Getting back to basics

Hailed as the UK's first Internet billionaire at the height of the dot-com boom, Autonomy chief executive Mike Lynch has ridden the stock market roller-coaster and so far lived to tell the tale.

Unlike some of the flakier home-grown tech firms and dot-coms, Autonomy has a business proposition that seems to fulfil a genuine need -- the ability to search and manage the mountains of unstructured data businesses are increasingly inundated with. This, combined with the fact Autonomy was born out of Lynch's research at Cambridge University on the probability theorem's of an 18th century mathematician, rather than a venture capitalist's chequebook, means of the 40 or so companies that claimed to do the same job as Autonomy at the height of the boom, Lynch's outfit is the only one still around.

Customers of Autonomy's various search and data management products -- including its Intelligent Data Operating Layer (IDOL) Server which the company claims uses pattern-matching techniques to understand information in context -- include Ford, the BBC, and RoyalSunAlliance. The ability to automatically search through thousands of text documents, voice files, and video as well as other kinds of unstructured data  has also seen the US Department of Homeland Security become a recent customer as it seemingly battles to monitor just about everything, everywhere.

ZDNet UK caught up with Lynch ahead of his talk at the European Technology Forum UK Technology Summit event (ETF is owned by ZDNet's parent company CNET Networks UK) in London this week, to get his views on the current state of the IT industry.

When you address a room full of IT managers -- what's the main message that you try to get over to them?
I was at a business talk recently and some one bought up the very first business computer which was Leo and the fascinating thing about Leo was that, although it was built by a load of boffins in Cambridge, it was built for Lyons to get their tea cakes out on time. The one thing you realise when you read about it is they did an incredibly detailed analysis of their business and how the computer could help their business and what questions they need to get answered. I think that's a crucial thing for IT managers in the current climate: don't think about the CRM system but imagine you're the operative working that system -- what are you doing to make that work. Then you can get a real idea of the benefit or otherwise of using that technology.

But a lot of people put the failure of CRM implementations down to issues around company culture, specifically sales culture, rather than problems with the technology itself?
I'm not sure that's true. I don't want to get into a religious war about whether it's technology or not technology but what I don't think people sat there and asked: "Right I am going to be an operative in a call centre using the CRM system today -- what do I have to do?" If you actually approach some of those CRM systems in that way it's easy to see why some of the ROI isn't being delivered.

The industry needs to focus on making sure its delivering real benefits to the customers. A company buys in a big CRM system for millions of dollars and ends up employing lots of people to take in all the information coming from customers, emails and phone calls and so on, and enter it into forms at the front of the CRM system. By the time you take that cost into account you wipe out the ROI.

Or you've got a supply-chain management system that looks great until you realise every so often an order has something written on it in pencil -- like make sure these screws aren't magnetised -- and the cost of dealing with that one order actually destroys the savings on the ten other orders that have gone through. There's often no way of dealing with that information in a supply-chain management system

So what's the solution?
A lot of it is about going back to first principles and actually listening to customers. A lot of this is about not understanding the information inside companies and the other is the whole problem of computer systems. What you've got inside a modern enterprise looks like a spaghetti rat's-nest of different apps taking data from each other and all trying to be connected. What that means is that as soon as anything goes wrong or changes you have to get loads of consultants in and there is this explosion of complexity. What we have to do is make sure that the industry sees this as a problem. So if the server goes down -- the software just keeps going on another server.

There's a lot of work in what's called self-aware software. You'd be surprised how many enterprise systems just fall over if the disk gets full. A self-aware system realises the disk is getting full, obviously tries to tell someone, but if the situation doesn't get any better makes alternative arrangements. I think the IT industry has almost been in a conspiracy not to face this problem because at the end of the day it means you've got more billable hours if you have to go in and sort it out. I think we, as an industry, have to face that that's not the way to treat customers.

So you've ridden the stock-market roller coaster through ups and downs -- are you seeing any sign of a recovery in tech spending?
The situation we are seeing in Q4 is that there's definitely a recovery in activity levels. But what you don't know is, is that robust or fragile? As soon as we get a bad bit of economic news will everyone crawl back in their shells?

We've been very fortunate on the stock-market roller-coaster. We have been profitable all the way through it and have plenty of cash - so we don't actually need it. But if you're company was loss-making and you had to raise money on the stock-market. It would have been a nightmare. The main thing is just head-down and carry on and it all sort's itself out in the end. It's actually quite a nice environment at the moment. In bubble days everything was so crazy. We had something like 40 companies that at some point said in their marketing they were like Autonomy and they've all gone out of business now.

The next big IPO everyone is waiting for is Google -- do you think they've been right to be cautious about going public?
Google has been very fortunate in being private through the downturn and I think they are quite wise to this and have tried to hang it out to the last minute but I think there are all sorts of reasons why they might have to go public very soon.

Increasing competition is probably one of those reasons. Microsoft has publicly voiced its intentions to get into search in a big way with Longhorn -- the next version of Windows.
Search is a very small part of our business. The only thing I would say is that having been in the industry, I have seen Lotus and Oracle and Microsoft announce initiatives in non-structured data for the last 12 years and the announcement seems to carry more weight than what arrives.