Shares for the future: lessons from a famous tech company

Although it is unusual for me to write a book review, bear with me - the Decision Engine table follows.

There are so many reasons to read The Curious Case of Mike Lynch: The Improbable Life of a Tech Billionaire it is difficult to know where to start.

• Invest with ii: Open a Stocks & Shares ISA | ISA Investment Ideas | Transfer a Stocks & Shares ISA

Mike Lynch was the co-founder and chief executive of Autonomy, one of Britain’s most famous technology companies and one of its most infamous. It is a rags to riches tale, a David against Goliath story, and a corporate and legal thriller with a bizarre twist at the end. If, like me and Mike Lynch, you came of age in the 1980s, grew up with Tomorrow’s World and the Sinclair Spectrum, and live in Cambridge, it is a trip down memory lane, and many of the locations are just around the corner.

Mostly, though, I got into this book by Katie Prescott, The Times’ technology business editor, for the accounting, particularly the accounting of software companies. This is a complex subject and there may be lessons from history now Autonomy’s accounting has been scrutinised by various courts.

Autonomy was born in 1996. Among the early red flags was a tank of cannibalistic piranhas at the centre of the company’s logo in reception. The company had grown out of Lynch’s first venture, Cambridge Neurodynamics.

Cambridge Neurodynamics employees would assure suppliers their questions about payment would be dealt with by the company’s unresponsive finance director Frank Bridges. “His elusiveness wasn’t that surprising. Lynch never paid anyone until they threatened to close his account, which was a way of maintaining cash flow. Also, Mr Bridges did not exist.”

There was also the suspicion that Cambridge Neurodynamics had stolen source code, or at least ideas that may have ended up in its Dynamic Reasoning Engine. The DRE was the technology Cambridge Neurodynamics spun out as Autonomy. It was a search engine that could index data, categorise information, and link to it.

Mike Lynch comes across as a man for whom the ends justified the means. He bullied staff and made claims only loosely related to the truth to customers.

The proximate reason for the company’s undoing though, was the financial shenanigans required to meet quarterly estimates, revenue and profit forecasts made by analysts for investors. When companies fail to meet estimates, their share prices fall. Autonomy went to extraordinary lengths to make sure that did not happen.

One theory is that Autonomy needed to keep its share price high so it could buy other companies with shares, which it would upgrade with its technology. But it was difficult to know whether the company was growing because of its superior software, or it was just buying customers. “One thing was certain – each new company in its stable made Autonomy’s accounts that much harder to decipher.”

• Ian Cowie: are you braced for AI apocalypse or buying into AI optimism?
• ii Tech Focus: Broadcom, Palantir, Oracle
• The Analyst: AI stocks, trusts and funds still worth buying

One tactic the company used was to buy hardware, which it would sell on to customers at a loss. The sales, though unprofitable, helped Autonomy meet revenue targets. Autonomy did everything it could to hide the source of the revenue.

It also accounted for hardware sales in a “fishy” way, as marketing costs. This meant that the cost of the hardware did not come out of gross profit, which would normally include the cost of buying the hardware, but operating profit. Gross profit margins, an important metric for analysts, were inflated.

Hardware sales peaked in Autonomy’s third-quarter results in 2009 when revenue was elevated 20%. The company’s auditor was uncomfortable, but after handwringing and a few glasses of wine, found an obfuscating form of words. Much later, Deloitte was fined for its audits. Its Cambridge office had subjugated itself to its biggest customer, the regulator deemed. The board’s Audit Committee asked questions but also caved.

You may wonder how Autonomy handled ballooning sales and marketing costs, which were depressing operating profit. Lynch said they were associated with a “whizzy” new product, “The Structured Probability Engine”. The product was so novel, the company said, it might be some time before it generated a return.

This was fiction. Autonomy had developed a product, or feature, with this name, but a tiny fraction of the increase in costs had been spent on marketing.

Mike Lynch, former chief executive officer of Autonomy, pictured in 2019 in London. Photo: Dan Kitwood/Getty Images.

Analysts also questioned cash flow, which was lagging profit. Software companies sell licences up front for a fee along with maintenance and support contracts, typically charged a year in advance. The revenue from these contracts is recognised over the course of the year, by which time, the company should already have been paid. Cash conversion, the ratio of cash flow to profit, should be high.

