Freedom to Tinker

I’ll be in a place with a possibly iffy Internet link until Monday evening. If you don’t hear from me in the next few days, I’m probably incommunicado; but please tune back in on Tuesday.

Yesterday’s New York Times ran a story by Randall Stross, marveling at the number of Ph.D.’s working at Google. Indeed, the story marveled about Google wanting to hire Ph.D.’s at all. Many other companies shun Ph.D.’s.

Deciding whether to hire bachelors-level employees or Ph.D.’s really boils down to whether you want employees who are good at doing homework on short deadlines or employees who are good at figuring things out in an unstructured environment. (Like all generalizations, this is true only on average. There are plenty of outliers.) Google is a bit unusual in opting for the latter.

What the article doesn’t say is that Google does not hire just anybody with a Ph.D. diploma. They’re pretty careful about which Ph.D.’s they hire. Google can afford to be choosy since so many people seem to want to work there. Google benefits from a virtuous cycle that sometimes develops at a company, where the company has an unusual concentration of really smart employees, so lots of people want to work there, so the company can be very picky about whom it hires, thus allowing itself to hire more very smart people.

The article also hints at Google’s success in integrating research with production. The usual model in the industry is to hire a small number of eggheads and send them off to some distant building to Think Deep Thoughts, so as not to disturb the mass of employees who make products. By contrast, Google generally uses the very same people to do research and create products. They do this by letting every employee spend 20% of their time doing anything they like that might be useful to the company. Doing this ensures that the research is done by people who understand the problems that come up in the company’s day-to-day business.

Sustaining this model requires at least three things. First, you have to have employees who will use the unstructured research time productively; this is where the Ph.D.’s and other very smart people come in. Second, you need to maintain a work environment that is attractive to these people, because they’ll have no trouble finding work elsewhere if they want to leave. Third, management has to have the discipline to avoid constantly canceling the 20% research time in order to meet the deadline du jour.

Google does all of this well. They probably benefit also from the nature of their product, which generates revenue every time it is used (rather than only when customers decide to pay for an upgrade), and which can be improved incrementally. Revenue doesn’t depend on cranking out each year a version that can be sold as all-new, so the company can focus simply on making its products work well.

Can Google maintain all of this after it has gone public? My guess is that it can, as long as it is viewed as the technology leader in a lucrative area. If Google ever loses its aura, though, watch out – when the green eyeshades come out, many of those smart employees will leave for greener pastures, probably for a company that bills itself as “the new Google.”