Know when to bail out

The systems manager is supposed to support Union Gas, Ltd.'s first client/server application - a customer information system based on distributed Windows NT servers and an Oracle back-end database. If, that is, the application ever gets rolled out. The Chatham, Ontario-based utility is about eight months into the project, and Lamb sees red flags everywhere.

"I get concerned when a project doesn't have any clear owner or sponsor," Lamb says. That's certainly the case with this application, which has tripartite management sponsorship from executives at Union Gas, IT outsourcer Information Systems Management Corp. (a business unit of IBM) and Union Gas' parent company, Westcoast Energy, Inc. "We've got a three-ring circus going here," he says.

The project also suffers from lack of accountability and communication and the classic critical failure factor - missed deadlines. "We're not hitting our milestones. We're three to four months behind, and that's after we pushed the deadlines out three months to start with," Lamb says. The budget is history at this point, too, he says, but even that is not such a big deal.

Time is the real problem. The floundering application has to be in production by the end of next year because it is replacing mainframe-based systems that aren't Year 2000-compliant."High-level management needs to step in. This is a cash register application that we need," Lamb says. About 800 Union Gas employees will be using the customer information system to support 1.2 million customers across Canada.

Failures are frequent

If the project fails, at least Lamb is in good company. A recent survey of 500 IT directors by Sequent Computer Systems, Inc. found that 76% of them had experienced a major project failure at some point in their careers (see graphic, page 47). Indeed, project failures seem to be a cost of being in the IT business.

The rate of project failures is accelerating, says Gopal Kapur, president of the Center for Project Management, a consulting firm in San Ramon, Calif."Project complexity is increasing because there are many more unknowns and technology changes are faster and more severe than ever before," Kapur says."Since the unknowns are higher, the risks are higher." That translates into more project failures and stress for those who work in the field. Sequent's survey revealed that a majority of projects fail simply because of a lack of project management, says Steve Wanless, senior marketing manager for Sequent, in Beaverton, Ore."Managers don't understand the scope of the project, or they don't manage it well," he says. A full 66% of the respondents blamed project failure on changing user requirements, a sign that the deadly "scope creep" has occurred.

Increased project spending late in the game should raise another red flag."Managers get the attitude they're going to roll it out come hell or high water. People are so committed at that point that throwing more money isn't too much of a problem," Wanless says. But this attitude can be fatal - a sign that ever-increasing amounts of money will be spent as the doomed project limps along.

The key questions to ask the sponsor prior to launch are"Under what conditions should the project be shut down?" and "What are the metrics of project success?" Kapur says. The first question is tough to ask because you're acknowledging that the project could fail. But if you never know the answer, it's difficult to know when it's time to pull the plug. By human nature, people hate to admit defeat and kiss goodbye the effort they've put into a project.

Jim Kinney, chief information officer at Kraft Foods, Inc., in Northville, Ill., admits that project managers at his company don't ask that question before launching projects."That would be a pessimistic approach," Kinney says.

Kraft does have an ironclad process for weeding out viable projects from half-baked ideas, another best practice that Kapur recommends. Kraft developed its Systems Development Process (SDP) project management methodology in-house with the aid of RWD Technologies, Inc., a consulting firm in Columbia, Md.

The cornerstone of SDP is "3X3"alignment, which means Kraft doesn't undertake a project until there is agreement among the project manager, the business manager from the unit that will benefit from the project and a senior-level executive sponsor.

"We only attack problems whose time has come," Kinney says."Since we have a clear alignment of objectives for starters, it helps curb scope changes."

However, even Kraft must pull the plug occasionally. When Kinney calls it quits on a project, this usually occurs right after the pilot stage, when it's clear the initiative didn't meet users' needs.

Users most often register their displeasure through specific feedback during the rollout or outright resistance to using the new system. If users don't agree with the sponsor's vision of the project goals, Kinney bites the bullet and shuts the project down instead of wasting excessive amounts of time and money trying to deploy it to a wider audience.

Still, Kraft's IS department is getting better at avoiding project failures and is often able to relaunch a project that wasn't initially well-received by pilot users. "Pilots sometimes reveal a fundamental problem. I don't think that's necessarily a failure," he says.

Adhering to solid project management principles has given Kraft recurring results, Kinney says. However, at this point it may take more than good project management to save the project at Union Gas.

For his part, Lamb is already numb to the possibility of failure."You sort of check out," he says."You make yourself heard and you're ignored, so you say,'Forget it.' "

In fact, Lamb was so fed up that he resigned from Union Gas; his last day is this week. But there's no getting away from the faltering application.Lamb took a job with the project contractor and hopes he can turn help turn things around.