1. Get Business-Side Buy-In

Stephen Rood, CIO of Strategic Technology, gives business managers a financial stake in determining which enhancements to ongoing projects they really need. Every year, each department gets a pool of developers' time for system enhancements. If the head of manufacturing has five projects, Rood calculates the hours for each and may report that he only has the resources for two of them."If they want to do a third one, I don't say no," says Rood, but the department has to pay for a contractor. "That gives them a certain degree of ownership," he says.

2. Be the Voice of Reason

CIOs must-gently and tactfully-deflate user enthusiasm for spanking-new systems, or the backlog can get out of control. Ken Kudla, CIO and VP of IT with The Queen's Medical Center in Honolulu, explains to his senior management peers why the time might not be right for a project from a cost perspective or a technological perspective. "Then we sit down as VPs together and talk about [the idea's] merits and whether we agree or don't agree," Kudla says. "Then it's not just one voice [making the decision]."

3. Don't Bend the Rules

Every business manager has a project that just has to get done yesterday. Although CIOs tend to want to show that they're helpful, it isn't necessarily a good idea to bump new requests to the front of the queue. "If a project doesn't meet the criteria of a high-priority project, I'm not going to reprioritize all of the IT resources [for it]," says Bob Holstein, CIO of National Public Radio. He also doesn't make casual assurances to anyone about their IT projects. "A promise is not a promise until I assign a date and a project manager," he says.

Also see The Number-One Problem of CIOs: The Project Backlog.