Francis Narin is no ordinary detective. So when a consulting firm called three years ago and asked him to investigate a division of a major chemical company, he didn't start following anyone or probing financial records. Instead, he fired up the vast data base at CHI Research Inc., the tiny research firm he runs in Haddon Heights, N.J., and began his line of inquiry: analyzing patent trends.
All Narin knew was that the chemical company's division had a reputation for innovation--and that the consultant's unnamed client was thinking of buying the unit. A search of CHI's data base of more than 1 million U.S. patent records turned up a surprising fact: Many of the division's key patents belonged to just two researchers. Then, a second survey showed telltale evidence that something was afoot: The pair had stopped filing patents under the divisional name and were inventing for the parent company instead. "In anticipation of divestiture," Narin surmises, "the company had moved their key people out." Narin calls this selling"the empty technological shell." Though he can't say for sure what role the CHI report played, the acquisition never took place.
The technique Narin drew on is bibliometrics, a form of statistical analysis used to scan patents and scientific papers to figure out which ones are most important. Increasingly, savvy companies are employing it to spot technological trends, unearth competitors' moves before they're apparent in the marketplace--and even help plan acquisitions, partnerships, and overall strategic direction. "You can spot right away the leaders and the followers in different technology areas, and you can detect companies shifting technological emphasis--or even strategies," attests Donald Milstein, manager of planning, coordination, and evaluation for Mobil Research & Development Corp., the oil company's technology arm.
Based on this technique, BUSINESS WEEK has teamed up with CHI Research to create a new Patent Scoreboard that tracks corporate technological power. It separates 197 of the world's top companies into 14 industry groups and ranks them by technological strength--a measure derived from the number and quality of patents (glossary, page 70).
As with any analytical tool, the new tables don't tell the whole story. They cover only U.S. patents--and while most important foreign inventions are also patented in the U.S., this isn't always true. Then, too, in order to keep research secret, some technologically strong companies restrict the number of patents they file. In addition, Japanese companies are being pressured by their backlogged government to curtail filings for minor patents--a trend that could mean lower Scoreboard numbers if they follow suit in the U.S.
Despite these caveats, the new Scoreboard shows key technological trends. The raw power of Japanese companies stands out, challenging the stereotype that they are mere copycats. Japan holds down the first four slots, led by Toshiba Corp.'s technology strength rating of 1,677. The best U.S. performer is fifth-place Eastman Kodak Co., at 1,186, with IBM a step behind. The top European: Philips Electronics, No. 11 overall.
COMEBACK KIDS? Within industry sectors, the story is more ambiguous. Given the prowess of Japanese auto makers, it's no surprise that only General Motors Corp. cracks the top five in tech strength. But Detroit may be catching up. Over the past five years, GM has gained an average of 10% a year, and Ford is up 16% annually. And while Nissan's rating has risen, Toyota's and Honda's have declined. Japan's carmakers downplay the numbers. True, they're patenting less, as requested. "But that doesn't mean technological development has slowed down," says Toyota Executive Vice-President Shiro Sasaki. Still, Detroit has recently gained market share, and some experts say its technological comeback may be a factor. "The changes that the domestics have made have been primarily in the technology," says Thomas F. O'Grady, president of Integrated Automotive Resources Inc., a Wilmington (Del.) consultant.
Among individual companies, General Electric Co. shows a seemingly ominous decline. Over the past two decades, GE has been awarded more U.S. patents than any other company. But since 1987, output has slipped an average of 8% a year--and its technological strength rating has fallen an average of 11% annually. GE attributes the decline to its downsizing from 30-odd businesses to just 13--and says it remains the world leader in most of those fields. On a brighter note, U.S. technology-strength leader Kodak has won 30% more patents annually since 1987, pushing into emerging fields such as electronic imaging. Its tech strength has jumped 44% a year. In 1991, Kodak trailed Canon Inc. by just 15 points.
The new Scoreboard will mark the first time many companies have heard of bibliometrics. But others are well aware of it. Kodak forages through patent data to track particular inventors--to see what competitors are doing and whether they are moving star scientists to hot new areas. Philips has used patent analysis to track 30 rivals in 39 fields, from software to optical recording.
The uses of bibliometric data go far beyond intelligence gathering, however. Mobil Research & Development first used the technique in 1987, as part of an assessment of its technological competitiveness. As expected, the analysis showed Mobil leading the pack in the discovery of zeolites, synthetic catalysts used in oil refining and petrochemical production. But researchers were stunned to find that competitors were often better at patenting applications for the very materials Mobil had discovered. "Obviously, we were missing out on a potential opportunity," says Milstein. Over the next two years, Mobil R&D upped by a third the budget and personnel devoted to exploiting new zeolites. Without giving specifics, Milstein says the moves will pay off with a slew of new applications.
EARLY WORD. These same kinds of studies--charting output from individuals, universities, and corporate labs--can be done with scientific papers instead of patent data. The pioneer in this area, the Institute for Scientific Information in Philadelphia, assesses the importance of papers by determining how often they're cited in other reports. Moreover, several groups in Europe and the U.S. use ISI's raw data for their own studies. Leonard Simon, president of the Center for Research Planning in Philadelphia, says it's possible to pick up hot new fields or shifts in the direction of research before they're common knowledge.
Whether it tracks patents, papers, or both, bibliometrics can reveal vital information. And even if it spots nothing experts in a given area don't intuitively know, the technique can provide hard data that corporate executives need in order to act on the judgments of their technical staff.
What's more, advocates say, now and then they see something fundamental. "We can show companies that their research universe is in the midst of a whole revolutionary process--and that down the line, all their products will be obsolete because the science will change," Simon adds. Even enthusiasts warn against basing decisions solely on such analysis. But this technique gives companies another way to guard against being blindsided by technological change.