8 min read
How to recover a failing project, without lying to the board
Most rescue plans fail because the assessment was never honest. Here is the triage model we use before we agree to take a recovery mandate.
A failing project rarely announces itself. The status board stays green while everyone in the room privately knows it is red. The date slips a week every week. The board has quietly stopped believing management, and a hard go-live is closing in. Recovery starts by ending that fiction.
Step one: get an honest baseline before anyone promises a turnaround
The first job is not a plan. It is a true read of where the program actually is. Read the artifacts, interview the people who ran it, and stress-test the plan that already exists. Within two to four weeks you should have a real baseline: scope, critical path, root cause, and a straight call.
The watermelon problem: green on the outside, red on the inside. A recovery that does not first break the false status reporting will inherit it.
Step two: make the rescue, rewrite, or kill call on remaining cost, not sunk cost
The recommendation has to be grounded in remaining cost against expected benefit measured from today, not in what has already been spent. A team selling a recovery mandate has a financial interest in the recovery going forward; that conflict is exactly why the kill recommendation has to be available. Independence is the product.
Step three: put one accountable operator in the seat
Recovery does not run by committee. One owner steps inside the program with real decision rights, re-baselines scope from reality instead of from what was promised, untangles the dependencies that make fixing one workstream break three others, and absorbs the political heat of the unpopular calls.
Step four: ship against a plan a board will actually accept
Delivery then runs on cadence and accountability against the re-baselined plan, with milestones a board, a lender, or an investor will accept. The work ends with the program shipping, not with a report someone else is left to execute. Advisory is how some engagements start; execution is how every engagement ends.
- Baseline first, plan second.
- Decide on remaining cost, not sunk cost.
- One accountable seat, not a committee.
- Milestones a lender would sign.
Project recovery is one of five capabilities at RAVENTRYX. If a program is already in distress, the triage above is where a mandate begins.
Common questions
When should a failing project be killed rather than recovered?
When the remaining cost to finish exceeds the expected benefit measured from today, not from what has already been spent. Sunk cost is not a reason to continue. A straight rescue-versus-rewrite-versus-kill call is the first deliverable of a credible recovery, and the kill recommendation has to be on the table for the assessment to be honest.
How long does a project triage take?
Two to four weeks for most programs. The triage reads the artifacts, interviews the people who ran it, and stress-tests the existing plan to produce a real baseline: scope, critical path, root cause, and the recommendation. It is the discovery phase, and it is built to move fast.
Who owns the recovery once the plan exists?
One named operator with real decision rights inside the program, not an advisory layer beside it. Reporting by committee is part of how the project failed. Recovery runs through a single accountable seat that re-baselines scope from reality and absorbs the political heat of the unpopular calls.