After 38 years with British Telecom, Peter Cochrane resigned last November as the giant telco's chief technologist and head of research. He left to work full time of ConceptLabs, a technology incubator that he helped to found in 1998.

On the fact of it, this was an unusual move. At age 54, Cochrane was at the peak of a successful career. He had just capped a long string of awards for technological innovation by receiving the Order of the British Empire and the Institute of Electrical Engineers' Millennium Award. He held a professorship at the University of Bristol and was a fellow of several British and international engineering and scientific associations.

Nor did the economy look right for such a step. The dot-com rose that had been blooming so brightly in 1998 was looking more than a little faded two years later. So why leave BT? For one thing, the move was not as sharp a departure as it appeared to be. Cochrane had been a part-time technology incubator for years, at BT and on his own account. In a sense, founding ConceptLabs was as much a continuation as a break.

But it also was a break. He'd joined the company in 1962 as a lineman and maintenance technician and spent his entire adult life there. He compared walking out of BT to cutting off his right. But he has also compared it to removing a very tight pair of shoes.

CSC World: Founding ConceptLabs isn't an entirely new direction for you.
No, it isn't. I'd been running all kinds of experiments at BT. But there were others I couldn't run, so I used my own money to experiment in setting up companies. Losing my own money felt entirely different from losing BT's money, and I learnt an awful lot.

Out of the first 20 companies I started I lost five, five are really very good, and a couple are crackerjack. One of the good ones is ebookers.com, which is on the Nasdaq and is going live on the UK stock market very shortly.

CSC World: How did ConceptLabs start?
In 1998, Rao Machiraju called to tell me that the advanced technology group at Apple was being pushed out and asked if I could help. BT wasn't interested, so I put my hand in my pocket and put as much money on the table as I could. Rao did the same. We got more money from friends and family. Then we got a million of venture capital and we started. By Christmas 2000, we'd got six companies into really good shape. They're in the United States, India, Singapore and the UK.

Depending on the state a company's in when we get involved with it, we take a slice of between 5 and 20 percent. When we add up all our fives and 20's and apply a caution factor of two, we see ConceptLabs with a market cap of about US$32 million. That's quite respectable for a company of 15 people.

CSC World: Very respectable. Which suggests that you have a method for incubating technology companies that works rather well. Care to let us in on it?
The questions I ask are very simple: "Who's got the money? What's their problem? Can I fit it? Can I do it at a profit? Can I do it in time?" It's like putting together a technological SWAT team: Helicopter up, look at the hole, identify a target, make sure no one else is there, get a customer lined up, drop in a team, and very quickly engineer a solution.

ConceptLabs has simple ground rules. First, we will not enter any space occupied by anyone else. We would not, for example, work on Bluetooth or WAP. Second, we will only go with a technology that we can bring to market in nine months or less. Third, we will only begin a project if there's the prospect of building a half-billion dollar company at the end of the line.

CSC World: What advice would you give to people who are staring a company?
That you have to be cruel to be kind. Sometimes you put a team together and find that actually they're a car with only three wheels. The missing fourth wheel can be anything: a CEA who's not sufficiently mature, people who aren't technically astute, a CFO who's never done the job before. If I err on the kind side, as I have in the past, what happens is they stagger into being, stumble around for a while, then die a cruel death. So it's better to be cruel right up front and say "No. Either we fix the problems now or we're not going to do this."

CSC World: Is the current crop of teams showing the same kinds of deficiencies?
It's getting a lot better. The UK starts twice as many companies per year per capita as the United States. But the failure rate is horrendous. Here's the UK attitude, "I'll start a business and if I can get it big enough, I'll go off to the golf course every Wednesday." That's not healthy.

The Americans have grand ambitions. They seem to start from global domination and work backwards. We have some people in the UK now who are following that model. They really have the bit between their teeth and make excellent teams.

CSC World: You've also written about the need for "smart funding" for start-ups. What is that?
There are two dangers: not having enough money and having too much. A slight lack of money makes people very innovative, but too much money makes them wasteful. Dumb money is money put into a project without any due diligence. Or investing in a project, then pulling out as soon as the project is a bit stretched.

Smart money brings with it a business network. It say, "We'll wrap a blanket of support around you, we'll find people for you to work with, help you find customers, help you do your pitch for contracts. And if you get in trouble, we'll come in and help."

CSC World: What about management structure? You've been outspoken on that topic.
Yes. Something I practiced in BT and practice now is driving down the number of management layers to the absolute minimum. The natural tendency seems to be, "I want to look important, so I'll hire more and more people underneath me." Ergo, people become disconnected and remote and everything tends to go pear- shaped.

If we were to set up a company that's going to create a revolutionary mobile phone network, we'd only want one or two layers of management. But if we were running an oil rig or a petrochemical plant or an automobile manufacturing plant, we'd have many layers of management and very formal systems. We don't want anyone innovation on the production line. But we do want innovation in our revolutionary company.

CSC World: One final question. What drives you?
There's a formula on my home page (cochrane.org.uk): Life = sft (sex, food, and technology). Those are the driving forces in my life. They're all intoxicating. We live and die by technology and we have a responsibility to live and reproduce. The single most important and most demanding job I've ever had is bringing my children into this world and trying to get them into shape.

I enjoy building and creating things. I envy young people the technological future that is before them.