The project healthcheck: what a truthful one looks like, and why most are theatre

Somewhere in most large organisations, right now, a programme review is concluding that a troubled programme is “amber, trending green”. The phrase should trigger suspicion on sight. Programmes do trend green occasionally; reviews conclude it constantly. The gap between those two frequencies is the healthcheck theatre industry.

The theatre has a consistent anatomy. The review is commissioned from a party with an interest in the answer – the delivery partner assuring its own delivery, the integrator hoping for the recovery contract, the internal PMO whose reporting is itself part of the question. The method is documentary: RAID logs are inspected for completeness, governance forums for attendance, plans for the existence of milestones. And the report is negotiated – findings softened draft by draft until the version that survives is the version everyone can live with. Every step is professionally executed. The product is a well-formatted document that measures the programme’s paperwork and calls it the programme’s health.

A truthful healthcheck differs at each step, and the differences are learnable.

It is independent in fact, not in name. The reviewing party must have nothing to gain from any particular finding – no recovery work to win, no prior advice to defend, no ongoing relationship the finding could sour. This is not puritanism; it is mechanics. Reviewers shade findings toward their interests not through corruption but through the ordinary human preference for conclusions that do not cost them anything. Remove the interest and the shading has nothing to grip.

It examines assumptions, not artefacts.Every programme rests on a small set of load-bearing assumptions: the integration will behave as the vendor describes; the data migration is a formality; the business will absorb the change at the planned rate; the contract’s incentives and the plan’s needs point the same way. Troubled programmes are almost never short of documentation. They are carrying a wrong assumption in a step everyone treats as obvious – and no completeness check will ever find it, because the defect is not an absence. A truthful review identifies the five or six assumptions the programme cannot survive being wrong about, and tests each against evidence rather than assertion.

It follows the information, not the org chart. Reporting chains in troubled programmes do not merely transmit information; they process it, and each layer has reasons to process it optimistically. The review must therefore triangulate: what the dashboard says, what the delivery teams say, and what the artefacts – the actual backlog, the actual defect trend, the actual environment availability – say. Where the three diverge, the divergence is the finding.

It reports in consequence order, with the seams showing. A truthful report leads with the answer a decision-maker needs, ranks findings by what they would cost if unaddressed rather than by ease of discovery, and labels its own epistemics: this we verified, this we were told, this we assumed. A report that blurs those categories asks the reader to extend the review team a trust the review team has not earned on the page.

The objection we hear most often: is this not needlessly adversarial – does a hard-edged review not damage the programme’s morale and its relationships? Our experience is the reverse. Delivery teams are usually relieved to be examined by people interested in the truth; they generally know where the bodies are buried and have often been trying to report it upward for months. The morale damage comes later, from the theatre version – from the two further years spent executing a recovery plan built on the same unexamined assumption.

The economics deserve stating plainly. A rigorous healthcheck on a significant programme costs weeks and a modest fee. The failure it exists to catch costs quarters and a career or two. Commissioned early – before the symptoms, when it feels least necessary – it is among the cheapest insurance an executive can buy. Commissioned late, it is still cheaper than the alternative, which is finding out from events.