• Home
• About
• Writers
• Landscape
• Wizards Index
• Fraud
• Wanna Be Like Larry?

Deconstructing HP Autonomy and Its Eight Answers

All Things Digital ran a story called “Eight Questions for Hewlett Packard Software Head George Kadifa.” Let me nudge aside any thoughts that the interview and the questions were presented as public relations and marketing. I want to view the comments or “answers” as accurate. Once I have highlighted the points which caught my attention, I want to offer some side observations from my goose pond in rural Kentucky.

First, there were two passages which addressed the $12 billion Autonomy purchase.

The first was information about a recent planning meeting. The Autonomy staff were on deck and ready for duty. The key statement for me was this one:

Basically when you look at Autonomy, the core unit is the IDOL Engine, which is the unique capability of meaning-based computing. We’re going to double down on that. In our labs in Cambridge, England, we have 40 or 50 mathematicians writing algorithms. And we’re going to build a team here in the U.S. to productize it and create a platform around it because it has that potential. Frankly, the way Autonomy was managed previously, they put a lot more emphasis into enabling applications, which was fine, but our belief is that there’s a broad agenda, which is creating a platform around meaning-based computing. So we will maintain those apps, but at the same time we’ll open up the capabilities to a broader set of players outside HP.

Makes sense. Pay $12 billion for IDOL. Leverage it.

The second was semi-business school thinking about how to grow Autonomy’s business. Here’s the passage I noted:

In Europe, they tend to make things complex in order to create more value. For example, they saw the IDOL engine as too complex to just give it to people. Instead they thought they should acquire vendors and then create value by enabling applications. Here we take something that’s complex and we ask how we might simplify it in order to give it more scale for a bigger market. So some of that difference was cultural, and some of it was that I think they fell in love with these acquisitions. … We think Autonomy’s technology has broader implications.

I urge you to read the full “eight questions” and the answers. Now my observations:

1. Productizing IDOL or any search engine can be difficult. When I use the word “difficult,” I mean time consuming, expensive, and timetable free. Buying a search engine and sticking it in a product or service looks easy. It is not. In fact, IBM has elected to use open source search to provide the basics. Now IBM is working hard to make money from its value add system, the game show winner Watson. There may be a product in “there”, but it is often to find a way to make money. HP has to pay back the $12 billion it spent and then grow the Autonomy business which was within shouting distance of $1 billion.
2. The notion that Europeans see the world differently from HP is interesting. I am not sure how European Autonomy was. My view is that Autonomy’s senior management acquired companies and did upselling. As a result, only Autonomy broke through the glass ceilings behind which Endeca, Exalead, ISYS, and Fast Search & Transfer were trapped. Before applying business school logic to Autonomy, perhaps one should look at how other acquired search vendors have paid off. The list is, based on my research, a short one indeed. Microsoft, for example, has made Fast Search a component of SharePoint. With Fast Search nearing or at its end of life, Microsoft faces more search challenges, not fewer. HP may find itself with facing more challenges than it expects.
3. The notion of “broader applications” is a popular one. Dassault Systèmes, acquired Exalead, which is arguably better and more recent technology than IDOL. But Dassault’s senior managers continue to look for ways to convert a more modest expenditure for Exalead into a river of revenue. Dassault has a global approach and many excellent managers. Even for such an exceptional firm, search is not what it seemed to be; that is, a broad application which slots into to many customer needs. Reality, based on my research for The New Landscape of Search, is different from the business school map.

HP is making an trip which other companies have taken before. My view is that HP will have to find answers the these questions, which were not part of the interview cited above:

First, how will HP pay off the purchase price, grow Autonomy’s revenue, and generate enough money to have an impact on HP’s net profit? My work has pointed out that cost control is the major problem search vendors face. It takes money to explain a system no matter how productized it becomes. It takes money to support that technology. It takes money to enhance that system. It takes money to hire people who can do the work. In short, search becomes a bright blip on most CFOs’ radar screens. HP may be different, but I am not sure that the cost issue will remain off the radar for very long.

Second, IDOL is a complex collection of software components. The core is Bayesian, but much of the ancillary IDOL are the add ons, enhancements, and features which have been created and applied to base system over the last two decades. Yep, two decades. In search, most of the systems which have figured in big deals in the last two years date from the mid to late 1990s. The more modern systems are not search at all. These new systems leap frog key word search and push into high value opportunities. HP may be forced to buy one of more of these next generation systems just to stay in the “beyond search” game.

Third, HP is a large company and it faces considerable competition in software. What makes HP interesting is that it has not been able to make its services business offset the decline in personal computers and ink. HP now wants to prove that it can make services work, but as the Inquirer pointed out in mid August 2012:

HP’s write-down of EDS might have resulted in just a paper loss – the firm didn’t actually lose $9bn in cash – but it provides an insight into how a decade of mismanagement has left HP in a bad situation. The fact is that HP cannot lay the blame on diminishing PC sales because its enterprise business, printing and services divisions all reported losses, too. For HP to write down the purchase of EDS, a company it paid $13.9bn for just four years ago, strongly suggests that those who were at the helm of HP in the run-up to that acquisition simply had no clue as to how much EDS was really worth and how to incorporate the company into HP. The value of any company can go down over time – just look at AOL, Microsoft or Yahoo – but for an established business such as EDS to be overvalued by almost $10bn just four years after being acquired is nothing short of gross incompetence by HP in both the purchase and the subsequent handling of the firm once it became a part of HP.

I don’t fully agree with the Inquirer’s viewpoint. But one fact remains: HP must demonstrate that it can manage a complex business based on IDOL, a technology which is not a spring chicken. The man who did manage Autonomy to almost $1 billion in sales is not longer with HP. In the history of enterprise search and content processing, Mike Lynch was unique. Perhaps the loss of that talent will continue to impact HP’s plans for a different approach to the market for Autonomy’s technology?

Life extension treatments are available, but these often do not work as expected and can be expensive. Most fail in the end.

Stephen E Arnold, September 25, 2012

Sponsored by Augmentext

Comments

Copyright © 2008 Beyond Search · ArnoldIT.com · Disclaimer · Revolution News theme by Brian Gardner · Log in