (primarily Professional Services but can include training and appliances)

Why do you not have a vertical business focus?

The IDOL technology is fundamental and horizontal - whether you are selling to Barclays or Del Monte it is the same product. It has a myriad of use cases, many of which are at the cutting edge of technology use within the Enterprise today. There are 3 reasons why it makes sense not to have a vertical focus: 1. we do not need to commit additional R&D (or services resource) to build out different use cases of IDOL. This keeps the model focused and ensures that the financial model is geared; 2. our OEM partners provide the vertical experience and targeted use of the IDOL platform; and 3. our business is so diverse that having vertical overlays in sales would be impossible to manage (selling to Government agencies one day and MTV the next) so we leverage the vertical expertise of our 400 go to market partners instead.

Please clarify your organic growth calculation?

Below is the calculation for 12 months through 31st December 2009.

There is clearly an exciting opportunity in regulatory but how long will this last?

We believe that there will be 3 phases of demand in the regulatory space. The first phase of regulatory deals (basic e-discovery and archiving solutions) is set to continue for some time, but the more interesting change is the acceleration we're beginning to see in phase two. This second phase is a new sudden focus on legal hold, early case assessment and advanced e-discovery, due to the regulators now making it clear that the traditional, manual, trust based models for legal hold are no longer acceptable. This affects every business across every sector and the total opportunity is expected to be larger than phase 1, albeit over a longer time period. We recently won a multi-million dollar contract from a major Wall Street institution to do this. Finally the third phase which is coming, in our view, will be focused on the kinds of things that ,for example the UK FSA and US President Barack Obama has been speaking about recently, such as information governance and supervisory needs. This will be a multi-year investment cycle that could be larger than the previous two phases.

Your business seems dependent on regulatory demand and as such is seen as a defensive stock. What will happen when this spending priority diminishes?

Although we have benefited from the increased regulation in the market this is only one of our pillars of growth. Our business model and growth opportunity is based on 3 very different and possibly out of phase cycles - the need to protect, power and promote the Enterprise. The protect cycle is multi-year and we believe splits into 3 phases [see previous question for definitions]. We are mid way through the first phase, and phase 3 we believe will be at least 7 years in length. Promote is focused around our Meaning Based Marketing product set and this will benefit from the increase in corporate spend as organisations need to leverage their on-line business model. Power is what has traditionally been our core business and is the critical infrastructure that enables any type of organisation to manage and process all of its data independently of where it is stored or created (structured, unstructured, and semi-structured).

At the second quarter results presentation you started to comment on 2nd and 3rd phase regulatory deals becoming visible. Has this trend continued and what should we expect?

Yes it has. Phase 2 has occurred sooner than we would have anticipated. Our first deal was in Q2 and this has continued into 2H of 2009.

Are we likely to see any more of the so-called "mega deals" in this area?

Yes we are currently tracking a number of large deals in this area. But it is now clear that the market for these solutions is coming in three distinct phases. The first phase evident over the last year or so and continuing strongly is driven by basics as people realize they aren't in compliance, such as archiving and standard eDiscovery needs.

Am I correct in assuming movements in either short or long term provisions are not included in the working capital movements as the provision is taken in the P&L and the cash movement in investing activities?

Correct. These all relate to restructuring provisions that were acquired upon acquisition and relate to events that took place prior to deal announcement. These are non-operating and do not relate to decisions taken by Autonomy.

Why doesn't Autonomy follow the standard industry practice of writing down acquired deferred revenues? Especially, in the case of Zantaz, where these were set to decline as credits were used up.

IFRS does not require write-down of deferred revenues only US GAAP does and as you know this distorts the organic growth rates for 2 years so is less than ideal.

Under IFRS deferred revenues are recognised based on their fair value at the time of acquisition. We are required to account under IFRS and hence record the values accordingly.

Can you comment in more detail on your revenue recognition policy?

