“When people hear your accent, they think you should be doing bone china, or castle tours,” says Autonomy’s founder and CEO, Mike Lynch. Apparently, people are still surprised that someone with a British accent – Lynch was born in Ireland and brought up in Essex – could have founded a successful software company, let alone a worldwide market leader.
The story of Autonomy’s rise to prominence is becoming something of a legend. It all started at Cambridge University where Lynch studied Electrical and Information Sciences and became Dr Lynch, with a Ph.D. in adaptive techniques in signal processing and connectionist models, as well as a research fellowship in adaptive pattern recognition. Some of his work was inspired by statistical pattern-matching algorithms developed by 18th century Presbyterian minister and mathematician, Thomas Bayes.
But were it not for a music promoter that Lynch met in a bar, who loaned him £2,000 to start the company, Autonomy might never have happened. Fast forward to 1991, and Lynch was being hailed as the UK’s first technology billionaire, as his sizable chunk of Autonomy’s stock rocketed in valuation amidst the dot-com euphoria.
Enterprise Search ‘War’
For its next trick, Autonomy won the enterprise search ‘war’. This was a market that sprung up partly in response to legislative and compliance challenges in the US and elsewhere (see end sidebar, The Rise of E-discovery). But also simply as a result of the frustration of many employees that they are unable to quickly find the corporate information that they want.
For a time, Autonomy’s closest search rival was Verity, until Autonomy bought the company for $500m in November 2005. After that, it was the Norwegian company, Fast Search and Transfer (FAST) that seemed to be the nearest rival. The two went through a particularly bitter phase after Autonomy bought Verity; FAST issued a press release offering, “Safe passage to Verity customers and partners who are facing uncertainty arising from the acquisition of Verity by Autonomy.”
Autonomy responded with more aggressive tactics of its own, putting out a release offering a, “Secure transition route for FAST customers and partners who may be concerned about Fast’s publicly-known accounting questions.”
FAST did indeed fall foul of its own accounting, when it was found to have been signing off so-called ‘Memoranda of Understanding’ as done deals, when in fact the revenue should not yet have been booked. In an interview with CBR at the time of the crisis, FAST CEO John Lervik told us, “It’s my responsibility, there are no excuses.”
But the damage was done, and in January this year FAST was bought by Microsoft for $1.2bn, though it is being operated as a subsidiary, of which Lervik is still CEO.
But even at its peak, FAST was not making anything like Autonomy’s revenue. In the last quarter as an independent entity before it was acquired – the third quarter of 2007 – FAST announced sales of $35.6m, up just 4%. In the same quarter of that year, Autonomy announced its sales rose 49% to $89.6m.
Autonomy’s growth has continued since then: in its third quarter of this year, announced in September, it posted sales of $127.1m, up 42%. In a recent report by IDC, the research firm described Autonomy as, “the clear revenue leader” in enterprise search. It said the search and discovery software market grew to $1.8bn in 2007, a growth rate of 28%.
The IDC report, written by Sue Feldman, vice president for search and discovery technologies, found that the market will continue to outpace the software market as a whole in 2008. “There are a number of reasons for the continuing growth we see in the search market,” said Feldman. “While the economy is slowing, the need to manage non-transactional information has become critical, and enterprises are investing in the next big information-centric computing platform.
“This market is also driven by government investment in intelligence applications,” said Feldman, “and by industries such as pharmaceuticals and energy that are heavily dependent on dependable, easy access to good information from multiple sources.”
Speaking to CBR this month, Lynch said he worries that FAST’s previous financial scandal will do little to help the credibility of European software firms: “This is the problem,” he told us. “We have FAST, Lernout & Hauspie, Baan: it doesn’t help.”
Lernout & Hauspie was the Belgian speech recognition firm found to have been cooking its books: founders Jo Lernout and Pol Hauspie, as well as former CEO Gaston Bastiaens, were arrested in what was considered one of the largest corporate scandals in history prior to Enron.
Baan Software was also embroiled in scandal, restating accounts to the tune of $43m after it emerged that it had been recognising sales to its own distribution company. The scandal brought Baan to its knees, and it was bought by Invensys.
Besides, Lynch still argues that FAST’s technology did not solve all of the requirements of e-discovery: “You have to run on Unix – to meet the regulations you have to run across everything,” he says. “You have to be able to produce any electronic document from any system, within 99 days. If you fail you will almost certainly lose the lawsuit and it could become a criminal case. We’re seeing customers who have bought one search technology having to buy another one.”
Lynch also asks, “If FAST had good technology why was it not doing well? This is a good business to be in if your technology works.”