At Autonomy, cash conversion was deteriorating because the company was tying up cash in hardware. It was also recognising revenue and booking profit for deals that were never consummated.

Autonomy pioneered this technique as it negotiated a deal with The Vatican, but it used it “again and again”. If it looked like a contract would not be signed before the end of the quarter, Autonomy would sell the software to a reseller to sell on to the end customer later. Contractually, the resellers took on the risk that the sale might not go through, so it looked legitimate. But there was a tacit agreement that Autonomy would recompense the reseller if a deal failed. For example, Autonomy would overpay for products or services from the reseller.

The tactic was unsustainable and this time Autonomy’s chief financial officer Sushovan Hussain, who went on to be the only Autonomy office holder or staffer to be convicted of fraud in a criminal court, was feeling the heat.

• The global funds abandoning the Magnificent Seven
• Sign up to our free newsletter for investment ideas, latest news and award-winning analysis
• 10 shares to give you a £10,000 annual income in 2026

In December 2010, he sent Lynch an email saying: “Really don’t know what to do Mike. As I guessed revenue fell away completely...There are swathes of reps with nothing to do maybe chase imaginary deals. So radical action is required, really radical, we can’t wait any more. Everywhere I look...it’s bad.”

I have focused on these accounting shenanigans because they are instructive for investors. At the end of civil proceedings in December 2022, a UK judge found that between 2009 and 2011 Autonomy had overstated sales by half a billion dollars.

There is much more to the book though. Mike Lynch’s story has a famously curious ending, he died in 2024, months after his improbable acquittal in a US criminal trial. His superyacht capsized one night in a freak storm off the Sicilian coast. This was just days after his single co-defendant Steve Chamberlain, a former vice-president of finance at Autonomy, was seriously injured in a road accident. Chamberlain had also been acquitted and died in hospital a day after Lynch died.

In addition to this twist in the tale, the case of Mike Lynch would not have been as curious had he been a one-dimensional cartoon villain, like the Bond villains the Autonomy meeting rooms were named after. But there were two sides to Lynch. He also comes across as a family man. He had a lot of time for children, and he was loyal to some friends in difficult times for them. He was a man who could charm as well as bully, which may have been what won over the US jury.

These stories, Hewlett Packard’s acquisition of Autonomy in 2011, its buyer’s regret, the civil and criminal proceedings that followed, and Lynch’s ventures after Autonomy, such as Darktrace - the AI cybersecurity platform that briefly lit up the stock exchange earlier this decade, are also well covered. 

One thing I felt was less well resolved, was the question of whether Autonomy’s software was any good. The book casts shade on it. A BBC Radio 4 documentary in 2012 could find no customers prepared to show how it benefited them. Although some said they had been happy with the software, they no longer used it.

Autonomy’s software may have been ahead of its time. It used Bayesian estimation, also at the heart of modern AI software, and its ability to analyse information “in human form” was “inspiring when it worked”. Mike Lynch is quoted (in 2017) as saying: “I’d like to think that when people look back, one of the things I did was produce one of the world’s first true AI companies.”

Whether he did, and whether Autonomy was the company he was referring to, is another curiosity.

30 Shares for the future

Here is the ranked list of Decision Engine shares. I review the scores at least once a year, soon after each company has published its annual report. The price scores are calculated using the share price prior to publication.

Generally, I consider shares that score more than 5 out of 10 to be worthy of long-term investment in sizes determined by the ideal holding size (ihs%).

Click on a share's score to see a breakdown (scores may have changed due to movements in share price). Key: qual is the share’s score out of 9 for the three quality factors (capabilities, risks, and strategy), price is the price score from -3 to +1, and ih% is the suggested ideal holding size as a percentage of the total value of a diversified portfolio.

Richard Beddard is a freelance contributor and not a direct employee of interactive investor.  

Richard owns many shares in the Decision Engine. He weights his portfolio so it owns bigger holdings in the higher-scoring shares.

For more on the Decision Engine and Share Sleuth, please see Richard’s explainer.

Contact Richard Beddard by email: richard@beddard.net or on Twitter: @RichardBeddard