Mike Lynch has steered Autonomy from tiny start-up to UK software giant. But what gives his company an edge? Technology that makes sense of 'millions of hours of human conversation'... and he thanks an 18th-century vicar for the idea
More than a decade after its FTSE-100 debut, Autonomy still divides opinion. Last October, at the tail end of a tough year in which the software firm's revenues grew by 18 per cent, Deutsche Bank analyst Marc Geall delivered an astonishing attack on the company's strategy. Geall told investors it was time for Autonomy to change its business model, derided the company's sales team as "hunters not farmers" and questioned the firm's structure as being "more representative of a start-up than a major global player".
It might seem churlish to suggest the City doesn't understand one of Britain's most prized technology assets. Geall is, after all, a former Autonomy employee. But if anything Mike Lynch, who co-founded the company in 1996, seems a little bemused by Geall's words, especially the suggestion that Autonomy should be anything other than entrepreneurial. "Should the organisation be run like one of the boring large software companies?" he asks. "We had to sit down and think about that. It took us about five minutes. Of course we should be entrepreneurial. Technology moves at an incredible pace. You have to be entrepreneurial if you want to adapt."
Autonomy is, of course, no longer a start-up. Remaining entrepreneurial with a staff of 1,800 and projected revenues of a billion dollars is a difficult trick to manage, especially with the City to please. Internally it's also a struggle, says Lynch, because employees change their behaviour as companies become bigger. "People get very good at playing the politics. And that's normally about not sticking your neck out," says Lynch. "When something happens, they want to be number three or four. Not number one."
Sticking your neck out could be Autonomy's motto. Could be... Lynch employs a multitude of other candidates. "Always take a gun to a knife fight" seems popular. It means: exploit your competitive advantage, ruthlessly and absolutely. Autonomy's advantage is also Lynch's lifetime work: a piece of software that allows computers to listen to phone calls or read emails and understand what's happening. "Our technology allows computers to make sense of human conversations. That's the unfair advantage that allowed two slightly nerdy people from Cambridge to create a FTSE-100 company."
Lynch likes to encourage risk-taking. He says it's one part of running a true meritocracy in which valuable ideas are always rewarded, whoever creates them. It's easy to criticise the person with a new idea, says Lynch. It's the English way. "What's much harder is to say, 'there's a seed of an idea there'." Meetings at Autonomy are split into two sessions, positive and negative. "We have meetings where you are only allowed to say why this is a good idea," Lynch explains. "You can then have a separate meeting about all the potential problems. But you've got to allow space for ideas to grow."
It might sound quirky, contrary even. But George O'Connor, director at investment bank Panmure Gordon, believes Autonomy's rise depends on a refusal to conform. "Autonomy is atypical of the software sector," says O'Connor. "It runs a lean ship", concentrating on sales and licensing rather than post-sales service to drive "very aggressive" margins. "It's easy for analysts to say they haven't grown the management layer and because of that they are vulnerable, but what are they vulnerable to?" The numbers, concludes O'Connor, reflect a "different operating model to [companies], such as IBM or Oracle. These are not apples and apples comparisons."
Lynch won't mind the comparison with industry greats. But he's keen to differentiate his management style, not necessarily from the competition, but from the ordinary. He ranks adaptability as highly as positivity, endeavouring to avoid what occupational psychologists call patterning—doing things the way they've always been done. "Many companies get this idea that this is how it is and this is how you do things. But in a fast-changing environment you can't afford to do that." Lynch calls it "the rule of properly". When people say, "we need to do this properly", he says, what they mean is, "they've no idea why they do it that way. You have to get back to first principles. Are those assumptions still valid?"
It's a scientific approach to business honed in the labs of Cambridge University, where Lynch first studied engineering. "I was always interested in the way things worked," he says. "When the electronic music revolution arrived I was desperate to be part of it." Like most students, Lynch couldn't afford a Fairlight synthesizer, even in the 1980s they stretched to £100,000. But he did have the aptitude to build one. He even managed to sell his designs to the trade, but it failed to make him rich. "People did it for love. But if they did software-writer credits for some of the hits of the late-Eighties I'd be on there."
One part geek, two parts entrepreneur, Lynch attacks life with the will of a man determined to be proved right. In Autonomy's early days, stateside meetings followed a pattern. Co-founders Lynch and Richard Gaunt would shake hands with a beaming American and then sit down for a patronising lecture about how different things were in the US. The implication was clear: what could a small start-up from Cambridge possibly know about the real world? In an attempt to counter the prejudice, Lynch started carrying a small card in his pocket, on which he'd printed the words: "It's different over here." Every time he heard that phrase, Lynch would flourish the card, like a bingo player.
