FORTUNE — Dear Annie: A colleague with a twisted sense of humor sent me your recent column about bouncing back from a big, visible failure, because I’ve just been put in charge of a project that everyone here thinks is probably doomed. Here’s what happened: A team of 12 has been trying for the past nine months or so to launch an internal capability that we have not had up to now. It would save us a ton of money if we can get it to work, but it has turned out to be more complicated than anyone expected, and senior management is so frustrated by the various problems and delays that my boss expects me to cut our losses by chucking the whole thing and starting over with a different team.
However, it seems to me that this thing is fixable, if we can just stop all the finger-pointing and concentrate on making a few essential changes. But I don’t really have much experience with this kind of situation. Can you or your readers recommend a good source of practical guidance on how to save a project that has gone off track? — Glass Half Full
Dear G.H.F.: One place to start would be a highly readable book called Rescue the Problem Project: A Complete Guide to Identifying, Preventing, and Recovering from Project Failure. Author Todd C. Williams, head of eCameron, a consulting firm based near Portland, Ore., has spent the past 25 years advising Fortune 500 companies on what to do about projects that are either headed for a cliff or have already gone over.
MORE: The biggest business turkeys of 2013
Sometimes he recommends scrapping them, but not always. “Deciding whether to go forward depends partly on how much the company has already sunk into it, and whether getting the project back on track can eventually make up the losses,” Williams says. “But the main question is one of strategy. How important is this project to the company’s strategic goals? If the capability you’re trying to build is critical, the project is worth fixing.”
Let’s say that the potential cost savings you mention are significant enough that it makes financial sense to persevere. Williams recommends these four steps to turning your project around:
1. Stop the blame game. “The finger-pointing stage is usually when I get called in,” says Williams. “But with any failure, a lot of the responsibility really belongs with senior management, for not providing clear direction, or not monitoring the project closely enough, or both. Once we point that out, bosses are more willing to shift the discussion away from assigning blame and on to finding solutions” — which leads to Step No. 2.
2. Focus on the facts. “Real data are your best friend right now,” Williams says. “Dig deep into the details of exactly what went wrong at each stage, and why.” Getting to the bottom of each failure point “is like peeling an onion. You need to work down to the center. Suppose, for instance, a critical component arrived late. Why? How can that be prevented in the future?”
He adds that it’s important not to be swayed by any opinion or assumption that can’t be verified. “Often people say to us, ‘When you do your audit of this project, here’s what you’ll find the problem was,’” Williams says. “My answer is, ‘Let me do the audit first and we’ll see.’ Anytime something has gone wrong, there’s a tendency to jump to conclusions — which often is what started the trouble in the first place.”
3. Keep the same team. If you decide to change the whole direction of the project, you might need to enlist people with different skills. But if the strategic goal is the same as before, “don’t fire anyone. Just work with them,” Williams suggests. He notes that this may seem counterintuitive, because management — your boss, for one — “often doesn’t see how you can get different results with the same people. But in most cases, you need to change what the team is doing, not who’s doing it.”
MORE: Big business and the art of lying
Williams says he’s seen many projects flounder because they’re understaffed in critical areas, or because a project manager “overestimated what could be done all at once. The scope of your project may need to be better defined, so that you reach your goal in stages, rather than trying to run before you can crawl.”
4. Communicate clearly and often. As the project leader, Williams says, “you need to make sure everyone understands what they’re building and why, and get everyone moving in the same direction.” That might sound obvious, but a frequent cause of failure is “well-meaning individuals who get so wrapped up in their own part of the work that they make decisions affecting the whole project, without checking with the rest of the team first.” Especially in the early stages of a turnaround, you may have to be a bit more of a micromanager than you’d like, just to ensure that doesn’t happen.
One further thought: Take a hard look at whatever technology you’re using. Does it serve the project’s goals, rather than the other way around? “For a project design to work, it has to be built around strategy, then people, then process, then technology,” Williams says. “It has to be in that order. Putting the technology first will just get you into trouble a whole lot faster and more efficiently.” Good luck.
Talkback: Have you ever been part of a turnaround? What helped get the project back on track — and what didn’t? Leave a comment below.