Blog - Commentary

HP and Autonomy -- lacking the innovation sauce

As technology luminaries row potentially all the way to the courts over their recent deal, Nigel Cannings says far more important issues threaten business models

In a stinging rebuke to California technology giant Hewlett-Packard, Mike Lynch, the former head of British software company Autonomy, said he believed in ‘innovation in all areas except accounting’.
That quip touches on the core of HP’s real issue in its high-profile row and possible legal action over its purchase of the Cambridge founded business.
A few weeks ago at HP’s flagship event, Discover 2012, the company’s chief executive and Mr Lynch’s chief tormenter, Meg Whitman, said: ‘It’s hard to kill founder DNA.’ That was a reference to Messrs Hewlett and Packard, who created the business in a shed in 1939. ‘Innovation is alive and well at Hewlett-Packard,’ she said.


There is an air of unreality to HP’s claims that Lynch & Co are responsible for a $5.5 billion write-down, which would make it one of the biggest corporate frauds -- or due diligence gaffes -- in history. Given the sums involved, it’s going to be a fees bonanza for any law firm and other professional who isn’t conflicted by having been instructed on the acquisition.
But that’s not the real story of this fiasco.
My father is a 45-year veteran of the software industry and a serial entrepreneur. His last business was acquired for a substantial sum by a North American company, so he was speaking from experience when he commented: ‘When will these large corporations learn the three golden rules of acquiring an entrepreneurial company? First, work out what it was that made it so successful and bottle it. Second, tie in the technical and marketing resources that achieved that success. Third, allow the founders to do what they do best -- don’t try to reinvent them.’
HP has ignored these rules.
Mr Lynch must be a fantastic marketeer and entrepreneur. Anybody that can take a company from zero to nearly a billion dollars revenue in fewer than 10 years is something special. Most entrepreneurs are difficult to manage; it’s not in their character to take orders. But it beggars belief that HP could let him walk away with all his money without using his abilities to continue to grow the business and at least prove its value.
Autonomy was at its grassroots, a highly technical company. In their early days, companies such as Autonomy have a hard core of developers -- perhaps fewer than half a dozen people -- that can interpret the entrepreneur’s market ideas and turn them into high-tech solutions. It is usual to ensure that these people have a financial interest in the company’s success to ensure that they are not head-hunted by competitors. Looking at Mr Lynch’s previous comments about technical people leaving, it seems that HP has lost the technical engine room he created.


My father made another interesting observation. Hewlett-Packard’s promotional catch-phrase is: ‘HP invent’. One thing my father has learnt over the years is that a business must not just invent, but re-invent and do it every five years or so -- or die. HP seems to have wanted to do that by acquiring Autonomy, and then promptly achieved the near impossible feat of shooting itself in the foot and the head simultaneously.
The computer industry is littered with seven-year wonders that shine brightly before they die or are acquired – and the pile has been growing exponentially since the first dot com bubble. However, there are some obvious exceptions:

  • Oracle is still driven today by its co-founding entrepreneur.
  • Microsoft still dominates the market and continues with technical innovation.
  • IBM was a computer mainframe company that re-invented itself as a PC manufacturer and again as a services company.
  • Apple – which rose to fame in the 1980s -- lost its impetus when the mercurial Steve Jobs left. On his return, he re-invented the company more than once before his death.

In every case, those businesses sought to re-invent themselves, not always succeeding at first, but sometimes doing it in ways that no-one expects (if you don’t believe me, you obviously don’t have a teenager with an XBox addiction).
Autonomy is not like any of these, and when you look closely at its products and market position, it could be seen as a one-trick pony. Its retrieval and analytical software is excellent, but after five good years, has it saturated the market rather than grown, meaning that HP has bought just a maintenance stream?

Voice retrieval battleground

Autonomy tried to make voice retrieval the new search battleground, leveraging technology that has been in its product set for years. However, what it has not done is try to re-invent core products to take account of new trends in technology that are about to blow its business model out of the water.
The Dodd-Frank legislation in the US will transform the need to store and retrieve telephone calls and voicemail. In processing terms, voice analysis is expensive -- hundreds of thousands of decisions have to be made every second just to process a single person’s speech. Imagine trying to do that for a whole bank with 50,000 employees, simultaneously.
The future HP should be looking at -- and probably would have been looking at if it had kept the key Autonomy players -- is in supercomputers. Not those room-filling machines of 1960s sci-fi films, but cheap graphics cards that run 4 trillion calculations a second (HP sells them for $7,699 a pop), and which currently power the world’s fastest supercomputer.
These cheap ‘co-processers’ allow the running of many tasks in parallel that otherwise have to be run individually. I have spent the last five years working with a team to develop robust voice and text analytics for the financial services industry, and we’re looking at improvements of up to 50 fold in speed by using these cards. Put another way, that’s potentially 50 fewer servers a business has to buy, which cuts down on space, cost and energy.

Supercomputer future

It will also be crucial to analyse other areas such as improving voice capture quality, better training of voice algorithms, and cleaning up audio files. Many of these will only work in real-time if supercomputer resources are thrown at the problem.
Autonomy has no obvious offering in this arena, and if its core technical people have left, there is little hope of producing one. With the implementation of Dodd-Frank and the need for all financial institutions to record voice traffic, a major slice of Autonomy’s business may not even be paying its maintenance in the not-too-distant future as other players step firmly into its core market.

Posted by:


13 December 2012

Related stories

Editor's picks