Lynch mob

Any aspiring criminals who want to avoid investigation should follow the lead of Autonomy, and do business with the investigating authorities.

A criminal investigation into Autonomy by the Serious Fraud Office has been thrown into confusion just as it began after the UK agency admitted that its £4.6m contract with the software-maker could present too much of a conflict of interest for it to continue pursuing the probe. Which is a shame, but particularly when the opinions voiced in America last year are considered.

Autonomy was founded as Cambridge Neurodynamics in 1991 by Michael Lynch, a Cambridge-educated computer scientist. Nothing wrong with that.  There are lots of us.

The company was based on the then-hot concept of Bayesian search, named after 18th-century mathematician Thomas Bayes, and ultimately developed an all-encompassing software package it called IDOL — Intelligent Data Operating Layer.

Hewlett-Packard say they "stand behind" IDOL.  And so they should; otherwise they would have to write off the rest of the $11 billion they paid for Autonomy.

But, in short, the naysayers say that there isn't really much to the concept, more just smoke and mirrors that Autonomy has used to make people think they have got something very impressive.  It's basically a search engine, like Google's and a host of others, but unlike Google's it just works on a company's own data, and doesn't have the portal/advertising business model of Google.

So how did Autonomy show such big numbers? They would go to customers and offer them a deal they couldn’t refuse. Say a customer had £5 million and four years left on a data-storage contract. Autonomy would offer them, say, the same amount of storage for £4 million but structure it as a £3 million purchase of IDOL software, paid for up front, and £1 million worth of data storage. The software sales dropped to the bottom line and burnished Autonomy’s reputation for being a fast-growing, cutting-edge software company a la Oracle, while the revenue actually came from the low-margin, commodity storage business.

Lynch’s management team also was practiced at the art of wringing attractive-looking growth out of a string of ho-hum acquisitions. The typical strategy was to bolt IDOL and other software onto a target’s existing products and try and convince customers to pay more for the “new” products.

If that failed, they’d milk the existing customer base by halting development and outsourcing support, my source says, using the cash from the runoff business to fund more acquisitions.

Which to come back to the SFO, tells us why they might feel a conflict of interest.  Not because they are a loyal and willing customer of Autonomy, but because they spent a fortune on software that they don't use as part of a ho-hum data storage contract.