A mind is a terrible thing to waste.

A few years ago, when NASA went to Congress for funding for its next-generation, heavy-lift launch vehicle, some wag on the committee asked NASA's administrator why the agency just didn't rebuild the Saturn 5 rocket. The Saturn, the largest and most powerful American rocket ever built, was the chariot that carried the Apollo space capsule and astronauts to the moon.

NASA, which invested about $50 billion of 1960s money in that rocket, said it would examine whether the Saturn could be rebuilt. A few months later, the space agency reported it would cost about the same to rebuild the Saturn as to build a new vehicle from scratch. The reason, the space agency said, was that no full set of plans for the Saturn could be located. Nor was there a complete set of tools or dies. And most of the engineers who built the rocket were now drawing Social Security or had died.

Since about half the rocket's cost went to R&D, and since some of the ideas, insights, and patents were used in other products, some of the rocket's intellectual content was preserved. But perhaps as much as half of that intellectual content was simply lost. My estimate is that NASA knows about $15 billion less today about how to build a Saturn than it did in 1970. That crude calculation is based on the amount of new R&D that would be needed to recreate the Saturn today. That not-insignificant sum gives new meaning to the bumper sticker, "If you think education is expensive, try ignorance."

Government agencies are not the only organizations that have a tough time valuing and keeping track of intellectual capital. Few companies are adept at managing what they know. In fact, a recent study of 80 companies conducted by the American Productivity and Quality Center found that 59 percent of respondents felt they were not doing enough to manage their intellectual assets. And most companies confuse data - which rushes dirty and unanalyzed from cash registers, engine sensors, and traffic counters - with information - which has been captured and analyzed.

But real knowledge is neither data nor information. It comprises insights that are, for the most part, deep and strategic. When Boeing's knowledge managers append Federal Aviation Administration advisories to the parts listed in the company's data warehouse of design specifications, the innocent coupling of those two types of information yields insights that save people's lives.

Not surprisingly, the companies that manage knowledge best are those whose livelihoods depend upon it. Several consulting firms, McLean, VA-based Booz Allen & Hamilton among them, now have chief knowledge officers who organize and structure the insights of the firm's best minds. Catalog companies, such as Minnetonka, MN-based Fingerhut, keep exhaustive records about their customers' buying habits. Skandia AFS, an insurance company in Stockholm, Sweden, uses a variant of Tobin's Q ratio to measure the value of ideas. The Q ratio is a Nobel Prize-winning set of insights that examines the replacement value of corporate assets and compares them with-the-value of corporate stock. By performing this analysis, and including the replacement value of insights, ideas, patents, and production processes, NASA could have seen that the amount it would have had to invest in memory - a Xerox copy of the Saturn's plans, one complete set of tooling and dies stored in a warehouse - would have accounted for a fraction of the rocket's replacement cost.

Yet it is not enough to store knowledge. Insights must be managed. "Trapped" intellectual capital leaves a company each time a worker is dismissed, laid off, or retires. The knowledge warehouse is confined to the worker's gray matter.

"For that reason," says John Clippenger, co-founder of Context Media, a company in Boston that writes knowledge-management software, "companies must create maps of their intellectual capital. These maps resemble thermal maps that follow weather patterns. They also track communities of practice - such as marketing or manufacturing."

In Clippenger's view, the software he creates must support the discussion process in each community. But it must do more. It must add to the discussion by finding new intellectual resources and by automatically scouring databases and the Internet, using intelligent agents. And it must capture the dialogue and the insights taking place in each community in ways that can be recreated for future reflection and use.

Had NASA had this type of system in place, the nation's rocket development program would never have had the equivalent of a 17-minute gap.

Joel Kurtzman, former editor of The Harvard Business Review, is an international business consultant and author. He is the director of the International Trade Program at The Manhattan Institute.