But that's actually something of a blessing. By relying too heavily on ROI metrics and projections, companies tend to opt for proven or "incremental" technologies instead of zeroing in on potential breakthrough tools that could deliver greater long-term returns. "You have to get over the sort of mindless application of metrics biased toward efficiency and cost reduction," Austin says. John Petrey, CIO of TD Banknorth, describes his financial services company as a fast follower. One exception came in 2003, when the company learned that its middleware vendor would no longer support the software integrating its online banking and backend systems.

At that point, Petrey saw an opportunity: Web services could provide an alternative to existing options. What's more, he figured, they probably wouldn't be obsolete before they were deployed. Rather than discuss the technology in terms of bits and bytes, Petrey pitched the business side by explaining how Web services, operationally, could pay dividends long term. Quantifying ROI posed a challenge. Instead, Petrey and his team analyzed the potential operational gains, as well as how to mitigate any implementation and execution risks. "The business first had to convince itself this wasn't just an IT organization that saw some cool technology it wanted to go after," he says.

Marriott tends to be conservative about adopting emerging technologies. "The only time we'd rapidly adopt [an emerging] technology is if it's low risk and wouldn't disrupt our business," Wolf says. When a new technology piques Wolf's interest, he and his team do an internal evaluation to classify the technology as "core" (fundamental to operations), "mandatory" (not crucial, but consistent with architectural standards), "optional" (could be used if a clear business case is presented) or "emerging" (under evaluation). Usually, he says, the new technology remains in the "emerging" bucket; occasionally it's moved to "optional." In rare cases, though, the technology is elevated to "mandatory" status.

To prevent missing the boat on technologies, Wolf recommends creating an internal "innovation center," staffed with expert developers and programmers who can test and evaluate technologies as they emerge. He's looking to develop this capability at Marriott.

"We're the largest hospitality company in the world, so we have everything to lose," he says. "Everybody's gunning for us, so we have to have enough technology to deflect those shots against us."