Commentary: Research Labs Get Real. It's About Time

While you stand there jabbing at buttons, opening side panels, and wondering why you can't get a simple 8 1/2-by-11 copy out of your Xerox copier, the deep thinkers at Xerox Corp.'s fabled Palo Alto Research Center are trying to improve your life. Not right away, though. Later this century, thanks to research on exotic stuff such as a robot named PolyBot or "smart matter" embedded with microscopic machines. They're even working on natural language processing--presumably, so that when you curse the copier, it will understand you.

They'd better work quickly, though, because PARC is up for grabs. As part of a companywide effort to raise money and cut costs, Xerox said on Oct. 24 that it would seek partners for the lab. PARC is famous for landmark inventions such as Ethernet computer networking and the point-and-click interface that did little or nothing for Xerox shareholders. Though PARC consumes less than 10% of Xerox' R&D budget, its enormous ambitions send the wrong message for a company that needs to refocus on making money. Now, co-owners will likely demand PARC researchers work on more practical, short-term business problems. And that's a good thing.

FERTILE. Xerox' woes might seem to provide evidence that there's something wrong with the U.S. system of corporate R&D. That's not the case. Innovation is rife in the U.S., from industrial labs to contract research outfits and Silicon Valley startups. Says Martin N. Baily, chairman of the President's Council of Economic Advisers: "We've been able to generate a fertile breeding ground because we have a range of funding mechanisms."

Start with the corporate labs, which once resembled ivory towers. Today, their PhDs are getting their hands dirty on real-world customer problems. But "Xerox has been relatively slow" to latch on to that trend, says Massachusetts Institute of Technology economist Rebecca M. Henderson, an expert on corporate R&D.

Companies that get more pragmatic about research aren't dropping the "R" to do more "D." Many tough business problems can't be solved without resorting to basic science--but these days it's a lot more focused. The National Science Foundation says growth of industry investment in research outpaced investment in development from 1994 to 1998, 64% vs. 46%. One product born of basic research is IBM's Microdrive, which is the size of a matchbook. It harnesses a phenomenon, giant magnetoresistance, that was confined to academic labs until IBM physicists caught wind of it. Says Paul Horn, IBM's senior vice-president for research: "We want to capture new ideas wherever they bubble up."

And they often bubble up outside of corporate labs. The fastest growing source of industrial R&D is startups funded by venture capitalists. The amount of VC funding soared from $5 billion in 1988 to an annual rate of $100 billion so far this year. Josh Lerner of Harvard Business School has found that VC-backed companies produce lots of patents that are frequently cited, showing that startups aren't freeloading off research done by others.

Smart corporations also are getting ideas from the outside. Cisco Systems Inc. has perfected R&D by acquisition, snapping up 20 companies this year alone. But even Cisco realizes that buying tech companies isn't sufficient. It spends 14% of revenue on its own R&D. Venture capitalists agree that acquisitions aren't the full answer: "If I were a CEO, I would want to own my corporate competency," says Mark W. Perry, general partner with New Enterprise Associates, a Menlo Park (Calif.) VC firm.

With corporate labs narrowing their vision, who will fund research that produces the Ethernets of the future? Government, for one, should boost support for basic research. For-profit companies may also pick up some of the slack. For instance, Judy Estrin, Cisco's former chief technology officer, has co-founded Packet Design Inc. in Menlo Park, Calif., to solve Internet problems.

There are many good models for industrial R&D. The one model that doesn't seem to work is PARC's: corporate lab as dream factory.

Before it's here, it's on the Bloomberg Terminal.