At the outset it is important to clarify that Autonomy uses IFRS as the formal recognition policy used to prepare its accounts. However, being a pure software business there are certain areas where IFRS does not have specific guidelines to cater for all situations but merely provides guiding principles. Thus for revenue recognition we voluntarily adhere to the principles set out in US GAAP SOP 97-2, which is far more detailed, prescriptive and conservative than IFRS, and that has made it the "gold standard" for other software companies like Oracle and Microsoft. In fact the guidelines to SOP 97-2 and SAB101 run to over 100 pages as they have evolved over the years in tandem with the much broader US software market requirements. IFRS on the other hand is far less detailed and less conservative when it comes to software companies. In short, we use SOP 97-2 because it is designed specifically for software revenue recognition and hence tends to be more relevant, but if IFRS for any revenue recognition point became more relevant or restrictive than SOP 97-2, we would use IFRS for revenue recognition since that is the formal policy.

Below are the basic requirements under SOP 97-2:

Software revenue should be recognized if the following criteria are met:

Autonomy has been very successful at maintaining consistently high operating leverage. Please can you help me understand the company's margin profile and what I should expect going forward?

Autonomy is one of the very rare examples of a pure software model. Many software companies are actually "pseudo-services" outfits that do an awful lot of customisation work on the product for every single implementation, while Autonomy ships a standard product that requires very little tailoring, with the necessary implementation work carried out by approved partners such as IBM Global Services, Accenture and others. This means that after the cost base has been covered, for every extra dollar of revenue that comes in, you simply take off nine cents to get to the gross margin, and then a further ten cents which is paid in commissions to our partner managers. That leaves 81 cents which fall straight through to the bottom line. We see no reason that operating margins of 45-50% should not be consistently achievable in the future.

Historically, companies that have achieved these levels often diversify into hardware or other areas, which has a big impact on operating margin. For Autonomy, it's hard to see what other areas could be valuable enough to warrant such a large departure from the pure software model.

In 2007 there was an increase of approximately $8.5m y/y in trade receivables (from $25.8m to $34.3m) that are over due. What proportion of this was related to the existing business, and what proportion was related to acquisitions? So while the number is constant as a percentage of sales it could mask deterioration in "revenue quality" at the core business, where DSOs appear higher. Is this a fair assessment?

No not at all. The age profile actually improved with 90+ day overdue receivables falling from 32% to 19% of total. Similarly, although the 0-90 day overdue increased the average actually age fell.

The organic growth in the business was 18% (from $251m to $295m) and thus the theoretical growth in TR would have been to $30.5m. The $4m above this is easily explained due to the acquisition of Zantaz which added c.$25m in TR some of which could have aged slightly due to the focus in shifting the business model to pay-as-you go.

I noticed that Autonomy's provision for doubtful accounts doubled from '07 to '08, and the charge for bad debts increased from -$0.2m to $6.8m over the same period. Please can you explain as I thought this should be fairly steady? Also, given that Interwoven had a lower provision as a percentage of amounts due, then should we expect the same metric for the group to go down year-on-year in '09 as we get the contribution / mix from Interwoven? Finally, if Autonomy's actual charge off of bad debts is so low compared to the provision then is it likely that the accounting for provisions will be looked at as being too conservative and not an adequate representation of risk?

Autonomy's revenues grew by 47% from '07 to '08, and as we got into the downturn people did slow up their payments a little bit, so the aging crept up which also had a small effect - doubling is about what you would expect. As for the charge, because the provision is automatic and conservative there have been occasions when someone finally does pay up where a negative charge occurs, the write off is a combination of positive and negative. The charge in reality on average runs at around 1%. Given our deal size it is not unexpected to see some variation here.

Post the Interwoven acquisition, Autonomy group policies are applied across the group (whilst this is not surprising it's also practical as IWOV is fully integrated and does not have its own finance department). Thus we take the 'IWOV' aged debtors book and apply our more conservative formula to the same IWOV data, which gives a similar provision as found elsewhere in Autonomy...ie the underlying debtor structure is similar, but our provision applied to their numbers is more conservative. So whilst provisions have gone up as a proportion of the combined revenues, the actual write-off is expected to increase incrementally in line with the growth of the business.

