Money, money everywhere for R&D, it seems—hundreds of billions, in fact, and more of it every year—but barely a trickle goes to services research. Yet, the services sector now accounts for between 60 percent and 80 percent of developed countries’ gross domestic product and worker wages. Does innovation improve only the proverbial widget, or can services firms leverage R&D to spur innovations and increase productivity?

IBM Corp. is launching a bold new initiative to find out. The company, which came in at No. 10 on our annual list of top R&D spenders, is the only one of the top 100 spenders that derives a significant percentage—slightly more than half—of its revenue from services. Although the U.S. Department of Labor’s Bureau of Labor Statistics estimates that 83 percent of all U.S. jobs are in services, many of the big tech-related services companies, such as Computer Sciences Corp., in El Segundo, Calif., and Electronic Data Systems Corp., in Plano, Texas, don’t even bother breaking out their R&D expenditures. Only Accenture, in Hamilton, Bermuda, touts its R&D expertise as a competitive advantage, and its US $243 million R&D budget for 2005 didn’t come close to cracking the top 100 R&D spenders, much less the top 10.

Productivity in services has generally lagged behind that in the manufacturing and agriculture sectors. One reason is that many services are too intangible to measure and thus to compare over time—think of psychotherapy, custom software development, or legal representation. Another is that separating the work done by the provider and the consumer is difficult. Then there is the requirement of consuming a service on the spot, rather than storing it for later use.

Above all, services are labor-intensive and therefore hard to automate, Even in information-technology services, where the firms and their employees are steeped in technology, labor still accounts for the lion’s share of costs. Of course, automation has boosted productivity in a few service industries, such as air travel. Nowadays, many travelers search for the best itinerary, purchase tickets, check in their luggage, and don’t have to deal with a person until they get to the security checkpoint. But there are limits to improving other services. How much more efficient can you make a haircut?

Still, the consensus is that plenty of services, particularly those that involve business consulting and IT, are ripe for productivity improvements. But is formal research the best way to achieve them? Henry Chesbrough, director of the Center for Open Innovation at the University of California at Berkeley’s Haas School of Business, says that for technology to make a difference, the business process must first be analyzed and redesigned. Services firms that changed the business to exploit the technology were able to gain a competitive advantage over those that focused only on implementing the technology. Chesbrough insists that formal research in services is a necessary complement to technology research.

The biggest company to back that theory is IBM. The services’ portion of its revenues swelled from $12.7 billion in 1995 to $47.5 billion in 2005. One of the milestones was Big Blue’s 2002 acquisition of a PricewaterhouseCoopers subsidiary PricewaterhouseCoopers Consulting, a $5 billion-a-year consulting business. Suddenly, services were a significant portion of IBM’s offerings and revenue. For Paul Horn, IBM’s director of research, it was an opportunity to push services to the forefront of the company’s research strategy. In just three years, services’ share in IBM’s total R&D spending rose from practically zero to nearly 25 percent, and Horn expects that share to keep rising.

Paul Maglio, a senior manager at IBM’s Almaden Research Center in San Jose, says his services group has grown from nine people in December 2002 to more than 70 now. He estimates that 550 of IBM’s 3000-plus researchers are working on services, either directly with clients or as part of teams running company projects. Maglio says the skill mix of IBM research will continue to shift as the labs hire ever more anthropologists, sociologists, and economists.

IBM isn’t alone. Last year Intel Corp., in Santa Clara, Calif., announced that it, too, was hiring anthropologists and social scientists to help in its product development. The change was so foreign to the company that it had to create non-engineering titles for the researchers.

Horn emphasizes that the shift does not mean IBM will gut its traditional strength in engineering and the hard sciences. Those disciplines will continue to be represented, he says, even as many of its practitioners get redeployed into services.

The company is also taking steps to actively shape its future workforce. It has been aggressively promoting a new academic discipline it calls services science, management, and engineering (SSME). There’s a precedent: IBM’s Thomas J. Watson Sr. encouraged Columbia University in New York City to offer the first computer science course in 1946. Today, IBM is encouraging a group of universities—including North Carolina State University, Pennsylvania State University, Rensselaer Polytechnic Institute, and the University of California at Berkeley—to create SSME programs.

