By: Peter Coffee

December 13, 2004, eWeek

Information technology has a well-hyped reputation for exponential rates of improvement, but productivity growth in delivering IT-enabled services has fallen short of what's been achieved in less glamorous sectors of the economy. As services come to dominate global markets, it's critical for U.S. systems scientists and professionals to find ways to apply capital and technology as effectively in the workshops of Web services as in the factories and fields.

Jim Spohrer, director of services research at IBM's Almaden Research Center, last month convened a gathering of 270 industry and academic practitioners in search of a science of service improvement. "Well over 70 percent of our labor force is in services," Spohrer told the assembled group at the Almaden campus. "If we're going to grow the GDP per capita, services productivity is what it's all about."

For purposes of comparison, U.S. farm productivity grew at 2.3 percent per year in the post-World War II period, as estimated by Wellesley College professor Robert Paarlberg. This is characteristic of what happens when capital and technology are effectively applied to labor.

Productivity growth in crafting service-enabling software has not kept pace. Randall Jensen, president of Software Engineering Inc., in Brigham City, Utah, found linear rather than compound growth rates in measured software development efforts between 1960 and 1990; his data indicates initial improvement at 1.8 percent per year, tapering off to 1.1 percent per year by the end of that period.

This ignores, moreover, the diversion of coding effort since the 1990s to increased security, to more elaborate user interface technology and to more diverse connectivity schemes. In addition, coding efforts have increasingly been devoted to enabling core business functions in real time and under intense regulatory scrutiny.

No business-regardless of industry sector-can sit out the services revolution, said Michael Radnor, professor of management and organizations at Northwestern University's Kellogg School of Management, in Evanston, Ill. Speaking as conference co-convener, on behalf of the Kellogg School's Center for Technology Innovation Management and the Management of Accelerated Technology Innovation consortium, Radnor warned that no amount of engineering or manufacturing prowess can elevate a company above the services fray.

"It doesn't matter what you make," Radnor said. "How long does it take to develop, perhaps, a wonderful high-tech controller? And how long does it take for that design to be commoditized, once it comes to market, by some other country?"

The only way to maintain the profitability that funds continued product development, Radnor said, is by offering "a total on-demand solution"-using the "on-demand" label that IBM has made a centerpiece of its current marketing efforts.

Traditional competitive analysis and forecasting won't suffice in multinational service-sector scenarios, warned John Sargent, a senior adviser to the assistant secretary for technology policy at the U.S. Department of Commerce, in Washington. "Nations that traditionally had to develop a lot of infrastructure to grow their economies can now bypass many of the steps of national economic development and jump right to the top part of the value chain," Sargent said.

Season goods with services
Industry practitioners are enriching their basic business models with service-based enhancements. At Deere & Co., the Moline, Ill., maker of John Deere agricultural equipment, the effective performance of the company's essentially mature farming technology is dramatically improved by the company's provision of differential GPS (Global Positioning System) services.

"We sell a worldwide correction signal for GPS that lets farmers position their equipment within 3 centimeters," said Tom Hein, Deere's manager of technology, during an Almaden conference session on services in the manufacturing sector. By reducing the waste of costly fertilizers and other chemicals, that precision translates to increased profit per acre more cost-effectively than by trying to get the same improvement out of other steps in the process.

Deere's leveraging of the base-line, freely available GPS technology should call attention not just to the power but also to the pitfalls of building service offerings on platforms provided by others. The edge that comes from using an outside provider's best-of-breed platform can be blunted by two hazards that the forward-looking IT architect must be prepared to address.

The first risk is losing competitive differentiation as one's competitors adopt the same sold-to-all-comers services. The second risk is that one will give away high-value knowledge to the service provider in the process of fine-tuning the service relationship, which the service provider will then be in a position to sell to one's competitors in the consulting role that many such providers are taking on.

"The issue of trust comes up again and again," said conference speaker Robert Johansen, president and CEO of the nonprofit Institute for the Future, in Palo Alto, Calif., "and I almost get tired of it because it always seems like a surprise."

The challenge of choice
Many enterprise developers will also find more than one viable supplier for key services, unlike the case with GPS, which is pretty much the only player in its game. In competitive situations, an enterprise team must strike a careful balance: It can try to make the most of a cultivated relationship with its service provider of choice, but it should maintain flexibility by confining itself, where possible, to common-subset and standards-based services. This preserves a competitive market for service provision.

That said, a development team then has an opportunity to assist the enterprise in strengthening customer loyalty beyond the point that's achieved by providing even top-tier goods alone. Services build on themselves; an application can use customer history to improve the accuracy of future service suggestions. In-house IT teams are needed to identify these opportunities, to evaluate and monitor outside services partners, and to maximize flexibility to adopt enhanced or alternative service offerings by properly using loosely coupled integration technologies.

Those who build services must see service consumers in many different roles, said Almaden conference speaker Scott Sampson, associate professor of business management at the Marriott School of Management at Brigham Young University, in Provo, Utah. "The customer may be a supplier," said Sampson, providing important input to the process-but that input may be unreliable, he warned, in timeliness and quality. The service delivery process must either absorb the resulting cost or find a way to help the customer do a better job.

The customer may be labor, Sampson continued, "but the customer-labor may ignore, avoid or reject technologies or process improvements." Achieving buy-in from customer-labor is even more important, therefore, than achieving it from the kind of labor that can be fired if it won't do as it's told. User interface design, ease of access to help systems and live support are all key IT contributions to this role.

In a service operation, the customer is also inventory-and expensive inventory at that. A doctor's waiting room and the waiting line at a theme park ride are visible examples of the idea and should serve to warn a service designer of the need to manage queues with care. Mere averages are not an acceptable figure of merit: A more sophisticated model of peak-to-average ratios in waiting times and service times is needed.

With services the key to continued economic growth, the Almaden conference participants agreed that service creators need to move beyond the hunter-gatherer stage-and figure out where, and how, to drive a plow.

Technology Editor Peter Cof-fee can be reached at mailto:peter_coffee@ziffdavis.com peter_coffee@ziffdavis.com.