New Challenge
The great irony in all of this is that Lynch does not want Autonomy to be pigeon-holed as an enterprise search company. It’s a bit of a tricky one though, since the company itself highlights pronouncements made about the enterprise search space by analysts, and at one time seemed to have picked FAST as its nearest standalone competitor.
In contrast, FAST CEO John Lervik was happy for his firm to be called an enterprise search vendor, telling CBR, “We think we can build a $1bn revenue company ‘just’ in search. After all, Oracle built a many-billion dollar company ‘just’ doing a database.”
IDC’s Feldman though says that, “At this point, it is clear that Autonomy should no longer be considered purely a search vendor. It builds search-based applications to answer market demands for better information-centric software.”
What does that mean? Autonomy’s website explains: “Autonomy’s software powers the full spectrum of mission-critical enterprise applications including pan-enterprise search, proactive information risk management, information governance, e-discovery, consolidated archiving, call centre solutions, rich media management, security applications, customer relationship management (CRM), knowledge management (KM) and BPM [business process management].”
Says Lynch, “If you can understand meaning, then you can create a very different search engine. But the real benefit is not searching for information, it is in doing something. We are seeing an explosion in software applications that can understand meaning [thanks to their licensing of Autonomy’s technology] and help you act on that.”
Lynch says Autonomy now has in the region of 500 OEM customers, writing applications that embed Autonomy’s Meaning-based Computing, or MBC. Their own software products rely on Autonomy’s pattern matching algorithms to extract ‘meaning’ from unstructured information.
For examples of organisations putting that Meaning-based Computing into action, Autonomy points to the US Department of Homeland Security, which relies on MBC across 21 agencies to monitor suspected terrorist groups, create a consolidated terrorist watch list, and alert authorities in real-time to potential terrorist activity.
The Ford Motor Company uses MBC to transform the text, audio and video files in its research libraries to meaningful reference material so around 150,000 employees can get up to speed on new projects.
“What we do is fundamental to computer-based information processing,” says Lynch. “When computers first emerged in the 1960s it was all about trying to take real-world information, which was complicated and subtle, and simplify it in a structured relational database for use by a computer system. Autonomy turns that on its head. Rather than trying to simplify the real-world to fit neatly inside a computer database, we allow the computer to understand the meaning of human-friendly, unstructured information.”
One of the differentiators over its smaller rivals in the space – including Endeca, IBM (smaller in terms of search, at least), Google Enterprise, Simplexo, Sinequa, Recommind and many more – is the list of supported file types that can be handled by Autonomy’s IDOL platform. “By supporting more than 1,000 different data formats, including structured, semi-structured, and unstructured data, located across 400 different content repositories, Autonomy can search all categories of information repositories in an organization,” the company says.
Today, with Autonomy likely to break the $500m barrier this year, Lynch confesses that there are days when he, “Wish[es] it was like the old days, with five people in a room. But in the early days we had no funds, no distribution, and nobody wanted to talk to us. We’re now in an incredible position that we have a company that can back up our ideas, a market that wants to buy them, and great people to do it all.
“Today, I don’t run Autonomy,” Lynch concludes. “We sit round a long table when we are working out what to do next. You have to in tech.”
Sidebar: The Rise of E-discovery
While previously it was acceptable for unstructured information – the likes of email, Word documents, PowerPoint files, images, spreadsheets, even voicemails – to be distributed throughout an organisation with little to help that company find them when they needed to, compliance demands particularly in the US put paid to this state of relative anarchy. A discipline that became known as ‘e-discovery’ was born.
In early December 2006, several key changes to the US Federal Rules of Civil Procedure (FRCP) involving electronic discovery became law. There was a fundamental shift from electronic data being included by exception only, to now being excluded by exception only. This meant that all discovery requests include electronic records unless otherwise specified.
According to ESG Research in 2006, 91% of organizations with a workforce of over 20,000 employees had been through some form of electronic discovery event over the previous twelve months.
Many companies have already suffered severe financial sanctions for non-compliance to an e-discovery request. For example, Bank of America Securities drew a $10m fine over SEC allegations that it had failed to produce an employee email thread related to an issue under investigation. Morgan Stanley lost a $1.45bn case against it, in part because the jury believed that the company had attempted to obstruct justice by failing to produce electronic documents.
In the UK, the Combined Code on Corporate Governance has also had a major effect.
Meanwhile the financial services industry has had particularly stringent regulations for many years, and UK watchdog the Financial Services Authority is widely regarded to now be growing real teeth. Also in the UK, the Data Protection Act and Freedom of Information Act have had an effect on the ability of citizens to request information that involves them, and has led to a heavy burden on local government.
Analyst firm Gartner predicts that the e-discovery software sector will be worth over £385m this year, up from £266m in 2007.
