I have been involved with attempts to rescue failed projects. The ones that seemed to turn around like magic are where there is (1) executive support for the recovery work, and (2) the new leader has the time and authority to understand the root cause of the failure.
I agree in general with the abstract notion that the effort to rescue a failed project is itself a project (DUH- wink), but beyond that, let’s discuss and define what is different about the rescue project in terms of its characteristics, especially the critical success factors: focus on the process failures and understand why the project failed, honest communications, adequate resource commitment, and executive support.
Morale, or lack of it, is usually a big issue for failed projects, as is effective communication between team members and the product owner. These have to be overcome immediately, or the rescue project will never make any progress, and the project may fail again.
Without the intial focus on determining the root cause, many second attempts fail again from “get thereitis”, the compulsion to see a late project meet some deadlines as a measure of the success of the turnaround. The most important focus for the recovery project is on fixing the process issues rather than the goals of the original project. I have had to explain the need to determine the root cause of the project’s failure to management, and I usually describe it as the ?best way to avoid repeating the earlier failure again?. Often, a second failure would be the final failure, so people are usually receptive to an experienced turn-around agent. A good technique to use to uncover root causes of a process breakdown is the ?5 Whys? technique of Six Sigma.
Using the root cause as the diagnosis, the prescription should be to get the project back on track, and make adjustments to the process to avoid repeating the failure. You may need to remove unreasonable expectations and to refocus project requirements, commitments, and the schedule. You will certainly need the executive support I mentioned earlier to accomplish this.