How to solve the cluster equation?
01 June 2008
In 1997, Microsoft opened its first big European lab – and it placed it in what was then Europe's only serious answer to Silicon Valley: Cambridge, in the United Kingdom.
Microsoft’s European R&D chief – on building clusters of innovation
The Cambridge cluster today remains special: it has more than 1,200 tech companies employing 35,000 people and has created more than 200 millionaires, according to the university at its centre. Microsoft's lab there has also grown – to more than 130 staff, and works closely with a further 200 researchers and several joint research centres in Microsoft's European collaboration network. In all, the company has invested more than £100 million in the lab, says its boss, Andrew Herbert, managing director of Microsoft Research Cambridge.
So if one Cambridge is a good thing, how do you make many? That's a policy question of growing importance in Europe. In Brussels, the European Commission is preparing to announce this year a new set of policies to promote clusters. Herewith, in an interview with Science|Business, are some of Herbert's ideas on what it takes to make more clusters.
Why do clusters matter?
Clusters tend to spring up around new technologies and therefore have early-stage companies - and that's an environment that's full of risk. Clusters help manage the risk. For instance, if the message to a prospective worker is, ‘come and join my start-up in some place you never heard of,' - you're going to think twice about it. If the message is, ‘come and join my start-up in Cambridge, and if it all goes pear-shaped there are 20 other companies hiring,' – that's a less-risky proposition. With clusters, there's a de-risking and a mutual support that allows companies to grow faster than they could do on their own.
What makes clusters work?
There needs to be something that acts as a hub, and it's often a university or a university department. There also have to be some entrepreneurial people who know about turning research ideas into business plans; just having a science lab doesn't work if it's not entrepreneurial. You've got to have a core of people who share the view that there is a business to be made out of a technology. Then once there's a culture of mutual support and usually some kind of organisation to promote networking, like The Cambridge Network here, by putting on events for local companies. When a cluster is working, people help each other in business. If someone knocks on your door and you don't sell what they want, you will point them to someone else in Cambridge rather than have them leave empty-handed. That's important.
What's the role of a big company?
We add gravitas to a cluster. It sends a strong message that interesting things are happening if a major company sets out its stall in town. Those technology start-ups are potential partners in our own R&D activity. The big companies help with all those concerns people have about employability.
How do you make more clusters in Europe?
One thing needed is early-stage money. In the UK start-ups are often funded by business angels, and they tend to be people who have been through the previous cycle in the cluster and want to re-invest and help others with the experience. They lend people amounts from tens to hundreds of thousands to get their idea out of the laboratory and help them through the first few years – something the VCs and banks don't do these days, because they want so much risk taken out.
Another key thing is what government can do by reducing the burden on young companies. Tax credits for R&D are important, and relaxing employment laws for young companies, so that it's easier to hire and fire and you can manage employment with your cash flow. That's for central government.
Then there's local government. We want to be somewhere that people have heard of and believe is a good place to work. It needs to be a place where people see other, equivalent job opportunities. It has to be somewhere that has a lifestyle that people like. What are the schools like? What's the shopping and entertainment like? Some governments try to create clusters in areas of industrial deprivation: a coal district with a high-tech cluster – you're going to struggle with that one.
Another point: what is the government attitude to foreign inward investment? For many start-ups, the goal is to be bought by a big company. How is the French government going to feel if it pumps €10 million into a cluster, and one of the companies ends up with a Microsoft or IBM badge on the door? In Cambridge, we'd declare that a victory. But in many countries, there are a lot of tensions around that question.
Finally, you need a university culture in which academics are encouraged to exploit their ideas. There is a tension about whether exploitation is done by the university or the researchers. Cambridge has a hybrid of the two: the university has a technology transfer office and by default owns the IP, but if you want to exploit it you can do so with the university as a sleeping partner. Generally, the ability to exploit IP in any way that really works – that's very important.