December 2001

Lean Mean R&D Machines

Continued from page 1

By Wade Roush

Making Connections

In the pharmaceuticals industry, where R&D costs have been rising for years without any commensurate rise in the number of new drugs reaching the market, it's long been clear that old research models needed revision, says MacKenzie. Implementing a discovery process that produces more drugs over less time requires freeing researchers from the "pressures of productivity" that can keep them from experimenting with risky new technologies. At the same time, even a state-of-the-art research center needs to keep its work aligned with business needs. To that end, Pfizer's Discovery Technology Center hires only researchers who "have scientific degrees but are also outstanding data miners or statisticians on the side," says MacKenzie. Such people tend to be "totally immersed in the business of drug discovery, not off to one side of it, which is incredibly important to what we do."

To guarantee a supply of such researchers, Pfizer cultivates close ties with the local academic community. Last year, for example, the company created a three-year fellowship program in computational biology at Cambridge's Whitehead Institute for Biomedical Research. Fellows are expected to work inside the center for part of that time, examining gene sequences or protein structures relevant to drug discovery, but are also encouraged to do independent research. And for more great ideas, the center isn't opposed to turning to smaller companies and technology suppliers. Cambridge, MA-based BioTrove, whose nanoscale liquid-handling technology allows researchers to mix tiny quantities of reagents with 10,000 or more separate drug targets on a single chip, is conveniently headquartered right inside the Pfizer center's facility.

One of the basic tenets of the Pfizer center is that research is more effective-and more profitable-if it's more connected to the world outside the company. And Pfizer is hardly the only high-tech company testing this hypothesis. Chip-making giant Intel, for example, is spending part of its $4 billion R&D budget this year to support a series of "lablets" adjacent to top universities, each directed by a faculty member who has taken a leave of absence for a year or more. Each of the 20- to 30-person lablets will focus on a promising young technology (see "Intel Revamps R&D," TR October 2001). For example, computer scientist David Culler, the founding director of an Intel lablet at the University of California, Berkeley, is developing the software infrastructure for networks of tiny sensing and communication devices. If such devices eventually permeate our surroundings-gathering and wirelessly sharing information that could be used in surveillance, environmental control or scientific measurement-Intel wants to be the firm that builds them.

Key to this program, says Intel's director of research David Tennenhouse, is the fact that the academic researchers heading the lablets have a strong desire to see their ideas applied in the real world. Intel, however, is discouraging the researchers from taking their technologies all the way to the commercialization stage or becoming business unit managers, which might keep them from doing what they do best-innovate. "We're saying, Work on this strategic research project for a few years, and if it succeeds keep moving downstream [toward the market] for a few years, but then cycle back to the lab until you have another idea that you want to foist on the world,'" says Tennenhouse.

Foot in the Door

f Intel and Pfizer are dismantling the old walls between corporate and university research, IBM is blurring another traditional boundary-that between corporate research labs and the company's own customers. The idea behind Big Blue's new Emerging Business Group is to offer small startup firms access to IBM's extensive research in information technology. In return, IBM may get a small amount of cash or equity, but the main point is to encourage the smaller firms to build their own new technologies on top of IBM software and services. "Basically, we get IBM platforms into those companies so that when they succeed, we have a growth market for our products," explains Dave McQueeney, IBM's vice president of emerging business. "If we pick the startup companies that will grow up to dominate new spaces, it could be very smart for us."
As an example, McQueeney points to work on online auctions at IBM's Thomas J. Watson Research Center in Yorktown Heights, NY. A team there is developing "combinatorial optimization" software that compares bids in Internet-based auctions where the bidders seek to buy distinct but overlapping sets of items. One bidder-say, an electronics manufacturer-may need components