'The initial development was done in a garage'

Roger Foster, 55, has just started his second business, Financial Objects. Here he explains how his first venture, ACT, which is now partly owned by Mitsubishi, survived in the rapidly changing world of computers.

There was virtually no high-tech sector in the West Midlands when I started in business. In the early 1960s the area was dominated by engineering and the motor industry. Information technology makes up a respectable slice of the economy there now. And the company I built, ACT - although part of it is now owned by the Japanese giant Mitsubishi and the rest is in the hands of Misys software - is still at the centre of it.

I left grammar school at 16 and qualified as a chartered accountant by the time I was 21. During my first job - two years at the engineering group GKN - I moved from the financial side to join the computer side. It was good timing - the third generation mainframes, like the IBM 360, were just starting to come in and the prospects for information technology looked brighter than ever before.

I set up ACT, which stands for Applied Computer Techniques, with two others in 1965. We did two things that were innovative. First we packaged programs that could be sold to any customer, instead of designing bespoke systems for each client. And second we set up a service bureau, running our accounting packages for companies that couldn't afford their own computers. We didn't have one either, but we rented time on a mainframe at English Sewing Cotton in Manchester. It meant that after working all day in Birmingham we had to drive two hours to Manchester, run all the data through, and come back again. As with so many of these stories, the main ingredient for success is hard work.

It was a slow start, and we lost money in the first year and had to raise money from our families and with a bank overdraft secured on my house. But after two years we were showing a profit. One of the nice things about the business was that it involved long-term relationships with customers. Once you had built it up it was a good solid business. So solid that we received a bid for the company in 1971 or 72. My partners were keen to sell so we put together a deal to buy them out.

The 1970s were really the era of the service bureaux. In 1979 I floated the company at a market capitalisation of pounds 3m. We were the first software company in Britain to float, and the shares doubled on the first day. But the computer industry changes quickly, and I could see that the next phase was in personal computers.

The first move away from the service bureaux came when Chuck Peddle, chief designer at Commodore, left to set up his own company, Victor. He approached us about writing software for the machine, which was called Sirius, and seeing the opportunity, we offered to distribute it in the UK. We recruited 800 dealers in the first year, many of which are still in business.

Sirius was a great machine, and accounted for 80 per cent of our sales by 1983, but Victor was a total disaster. It's now a Harvard Business School case study of boom and bust. It went from zero to sales of $250m to being bankrupt in four years. We were worried that we'd lose our main product, and decided to design and build our own - the Apricot. It scares you to death to think about it now but it didn't at the time - we were too busy making it happen.

The initial development was done in the garage of a house in Dudley, West Midlands, but within 12 months we had a full-scale research and development department and had built a factory in Scotland from scratch. The Apricot project went very smoothly; our biggest problem was keeping Victor from pulling the plug on us before we were ready. But in the end, the transition from selling Sirius to Apricot was seamless.

The best Apricot computer we built was the first one. It was regarded as a major advance and it's still in the design section of the Victoria & Albert Museum. Bill Gates was the first person to see the prototype and he adored that machine. We used Microsoft software in the days when it was a smaller company than we were. Our key contact was Bill Gates himself. Gates was a bit of a propeller head in those days, but as well as being technically brilliant he was commercially astute, skillful at marketing and a very tough negotiator.

Between 1983 and 1985 we sold more Apricots in Britain than IBM sold PCs. But we were still a relatively small company compared with the giants in America and the Far East, and were coming under increasing pressure. In the early 1980s PCs were individual machines, but by the late eighties it was becoming a commodity business, with prices falling and margins being squeezed. So in 1989 we sold out to Mitsubishi for pounds 39m.

By then we were already selling software into banks. We already had a healthy share of the market in the City with our Quasar package, which we had developed in 1980, and we used the money from selling the Apricot business to go on a buying spree. We bought BIS from Nynex, the US phone company, for pounds 95m and Kindle in Ireland for pounds 35m, as well as one or two others which gave us market leadership. Last year, three decades to the day after ACT was founded, we sold the company for pounds 212m.