These days, "they probably think we're a US company", he says. It's a backhanded compliment because although the firm now has a joint headquarters in San Francisco, it remains resolutely British in an industry dominated by US giants. How? It's an eccentric tale. Autonomy's "unfair advantage" is rooted partly in Lynch's synthesizer tinkering, which taught him the basics of signal processing, and partly in the fruitless attempt by the 18th-century vicar and statistician Thomas Bayes to prove the existence of God. While studying for his PhD, Lynch noticed that Bayes's probability equations could be used in a more contemporary context: helping computers make sense of information.
Computers have traditionally relied on databases to help them access and manage data. "A proportion of the world's information is organised into rows and columns, share prices in a database for example, which makes it easy for machines to interpret," explains Lynch. "But the majority of our data—emails, texts, or phone conversations—is unstructured and layered with nuanced meaning", making it difficult for machines to understand.
While most of the IT industry focused on databases, the value, thought Lynch, lay in interpreting that unstructured data. His first notable customer was the police force. In the early 1990s, police work consumed both time and resources. It would take 30 people up to three weeks just to match a fingerprint. Lynch says after he turned up "with a machine that could do the job in five minutes" it was an easy sell. "They wouldn't have cared if it was made by elves," he recalls. Autonomy's APCM software drives Holmes 2, which not only matches fingerprints but also spots connections in disparate witness statements and police reports.
Big business is now Autonomy's main revenue generator. The software's knack for spotting value amid reams of workplace conversations has led to lucrative knowledge management contracts with corporate titans, including Citigate and Shell. But an ability to reveal hidden patterns allows the firm a foothold in unexpected markets, too. "We're the arms dealer of British politics," says Lynch. Autonomy supplies technology to all three main political parties, which use it to monitor online, social media, and TV broadcasts. "Every time someone says something, the system finds out where and when they've said something different," explains Lynch.
Clients with call centres, such as Vodafone, use the technology to listen in on calls and automatically alert a supervisor if a customer becomes irritated, or even suggests appropriate solutions to the problem. In financial services, where the burden of regulation is heavy, the software can be used to make sense of critical phone calls and emails. It can "listen" to a million hours of conversation, says Lynch, and deliver the 10 hours most likely to mention your chosen subject. The evidence to convict Jérôme Kerviel, Société Générale's rogue trader, was collected using Autonomy technology.
Ten hours of information is normally overkill, says Lynch. "The reality is that once you get past 15 minutes none of it will be relevant [but] it's a million-hour job filtered down to minutes." Humans are slightly more accurate than machines at sifting data, but only at first. "Then they get bored," says Lynch. "The bigger the task the more accurate the machine will be over time."
Lynch isn't advocating the replacement of humans by machines, but he does imagine a world in which leaner, more productive companies use computers to perform menial, time-hogging tasks. It's worth reiterating that Autonomy is hardly flabby itself. "Our gross margins are 92 per cent. I would be no good at running a biscuit manufacturer," he says. That Autonomy is therefore the perfect embodiment of its own principle is a useful sales story.
Lynch is still the company's USP. There are no other founders running UK companies of Autonomy's size in any industry. The technology sector, in particular, is known for replacing founder-entrepreneurs with more experienced career chief executives on the assumption that their skills are better suited to corporate terrain. The UK tech scene is heavily populated with half-grown companies, anticipating foreign investment to take them to the next level. As investor Sherry Coutu put it at a Business Leaders Network event: "We don't need more start-ups, we need more finish-ups."
Lynch agrees—and disagrees. "Look at ARM. That's a leader. It's world-class. There are five or six other firms that are on their way. That pipe is unblocked." But, he adds, British technology companies could achieve a bigger footprint if they only learnt to appreciate the role of marketing. "We have wonderful technology in the UK. The best place to do R&D is the UK. But we struggle to produce businesses. Marketing is telling a story, presenting it. You almost get the idea at Cambridge that that's cheating."
Like all tech entrepreneurs, Lynch is passionate about the industry's role in spurring economic recovery. And he's optimistic about the crop of talent charged with driving the revival. "How do you create fundamental wealth? You can grow it, but there's not much space left for that. You can dig it out of the ground. The other way is to use our nation's incredible natural resource of creativity and knowhow to turn something into something else more valuable. We are sitting on a gold mine of talent."