R&D Envy

Wolfe enlisted the help of a consortium frequently used by the National Governors Association and state groups to do project-specification consulting, called Search (www.search.org). Search consultants, funded entirely by the Department of Justice, studied North Dakota's needs, provided proof of concept, and did an implementation plan for the integrated Law Enforcement Records Management System. It would have cost North Dakota $310,000. The state received $2 million from the departments of Justice and Homeland Security to implement the system. "This was our first major source of external funding," Wolfe says. Now he hopes to do it again. His next target is a system that links the state's 52 county prosecutor offices. Beyond that, Wolfe says the state's jails could benefit from a combined purchasing and jail-management system. There may not be outside funding for them, but Wolfe believes his efforts have shown a practicality to the state legislature.

Hilton Hotels Corp. also looks outside its own walls to stretch its research dollars. The company spends about $1 million a year, or a bit less than 1% of its IT budget, on research. Some goes for a few staffers exploring technologies such as ways radio-frequency identification chips might help a hotel--as an alternative to card keys for room access, for example. But until the cost of chips plummets to the 2-cent range, RFID isn't practical. So Hilton also piggybacks on the efforts of its neighbors in Memphis by contributing to the FedEx Institute of Technology, a center for the study of emerging technology at the University of Memphis.

The company also has a room at the Hilton Garden Inn Hotel at LAX Airport in Los Angeles, where it tests how new technologies, such as retinal scanners or fingerprint readers, work with real guests. (Frequent travelers will specifically request the room, eager for a sneak preview of the future.) In the past year, the research team experimented with various wireless devices connecting to its reservation system. The learning paid off in developing a handheld tool that staffers use to assist guests who have trouble with check-in kiosks, which Hilton plans for 45 hotels this year. It's that track record for experimenting with technologies relevant to the business that keeps the trust to throw $1 million a year at work without a hard-and-fast end goal. "It's not going to be us discovering some new technology," Hilton CIO Tim Harvey says. "It's figuring out what are going to be the really big hitters that will have an impact on our business."

That's what J.C. Whitney & Co. was trying to do two years ago when CIO Marty Wehrle wondered whether a commercial search engine would let the auto-parts retailer's Web site better serve customers than the existing engine developed in-house. The Web site's catalog includes 3.5 million parts, each with an average of 10 possible uses in various vehicles and models. Finding a specific item without a part number could be daunting.

Wehrle wanted funding to do comparisons, but his department was in a three-year capital project to upgrade all applications, so there was no available money. Wehrle's director of technology, Geoff Robertson, devised a cash-free way to research the idea. Robertson convinced Mercado Software to lend J.C. Whitney a copy of its search engine, then split the Web traffic between mirrored sites, one using Mercado and one using J.C. Whitney's search engine. Mercado Software loaned the use of its engine for the chance to win the business.

"It was a live test," Wehrle says, and the results showed that the commercial engine didn't make a big enough difference to justify its purchase to replace the internally developed one. Until Wehrle conducted the no-cost experiment, he didn't know whether his own search engine was good enough to convert site visitors to buyers.

One thing is certain as IT departments wrestle with these challenges: A research and development mind-set depends on the attitude communicated from the top.

Sabre CTO Murphy sends the message that good things can come from even random experiments. Sabre has been a pioneer in turning IT into business advantage, whether it was harnessing mainframe technology in the 1960s to gain a majority market share for electronic ticket distribution, or using large-scale, low-cost Linux servers and open-source MySQL databases today for a low-cost way to push out information on airline schedules or hotel availability as it faces tougher competition from growing Internet-based distribution models. Murphy recently canceled a lunch appointment to watch senior engineers cobble together a gigabit-speed Linux cluster. The experiment cost $15,000, and, even so, there are no immediate plans to use such a cluster.

At HIP, CTO Villalba believes the most-innovative IT staffers always will try out new technologies on their own--with or without permission. The difference, Villalba says, is that HIP "sanctions the experiment."

Yet these companies' IT staffers do their part by keeping their curiosity channeled to business needs. At Sabre, gigabit-speed cluster switches would be a logical next step because the current travel-shopping systems for Sabre's Travelocity site run on megabit-speed cluster switches. At HIP, the concept for the nurse's information-gathering application emerged after chief medical officer Ronald Platt challenged Villalba and his team to find a way to eliminate delays in gathering patient information.

For companies like these, supporting experimentation in IT isn't expendable. It's tied to the fate of the business.

Illustration by Marc Rosenthal