The magic numbers for Ruffolo are 12, 8, and 2. That’sa three-link chain of the annual number of compounds entering development, the number of investigational new drugs entering clinical trials each year, and the annual number of new drug applications submitted to the US Food and Drug Administration. He says that his approach has helped turn around what he calls the company’s “pathetic” track record of submitting new drug applications in the years before he arrived. And best of all, it’s proven to be sustainable: Wyeth has met the targets every year since 2001, he says.“That’s the most important point. It’sa steady-state model.” Ruffolo admits that approach didn’t win him any popularity contests at Wyeth.“Scientists hate this approach,” he says.“When I was a scientist, we used to say that you can’t manage science. But it needs to be.” Those who didn’t buy into the approach left the company, he says—and those who have remained appreciate knowing where they stand. Richard Scheller takes a very different approach as executive vice president of research at Genentech, which has eschewed large acquisitions and does all research at its ever-expanding South San Francisco, California, campus. A neuroscientist and former Howard Hughes Medical Institute investigator at Stanford University, Scheller came to Genentech in 2001 after deciding that its culture meshed with his own philosophy of doing science. Genentech’s corporate strategy, labeled Horizon 2010, does include research goals for its more than 600 scientists over the next 5 years. But although they specify the number of new products to be moved forward for each of the company’s three major therapeutic areas, some goals omit key steps in the process. And they aren’t linked together in a formal manner.

Sitting in a top-floor office overlooking San Francisco Bay—and the pier that was allegedly the favorite fishing hole of cofounder Herbert Boyer—Scheller describes an ongoing study of Genentech’s attrition rate and the nature of its pipeline in a way that suggests he doesn’t view it as quite the priority that Ruffolo does.“It turns out that different types of projects fail for different reasons,” notes Scheller, who says that he “doesn’t know very much about big pharma” despite the fact that, based on the value of its stock, Genentech is the fifth-largest drug company in the world.“For example,” Scheller says,“I’m expecting small-molecule throughput rates to be lower than for protein therapeutics. I’m also leading a project to understand the bottlenecks. And I think that they will turn out to be what you’d expect: Some projects will be underresourced, some will suffer