How Google Fuels Its Idea Factory

CEO Eric Schmidt describes the simple principles driving the company's steady stream of innovations

Leading up to Google's first-quarter earnings report on Apr. 17, investors couldn't have been more bearish. They had knocked the stock down 35% since the start of the year, concerned that a weak economy would finally hurt the search giant's advertising business. But Google (GOOG) defied skeptics—and the economic downturn—with a surprisingly strong showing that sent the stock soaring 20% the next day. More than anything, Google's continued prosperity is a testament to its ability to keep innovating, both in search and advertising operations and in new lines such as online office-productivity software.

Many companies, says Chief Executive Eric Schmidt, can skirt downturns entirely by coming up with innovations that change the game in their industries—or create new ones. (When asked if Google's strategy would change as the economy heads into a likely recession, he replied: "What recession?") In a recent interview in a tiny meeting room next to his Mountain View (Calif.) office, Schmidt told BusinessWeek Silicon Valley Bureau Chief Robert D. Hof how Google manages the tricky process of innovation.

Do companies have to manage innovation differently in a downturn?

Innovation has nothing to do with downturns. A hot product will sell just as well in a recession as it will in a nonrecession. Let's imagine that we invented a better advertising product for television. What would our revenue growth be for that? Well, you're into a $50 billion market, so it will be driven not by whether there's a television ad recession but by what degree we can get people to substitute [our product] for the other. The strong companies understand this, and during a recession, they invest.

Can other companies emulate Google's famous model of letting engineers spend about 20% of their time on projects outside their main job?

The story of innovation has not changed. It has always been a small team of people who have a new idea, typically not understood by people around them and their executives. [This is] a systematic way of making sure a middle manager does not eliminate that innovation. If you're the employee and I'm the manager, and I sit down and say, "Our product's late, and you screwed up, and you gotta work on this really hard," you can legally say to me, "I will give you everything I've got, 80% of [my time]."

It means the managers can't screw around with the employees beyond some limit. I believe that this innovation escape-valve model is applicable to essentially every business that has technology as a component.

Why aren't many other companies doing this, too?

I think it's cultural. You have to have the culture, and you have to get it right.

What obstacles does Google face in continuing to innovate?

A problem that we face now is that we have people in multiple sites. It's a problem that everybody faces, but we're going to face it bad. We have, like, 50 locations.

So you still need that face-to-face contact?

The best programming team is a "telephone call," which is two people, you and I, programming together. The second-best programming team is, everybody fits into a single room. All other variants are bad.

Google has a reputation for doing a lot of R&D in-house, as well as buying companies to bring in other technologies. Are you shifting that toward getting outside talent to contribute in a bigger way than they have?

I don't know yet. The biggest acquisition we did was DoubleClick. We'll see how successful that is. We really like the small-team company acquisition model. My guess is, with the alleged recession happening, there will be a lot of good targets.

So we're likely to see even more acquisitions by Google?

I would think so. But small. The likelihood of us doing big things is pretty low because we'd have to assimilate the culture. Nobody works the way we do. The Google culture makes sense if you're in it, and no sense if you're not in it.

Does this mean it's tough for people to adjust, and will that limit the number of companies you can comfortably buy?

First, it is a fact that they adjust. Second, there is selection bias. Usually, people want to [be acquired by Google]. It's always very friendly. Because they have choices, and they choose us.

How does Google make sure it's producing innovations that change the game enough to create big new markets but also continue to appeal to its main customers, who might not want so much disruption?

We make an explicit decision to favor the end-user. [We] do not say, "Newspapers should be happy. Advertisers should be happy. Telcos should be happy. Competitors should be happy." Those are fine if we can do it. But it's all about end-users.

How do you make sure all these Google engineering projects actually turn into useful services?

The No. 1 thing we do require is: You can do whatever you want as long as you track it. We have very sophisticated measurement systems at every stage of launch. We have what is called trusted testers. Then beta test, which is forever. We do these 1% launches where we float something out and measure that. We can dice and slice in any way you can possibly fathom.

What's more important than the absolute number is the relative growth rate. High growth solves virtually all problems. If the growth rate is low, or negative, you've got a serious problem.

How much of a concern is it that Google has seen both executives and engineers depart for other companies or startups?

Let's do some math. We have 18,000 people. What is 1% turnover [per month]? 180. Do you think 1% turnover is reasonable? In this area, it's quite low. Ours is some small percent, 1, 2, 3%.

What bothers me is that some people write: "So-and-so left the company." Well, they don't also write that we hired 120 people that week, five of whom have Nobel prizes, three of whom have PhDs, and so on, who are beginning their career here now.

Can innovation really be managed, or is it a case where you have to keep the company and its managers out of the way?

I disagree with the word "managed." You have to have a set of necessary conditions for innovation to occur. To start with, you have to listen to people.

Pretty basic, no?

But not often practiced. Innovation comes from places that you don't expect.