The quantitative evaluation of the risk of doubtful accounts is automatic, and it errs on the conservative side because that's generally mandated by IFRS. While we take your point that the policy is conservative, it's always better to be prudent. The company prefers to be conservative in these things, for example, adopting a much more conservative revenue recognition policy than found in UK software companies.

Can you help me reconcile the movements in trade receivables and cash between Interwoven Q408, IWOV deal closure and your Q109?

Here are some guidelines to help make this reconciliation. IWOV debtors in the closing balance sheet (4Q08) was $42m. On the 1Q09 conference call it was stated that these were approximately$36m when the combined books were closed for 1Q09. The confusion probably arises because published IWOV "other assets" at deal closure was only $27m. In understanding these movements there are a number of effects that need to be taken into account:

1. As reported for 4Q08, IWOV "Other Assets" (Account receivables, fixed assets, Deferred Tax Assets and Other assets) totalled $62.6m.
2. The bulk of the assets relate to: Accounts Receivables $42m and Fixed Assets $16m (4Q08). DTA and other each being c.$3m each.
3. The balance sheet at closure has IWOV "Other Assets" $26.9m. Although not given fixed assets fell significantly to c.$4m (reconcile with CapEx in 1Q09 cash flow). This would imply trade receivables have fallen to c.$21-22m.
4. By subtracting the IWOV trade receivables at closure from 4Q you get a c.$20m collection that occurred in the run up to the acquisition. Yet IWOV cash is flat ($185m at Q4 and $184m at deal closure).
5. The reason why IWOV cash is not higher at the time of the acquisition is that Bankers fees and self help "restructuring costs" totalling c.$15-20m need to be adjusted for.
6. The final movement in IWOV receivables from deal closure (at c.$21-22m) to quarter end of $35-36m represents an increase of $14-15m. This higher than the stub revenue approximation in Q1 of $8m for the following reasons: 1. PO's that have been received and invoiced but where revenue recognition requirements could not yet be met (normally the case just post an acquisition); 2. the maintenance element of recognised licences; 3. any maintenance renewals that occurred during the period.

Can you explain why Other Receivables has increased?

Other Receivables will tend to grow in-line with revenues and for Autonomy represents 5-6% of annual revenues. This includes prepaid insurances (such as Director & Officers - based on market capitalisation of the business – 1 yr paid up front, D&O tail from acquisitions – 3 years paid upfront), employee insurance, prepaid sales taxes, prepaid office rents etc. all of which have increased with the size of the business. It also includes a small amount of accrued income which is less than 5% of sales in any quarter, and includes hosted revenues from pay as you go billed a day after the quarter end.

Can you explain why Other Payables has increased?

Other Payables includes accrued bonuses, expenses, income tax payables, spend with external suppliers, which have not yet been invoiced etc. There are two factors that have meant that this has increased from 10% to a normalised rate of 15% of costs: the first is due to the Interwoven acquisition where payables as a percentage of costs were higher than for Autonomy; the second is that in Q3 expenditure on IDOL SPE was made in the second half of the quarter, but is invoiced and paid in Q4.

I saw an article which said that Autonomy only converts 75% of its profits to cash. This seems surprising from what I know of the company. Can you explain?

The article contained a number of factual inaccuracies, which are listed here.

The article said: "The company announced a cash conversion rate - calculated as Q3 operating cash flow over Q3 earnings - of 131pc, which it duly trumpeted throughout its presentation.What Autonomy failed to flag was that it had not yet paid for $44m (£27m) of products and services, which were accounted for as a cost in the quarter. This obviously flatters the cash conversion rate as the cost is counted against earnings while cash flow remains unaffected."

This statement is inaccurate. If the effect of the extra creditors above the normal level was removed, the cash conversion was still over 100% in the quarter (using the standard method of current quarter CFFO on current quarter EBITDA).

