What Money Can't Buy

And what of its megastar researchers? They aren't toiling on the latest Longhorn bell or whistle -- but neither are they pursuing anything with apparent immediate commercial potential. Jim Gray is developing the "World Wide Telescope," an Internet database that, Microsoft says, will eventually host "the entire world's astronomy data." Gary Starkweather is refining his design for a circular, "super panovision" computer monitor. Gordon Bell has spent his time recently on MyLifeBits, an online database aggregating all of his personal information, including 20,000 documents, 60,000 emails, 15,000 photos, and all of his music and videos.

Gates says he's content with that approach. "The market will tell us if we're ahead of our time," he says. "That's okay." But is it really? Experts in innovation argue that Microsoft just isn't investing enough in the sort of offensive innovation that should define its future. And the research it does do seems wildly inefficient. Over the past five years, Microsoft has spent an average of $9 million per patent, nearly twice the average for its software peer group. (In July, the company told analysts it would aim to double the number of patents it seeks. But in light of Microsoft's campaign to reconcile intellectual property disputes in the European Union and with competitors such as Sun Microsystems, this appears to reflect legal strategy as much as a renewed zeal to improve the research effort itself.)

Of course, it's impossible to say exactly what the right balance is between offense and defense, or between the short and long term. Ultimately, observes Gary Hamel, innovation guru and chairman of Strategos, "really good ideas are just few and far between. Forgive the metaphor, but it is a little like the process of sperm trying to fertilize an egg. Increasing the number of really strong swimmers doesn't increase your success rate. That's not how biology works. Only one [sperm] gets to fertilize the egg."

In other words, innovation is often just a happy, creative accident, where the laws of numbers can actually get in the way. In this sense, Microsoft's size and wealth become obstacles rather than assets. Venture capitalists typically pore over 50 to 100 deals to find a good $20 million software investment. By that logic, Microsoft, with its $6.8 billion annual R&D budget, must consider as many as 35,000 new ideas just to find a few hundred worth investing in every year. Is it any wonder the company invests in so many dogs -- or, for that matter, toilets?

No surprise that the parking lots of the Courtyard, Fairfield, and Residence Inns opposite Microsoft's headquarters are always full. The rooms there are choked with engineers who have been summoned by Microsoft on the basis of particular knowledge or expertise that may someday turn into a commercially viable innovation. Some will sign on as consultants and a few others as employees. But the motels are always booked: Microsoft is that starved for ideas.

Even then, there just might not be enough legitimate opportunities out there for Microsoft to employ its vast resources effectively. "There are only about 10 software companies in the world with more than $1 billion in sales," observes Brian Skiba, managing director of the San Francisco-based hedge fund Viant Group. In the software industry, a research project can be considered successful if it spurs annual revenue growth of 10% to 15%. By this metric, "Microsoft has to create the third- or fourth-largest software company in the world every year to be considered innovative," says Skiba. Even if Microsoft had been able to replicate Google's dominance in search technology, Google's $1.5 billion in revenue would have lifted Microsoft's own top line by just 4%.

That's why it serves Microsoft to avoid the connection between research and revenue growth. "I think you're too focused on the top line as opposed to the bottom line," CTO Mundie said, when we put this calculus to him. Mundie argues that successful R&D is judged ultimately by profit growth. How much? Mundie was vague. "It's the maximum that we can achieve." Later, he added, "We don't measure these things."

This is the luxury of entrenchment. So long as Microsoft enjoys its dual annuities from Windows and Office, it needn't worry much about the revenue (or lack thereof) generated by its long-term research investments. It simply plows money into ideas good and bad, then sits back and waits -- even decades -- until the market is prepared to validate these investments. "Because we've got Office and Windows," Gates says, "we can afford to take some risk in these new areas and make sure we're out in front. The real sin for us is if we miss something."

But the reality is, Microsoft has already missed plenty. And that's the downside of entrenchment: Even as the company throws resources at the future, its strategic gaze rests squarely on the past. As Microsoft forfeits future revenue growth for current income, it continues to cede genuine innovations and important new markets to future upstarts with bigger ideas -- and far less to lose.

Carleen Hawn is a writer based in Los Angeles.