What I Look For When a Programme Says It Is Recovering

There is a moment on a lot of troubled jobs where the reporting turns optimistic — “we are recovering” — and the question is whether the evidence agrees. Reported recovery is a statement. Demonstrated recovery is a trend you can see in the data. I look for the second, not the first.

The trend, not the snapshot. A single good month is not recovery. I look at the trajectory over several cycles — is the gap to the plan actually closing, or is it holding steady while the narrative improves? Recovery is a direction sustained over time; one cycle tells you almost nothing.

Milestone movement. Talk is cheap; milestone dates are not. Are the interim and contractual milestones actually pulling back towards their targets update on update, or are they static while activity-level percentages tick up? A programme can be very busy without any of that activity reaching a milestone that matters.

Critical-path movement. This is the one that gets missed. If the job is genuinely recovering, the critical path should be shortening or holding. If forecast completion keeps slipping while everyone reports recovery, the recovery is happening somewhere that is not on the driving path — which means it is not recovering the finish date at all.

Float behaviour. I watch float across the network. Recovery that is real tends to rebuild some float or at least stop it eroding. If total float is still going negative while the programme claims to be recovering, the network is telling you the truth the narrative is not.

Productivity and remaining scope. I compare the productivity the recovery assumes against what is actually being achieved, and I look hard at remaining scope. It is easy to “recover” by burning through the straightforward remaining work while the difficult, high-risk scope sits untouched at the end. A percentage-complete that is racing ahead of physical progress on the hard work is a warning, not a comfort.

Forecast reliability. And I ask whether the forecast the recovery produces is honest — whether it follows from real progress and logic, or whether the completion date is being held in place by a constraint while everything behind it slips. A forecast that never moves despite visible slippage is not stability; it is a date being defended rather than calculated.

The reason this matters is that reported recovery, accepted at face value, delays the decision that the job actually needs. Every cycle that “we are recovering” goes unchallenged is a cycle in which the real position is not confronted — and the options to do something about it narrow. The most expensive words on a troubled programme are the ones that let everyone avoid the harder conversation for another month.

The takeaway: do not accept that a programme is recovering because it says so. Test it against the trend, the milestones, the critical path and the float — and if the evidence and the narrative disagree, believe the evidence. The framework behind reading this is in Forecast Integrity and the Director’s Playbook.