Pumping out graduates schooled in services science could form part of the United States’ or Europe’s competitive advantage in the coming years, Horn says. With the migration of manufacturing and R&D to countries with lower labor costs, aspiring engineers studying in developed countries may be able to carve out a lucrative career in services science by applying their skills to a different set of problems and working on more multidisciplinary teams.

The U.S. Congress is listening. The National Competitiveness Investment Act, introduced in September by a whopping 38 senators from both political parties, includes a section supporting government initiatives in services science. The legislation is a direct result of high-profile reports issued by business and academic leaders who have expressed concern about growing competition in high technology from India and China. The bill, which includes such wide-ranging policy initiatives as increases in physical-sciences and engineering research and a buildup in the number of K–12 science teachers, also calls for the federal government to develop strategies to leverage services science to improve U.S. competitiveness.

All of this begs a simple question: What is services science? The legislation defines it vaguely as “curricula, training, and research programs that are designed to teach individuals to apply scientific, engineering, and management disciplines that integrate elements of computer science, operations research, industrial engineering, business strategy, management sciences, and social and legal sciences, in order to encourage innovation in how organizations create value for customers and shareholders that could not be achieved through such disciplines working in isolation.” That’s not a definition, just a laundry list.

For a better idea of what the field is about, Maglio points to one of his group’s projects at IBM Almaden. Members of the group went to data centers and interviewed systems administrators in their natural, air-conditioned habitats. Maglio’s team learned that the administrators found their technologies inflexible, and to compensate for the problem, they informally created and shared custom tools. The research has already resulted in a new software tool for system administrators, and Maglio expects more products to follow.

But services research is not only social science–oriented. Horn says that IBM’s research in business consulting “tends to be much more mathematical,” tailored to specific industries and involving supply-chain optimization, logistics optimization, data management, and data analytics. All of those, he says, are “aimed at helping our clients improve their business processes and their approaches to markets.”

In determining success or failure, IBM is holding its services research projects to the same standards as its other R&D initiatives. The research division justifies its budget by demonstrating return on investment through hard examples of how it reduces the cost of services delivery. Many of those improvements are embodied in new tools and technologies. Ultimately, Horn says, customers must be convinced that they are getting better value—through some combination of reduced costs and improved quality—from IBM than from the competition. So far, the numbers suggest that Big Blue’s top management is satisfied with the performance of its services R&D projects and looking for more of the same: IBM’s R&D investments are growing much faster in services than in software and hardware.

Ralph Gomory, president of the Alfred P. Sloan Foundation in New York City, sees some parallels to the emergence of services research and the transition IBM faced in 1970, when he became its head of research. The big difference, he says, is that IBM then had large and important internal customers—for example, the division that developed and manufactured semiconductors and packaging. So if the manufacturing group had materials problems in developing multilayer ceramic packaging, the group could work together with the research division’s physicists to solve the problem. Relationships could be built that lasted many years and were even solidified into long-lasting interdivisional organizations called “joint projects.” Today, the final customer for services research tends to be one that is outside IBM, making it more difficult to have close and lasting relationships. But Gomory says he thinks it can be done.

Now, however, the internal customer in the services business is closely tied to external customers, making it trickier to demonstrate any tangible contributions from IBM’s research division. Perhaps the solution is so specific to the customer that it cannot be reused with other customers. “The [services] exchange is generated by both parties,” Berkeley’s Henry Chesbrough says, adding that the adopter or customer is also a co-producer, “intimately defining, shaping, and integrating the service into his or her organization.”

Such close collaboration with customers is also changing the way IBM researchers work. Doing research in services isn’t anything like, say, spending some quality time with a molecular-beam epitaxy machine. Field research at customer sites, such as that conducted by Maglio’s group, is expected to become more common. The way we think about the physical infrastructure of a research lab may need to be expanded, Maglio and Chesbrough say, as researchers spend less time with costly machines and more time at customer sites.

Whether IBM can successfully refocus its R&D enterprise and lead the services sector in breaking new ground in the formal research of the post-industrial economy—and whether its services competitors can or will follow suit—remains to be seen. But considering IBM research’s storied history in hardware (hard disk drives, DRAM, and reduced instruction-set computer architecture) and software (relational databases and Fortran), this is certainly one strategy worth tracking.