Google vs Xerox on R&D

September 25, 2008 4:10 PM EDT
Google does R&D but has no dedicated researchers or formal lab organization. Every engineer is considered part of the virtual R&D team and is expected to donate 20% of his time – about one day a week – to research.

Xerox takes the traditional approach, investing in a separate R&D function that includes basic research. I spoke with representatives of both companies at MIT’s EmTech08 confererence this week and asked them to explain what makes each approach successful.

Here's what they had to say.

The Google View

Matt Glotzbach, enterprise product management director at Google, thinks having all engineers included in a virtual R&D organization works well. “Twenty percent of the time as an engineer you’re expected to spend on projects outside of your domain. My job is to look across the R&D and look at interesting projects we can bring to bear,” he says.

One advantage to this approach is that researchers are skilled at bringing new technologies to market, they are application focused, and they may get new products to market more quickly. The downside is that engineers have very real deadlines and continuous interruptions from the day-to-day business that can keep them from thinking the big thoughts or focusing on basic research.

“When you’re pushing a product out the door, am I really supposed to believe that the engineers go away one day a week?” Perhaps not, but they still get their 20% time commitment in, Glotzbach claims. “The reality is an engineer may be heads down on a [product] initiative and then might spend a month or two working on some R&D initiative,” he says, adding that new innovations are constantly coming out. He points to Google's new browser, Chrome, as an example.

He says existing Web browsers were too slow in the areas of rendering and Javascript execution for the Web applications Google is developing. Chrome, he says, was "our attempt to push the browser market a huge giant leap forward." The open source Chrome V8 code, he claims, has Javacript capabilities that are "two orders of magnitude better than those of existing browsers."

The Xerox PARC tradition

Sophie Vandebroek, Xerox CTO, thinks not having a dedicated R&D research function makes it hard to conduct the kinds of basic research that she sees as key to major breakthroughs. “It’s critical to have centralized R&D, to have a separate organization where the number one priority of the people is to invent and create high impact new technologies,” she says. By contrast, she says, the first priority of the engineering division is always to "get the product out.”

At Xerox, 25% of the Lab staff at any given time is involved in what she calls “exploratory” research. Another 25% are engaged in the “incubation phase” of a project. Of those, about 40% move into product development, and the R&D engineers move forward with it, collaborating with the engineers on the product development team through the first two phases. “The same researchers … are basically embedded in the first two phases,” she says.