As specified on the Q3'09 earnings call, there was a one-time investment in a fundamental new product that led to a negative impact on operating cash flow in the quarter and a corresponding positive impact on cash flow from operations. Adjusting for this effect the underlying cash conversion was approximately 105% (using the standard method of current quarter CFFO on current quarter EBITDA).

The article also said: "When asked about this after the presentation, chief executive Mike Lynch conceded that it would indeed be reversed in the following quarter."

This is inaccurate. It was stated by management on the Q3'09 earnings call that the cash flow number for Q4 will be lower due to the fact that it is Q3 business that generates the Q4 cash flow, which is numerically lower than the Q4 generated EBITDA. This is a normal seasonal effect for the business, and is not due to the reason mentioned in the article. The last twelve month cash conversion is still expected to be good.

The article went on to say: "If that one-off boost to cash flow had not occurred, Autonomy's cash conversion rate would be closer to 73pc."

No, it is approximately 105% adjusting for the investment in IDOL SPE (using the standard method of current quarter CFFO on current quarter EBITDA).

To calculate the underlying cash conversion it is not appropriate to remove only the $44m payable move - as was done by the author. There was a c.$25m impact due to the investment, but there was also a corresponding negative impact on the profitability of the business. It is not appropriate to adjust only one side of the calculation.

The article then commented: "That is roughly in line with the company's cash conversion for the past couple of years and should ring alarm bells for investors."

No, this statement is not correct. In fact, the trailing 12 months cash conversion is approximately 90% (using the standard method without any lag), which is as expected for a business that has shown the growth rates Autonomy has done.

Finally, the article states: "If a company is consistently not converting a quarter of its profits into cash, those profits arguably cannot be described as profits."

Autonomy is converting almost all of its profits into cash. There is a lag of one quarter due to the growth in the business. This would only unwind if Autonomy was to stop growing.

The above statement suffers from two fatal flaws:

1. The linking of cash conversion percentage to the amount of profits. If indeed a company's cash conversion number is 75% but it is growing strongly, it will still have cash conversion of 75% but can be converting 100% its profits into cash. The business signed in the previous quarter gives rise to the cash flow in this quarter. That number is divided by the EBITDA in this quarter. If the company is growing, the number is less than 100% as this quarter is bigger than last, but the company may still be converting all of its profits into cash.
2. It is simple to show Autonomy is converting its profits into cash as its trailing twelve months cash conversion (unlagged) is running at approximately 90% not the 75% stated.

Can you confirm whether the difference in acquisition costs for Interwoven reported between the second and third quarters relates to restructuring costs?

We have seen some analysis that attempts to reconcile the audited statements made at the time of acquisition, with the note to the accounts in 2Q09, and finally with the cash flow costs as per 3Q09. This analysis concludes that the difference of $10m relates to post-acquisition restructuring costs that have been taken through the cashflow. This is not the case (see below). Additionally, the analysis contains some fundamental errors, including the incorrect assumption that fees paid by Interwoven were post-acquisition, when they were in fact paid pre-acquisition. The full reconciliation is shown below:

The transaction value as per 22nd January was an initial estimate based on the share price paid and the estimate of purchasing in the money stock options at Interwoven on this date. It did not include any additional vesting of options in the interim or any associated legal and bankers' fees.

Delta A: The $29m of additional costs relates to $16m of additional cash paid for the equity of Interwoven giving the total consideration paid of $791m (as disclosed in Note 3 of 2Q09 release and shown above) and bankers costs and legal fees for Autonomy for $13m. At this stage the costs relating to fees were still estimates as any currency moves between deal announcement and final payment were still to be adjusted for. It is also worth noting that fees were paid partially in Q2 and finally in Q3 (as per cash flow statements and is a result of when advisors invoiced Autonomy).

