David Roper, managing partner at CIO Partners, has been consulting to CIOs as an independent IT management consultant for the past couple of years. He says in exiting a program, you should make use of all communication channels, give the bad news up front, then concentrate as much as possible on the positives.

In one case, he says, when an organization where he was previously employed decided to abandon a $20 million project due to a shortage of capital, it eased the pain of pulling the plug by getting staff to focus on the future.

"I don't think they would have ever felt that [exiting the project] was a better outcome than had the project continued, but it was definitely better than simply being assigned to do something totally different or alternatively losing your job or having to go somewhere else," Roper says.

"During the decision as to whether the project should or shouldn't go ahead there was lots of discussion by lots of people about whether the project was successful or wasn't successful, and for what reason. Ultimately the organization knew that it had to cancel one or more initiatives and was looking for good reasons to stop some or all of those initiatives so there was the usual politicking that went on during that process.

"I suppose what's important to survive that process is having a very good handle on what it is that the project is trying to achieve and how you are going at any point in time against achieving those objectives, just so that if anybody does start questioning the success or otherwise of your project you've got the answers at hand to quickly pour water on any of those perceptions. Whereas if you just have a gut feel or haven't got ready answers to political-type issues, then you can easily find yourself ambushed and decisions get made before you can present those facts," Roper says.

Warning Signs

For Wolken's Thew, one sign of an urgent need to exit is when project meetings become acrimonious. If you are working with a third party when people start sniping at each other across the room, he says, you should immediately exercise the clause in the contract that allows you to go to mediation. If mediation fails you should go to arbitration.

"When meetings are consistently acrimonious, when the parties can only think about slanging across the table at each other, and hurling blame, then you've got a real problem. That's why I say you bring in mediation because mediation quietens it down and focuses attention back on what they are supposed to be doing," Thew says.

If mediation does not work you go to arbitration, and if arbitration does not work the project ends.

"The issue is you don't pull the plug too early," he says. "You look first to resolving it. Sometimes changing the players can help that: there's a thing in our industry called the disaster specialist project manager. And the disaster specialist comes in, takes over the project, takes over the problems and is a specialist at working inside those environments. The joke about them is that if there is no disaster they will create one, because you have to give them something to do that's usually challenging; they can't work under normal circumstances.

"And the other thing is you give them absolute authority, reporting directly to the most senior management or the board. So they have the authority, they have the funds. It's sort of like the reverse of military strategy: you don't ever hold reserves, throw everything at it up front to get it solved because it's only going to drag on. You throw as much at it up front as you can; you must resource the project adequately to get the thing fixed."

When that does not work, then as long you have based your decision on hard analytics, you can exit with dignity, knowing you have left your biases behind.