Steve Ballmer's Big Moves

Even if you look at infrastructure companies, who can you really count on to be around in five years? VeriSign might be a safe bet. But look at guys like BroadVision, Vignette, or Interwoven. They have nice market capitalizations, but are they going to be around in five years?

All the same, you've said repeatedly that the biggest challenge to Microsoft isn't likely to come from an entrenched company; it's likely to come from a startup that's taking a dramatically different approach to some big issue. So I have to ask: If you weren't running Microsoft, and you were just starting a company, what would you set out to do?

Do you mean "What's a good business idea?" or "What would I do to compete with Microsoft?"

Either one.

The first thing I would ask myself is "Can I afford to build a company for the long term? Or is my strategy to build something that, 7 chances out of 10, I would want to sell to Cisco, Microsoft, or Intel because it would dovetail with their business?" If I couldn't afford to be patient, I would probably pick an area that's very enterprise-oriented, where I could get real customers quickly with a modest investment. There are plenty of good enterprise ideas rattling around. I wouldn't pick a dotcom area, because that's too dependent on brand. I'd probably choose some enterprise infrastructure area -- and then attach it to someone else's big sphere.

If I could be patient, I'd probably take a look at one of the real platform shifts that is coming along. That might mean looking at Microsoft's dot-net initiatives and asking, "Do I want to bet with Microsoft, or do I want to bet on the opposite side?" I might bet on mobile wireless technology, assuming that this will really be a major shift. It may be less of a shift than some people think, but it's an intriguing area.

So how do you make sure that you're getting fresh new perspectives, given the job that you really have?

I try to have at least two or three chats a month with venture capitalists. We invest with Accel Partners here in the United States, and we have other venture funds in England and in Israel. It's a way to keep in touch with interesting companies and to look for interesting opportunities for partnerships. I spoke at a venture-capital roundtable we put together in California not long ago. I spoke at a similar event in Paris around the beginning of November. When I talk to VCs, their top issues always seem to be "Where are you going? What are you avoiding? Where is it clear that you need partners?" That's very important stuff.

Also, I've always encouraged -- and sometimes ordered -- our engineers to get out and talk to customers. We can believe that we know where the world should go. But unless we're in touch with our customers, our model of the world can diverge from reality. There's no substitute for innovation, of course, but innovation is no substitute for being in touch, either. We have Project Bluestar, which we're doing with financial-services customers, primarily in New York. We've dispatched a number of our key development people to New York so that they can spend time with Merrill Lynch, Goldman Sachs, and other businesses that will fundamentally shape our product strategy.

What's your advice to the CEO who's running a blend of businesses? Even if an executive isn't in high-tech, she is likely to have some big, profitable warhorses that were founded long ago and some small but promising new areas that clamor for resources. How do you manage the two?

One of our biggest challenges is managing a portfolio of multiple businesses. There are many CEOs who can say that each one of their businesses has almost total independence, so all that they need to do is juggle how much investment they put in. But that's not our situation. We have common technology platforms across nearly all of our business, so we can't let people at each operating unit just do their own thing. It's not enough for us just to decide the investment streams. We need a common road map.

If you think about it, there's a lot you can learn from the automotive industry. You have to look at the people in this world who have big R&D budgets -- and that basically means pharmaceuticals, autos, chips, and network-equipment makers. At Microsoft, we're spending $4 billion a year in R&D. If you want to find people up in our range, they're basically almost all within those industries.

Now, what's striking is that in pharmaceuticals, the company that leads a therapeutic category in one generation is very seldom the leader in the next generation. So the guy who is best at treating heartburn this year generally isn't the guy who will do best in the next cycle. That's not the dynamic we want to be in. We want to keep building a leadership position. In other cases -- General Electric, for example -- it's more that they're in five or six independent businesses than that they're trying to support a common platform. (Although any company with GE's market cap and an ability to grow earnings 15% a year is worth studying.)