Delta B: The difference between 2Q09 Note to accounts and the 9 month cash flow relates largely to timing differences and falls into 3 areas:

1. $6.5m Interwoven self help provisions: at the time of the acquisition it was stated that Interwoven had previously undertaken a restructuring plan. This included a RIF and some office closures. $6.5m of provisions were made by Interwoven and recorded in their P&L prior to acquisition. Autonomy assumed those provisions, resulting from decisions made by Interwoven management prior to the acquisition, and the cash costs of settling those provisions goes through cash flow from investment activities as per IFRS. These could continue to unwind in future periods;
2. $2m FX move: Costs are recorded at the time of the acquisition (17th March) but were not paid until 2Q and 3Q. The £:$ exchange rate moved from c.$1.4 at time of deal closure to $1.6-1.7 in Q2 and Q3. This represented a 14-18% depreciation of sterling and increased the effective costs of the transaction.
3. $1m Additional Costs: these related to additional legal fees.

I came across some research which purports that IWOV bankers' fees could not have occurred pre-acquisition (because cash balance at 17th March should have been higher than $184m) and that the acquired deferred revenue balance should have been $73m not the $60m stated on the Q1'09 call. Is that right?

Unfortunately the analysis you mention uses a number of assumptions that are non-evidence based. Trying to extrapolate from the Q4'08 cost structure to give 1st January and 16th March cash flow impacts at the time of the acquisition is impossible without knowing how the business was being managed. Also, we would draw your attention to the analysis further down this page relating to why $184m of acquired cash is consistent with the cash flows that occurred within IWOV. The inaccuracies in the analysis we have seen are as follows:

A. Gross Margin of 60%: IWOV gross margin was 75% in Q4'08. Assuming professional services are a fixed cost and that the bulk of revenues recognised related to support and maintenance and services then the gross profit in this period was $20-22m;

B. Opex assumptions: IWOV opex in Q4'08 was c.$36m (ex SBC). This would have included discretionary bonuses, discretionary marketing spend, and commissions all of which would have been on a much lower run-rate in Q4'08 (seasonality, and cost containment ahead of acquisition). There was also the self help restructuring that needs to be adjusted for. Opex for the period was thus in the low $20m's not the $32m alluded to.

C. EBITDA: making these adjustments IWOV EBITDA in the period was between $2-3m.

D. Cash Flow movements: we have explained that there was a positive c.$20m receivable move and also a c.$2m payable move due to better cash management ahead of the acquisition. Deferred release in the period is likely to have been $12m. This means that the business would have actually been quite cash positive in the period and certainly in excess of $12m.

E. Bankers fees: there were cash costs in the quarter as mentioned in response to an earlier question relating to c.$15m. Netting all these effects leaves the cash balance broadly unchanged from Q4'08 to time of acquisition.

F. Receivables move and impact on deferred revenues: reconciliation of receivables (as per "Can you help me reconcile the movements in trade receivables and cash between Interwoven Q4'08, IWOV deal closure and your Q1'09"?) shows that there was a delta of c.$5m not the $12m that has been suggested, and this is what has been incorrectly analysed.

What is normal seasonality in revenues for the business?

Excluding the overlying growth within the business seasonality would be expected as follows:

Q1 = Q2

Q3 < Q1 or Q2

Q4 > Q1

Please can you explain your seasonality in operating margins

Because our cost based is largely fixed (S&M is the only element not fixed) the operating margin seasonality will tend to follow that of our revenue stream with a weaker margin in Q3 and the strongest in Q4. S&M costs will tend to follow revenues because a partner manager margin of approximately 10% is paid for every additional dollar of revenues above the S&M cost base. This can be seen below:

What is the normal seasonality associated (balance sheet) with deferred revenues during the year?

Our seasonality follows the norm in the software industry. Deferred revenues will tend to be strongest in Q1 and then released during the year. Q2 will be seasonally weaker with Q3 flat to slightly down. Q4 will normally show a slight pick up in deferred more in-line with Microsoft and Oracle.

As a point of comparison the following table shows this for other enterprise software companies: