Blog article
See all stories »

Why HP might have underpaid for Autonomy

In October 2011 amid a blaze of publicity H.P., a company which mainly produces hardware and devices like inkjet and laser printers, announced it had agreed to pay $11.1 bn for Autonomy, a UK company that produced sophisticated software. Dr Mike Lynch the CEO of Autonomy, said the deal had opened a “fascinating new chapter” for his company as part of a bigger, global corporation.

But a little over a year later H.P announced that it had written off $8.8bn of the $11.1bn it paid for Autonomy, of which it says $5bn was "linked to serious accounting improprieties, misrepresentation and disclosure failures". Really? Looking at matters in a realistic manner $11.1 bn for Autonomy was probably a bargain if you had studied and modelled the long term value and potential scope of Autonomy’s extraordinary patent portfolio, Especially if you also understood the stickiness of Autonomy’s software and its effect on the businesses  and government agencies that used Autonomy’s technology.

Here in the UK amongst the technorati it is fairly well know that Dr Mike Lynch did his PhD thesis on Bayes’ Theorem and in 1996 set up a business with Richard Gaunt that eventually became Autonomy. Along the way in 2004 Richard Gaunt on behalf of Autonomy filed a provisional patent application for “Methods and apparatuses to generate links from content in an active window” which later became a full patent filed by Dr Mike Lynch and two other Autonomy staff that was granted - Method and apparatus to link to a related document

This granted patent is one of what is believed to be around 170 patents owned by Autonomy in countries around the world which cover all sorts of core technologies which are going to lock up this area for the next twelve to fifteen years. US Patent 7272594 makes reference to Bayes’s Theorem (how you make choices as data becomes available) and Claude Shannon's principles of information theory (mathematical limits of certainty). This patent, in my view. is a fundamental building block for all technology in this sphere of activity. No rival company (Google, Microsoft, IBM, Amazon etc) could afford to enter this field with any cloud service offerings because Autonomy's patents will enable them to be injuncted with dire consequences for all of them - billions in damages etc. This is not technology where it can be claimed the consumer is being denied rights because of a lack of a competitive market in the patents through licensing - Autonomy is a business to business offering whose technology falls full square within the lawful monopoly for patents permitted under the US Constitution - a monopoly right which is respected across the world. Rivals who choose to try and compete with Autonomy are likely to suffer the same fate as Kodak did when it tried to produce an instant print camera in breach of Polaroid’s patent portfolio - global injunctions and heavy damages

So why are these rights so valuable even if Autonomy is not yet fully enforcing its patents? The issue is one of stickiness. Around the time of the takeover by HP Autonomy had  20,000 clients, with management contracts for giants such as Citigate and Shell. Autonomy also drives the UK police's Holmes 2 system, which can tie together fingerprints, witness statements and police reports. It can sift emails, documents and even phone calls and elucidate the meaning inherent in them. It allows customers to search and categorise unstructured information - such as e-mails, phone call logs, pictures, film clips, anything that has not been organised into a database. It has been heavily used by banks and other large corporations preparing for class action lawsuits, helping them find all the documents needed for trial. Société Générale, for example, installed the software to trace the actions of rogue trader Jérôme Kerviel. Any company which is hit by a "litigation hold" notice had better turn to using Autonomy's software or face the consequences.

Stickiness means that once a company starts using Autonomy's software it cannot stop - the cost of leaving is too great. Even if someone could lawfully design around Autonomy’s patents it will not be able to capture an existing Autonomy customer because if the customer wishes to move to a rival supplier the cost of doing would be  prohibitive. If your business depends upon having a computing network running 24/7 (which today covers all financial and insurance services) and you have purchased an indexing package from a company which you have run for some time - then you may never be able to move. This is because of what Donn Parker of SRI International called MTBU - maximum time to belly up. He found that businesses that relied on computers for the control of their cash flow fell into catastrophic collapse if those computers were unavailable or unusable for a period of time. How long? By the late 1980s it had fallen from a month to a few days. That’s not a good thing; it meant that a collapse of the computers that any UK clearing bank relied on would destroy it in less than a week.

Today moving from Autonomy to a rival supplier (even if one lawfully existed) would cause the company to have to survive without its essential computing services for a protracted period of time during which time the business would be vulnerable to collapse. No business which uses e-mail in its day to day operations could afford to move. It would be like being told as a householder that you could switch electricity suppliers if you were prepared to live without electricity for 3 months - not many households would be prepared to do so.

But the opportunities and benefits which arise from using Autonomy’s software in a financial or services business outweigh all the risks - imagine compliance departments running Autonomy's software to stop another financial collapse, another Barings or UBS case. Autonomy was never holding its major clients to ransom with its sales contracts but it had been able to use a very aggressive way of estimating the value of each customer - to take account of the fact that every customer Autonomy acquired will never leave - i.e. there will be no churn. This is what H.P present management team of hardware salesmen and technocrats have failed to understand. And this is why H.P. may have got a bargain in paying just $11.1 Bn for Autonomy.

Alistair Kelman
Barrister and Forensic Computing Expert Witness

For the avoidance of doubt I say the following:

  1. I am not employed by either H.P. or Autonomy nor have I ever had any business dealings with them - but I am commenting as a forensic computing expert witness and barrister.
  2. I do not own nor ever have owned shares in either Autonomy or H.P.

Autonomy - now part of the HP family

Comments: (1)

Matt Scott
Matt Scott - Fiserv Inc - London 04 February, 2013, 08:12Be the first to give this comment the thumbs up 0 likes

I believe a fair share of the value within the company was in the expertise of the staff - most of which have left post-acquisition as the division was not managed in line with their expectations.  I wonder how long it will be before we see Autonomy 2 the sequel (probably after a period of contractual restriction has elapsed no doubt)?