Nobody decided to be here
Programmes rarely fail in one meeting; they fail across eight unremarkable ones. Each extension was individually defensible (a dependency slipped, a vendor re-planned, a scope item turned out to be three scope items) and each approval was smaller than the courage required to stop. The cumulative position was never on any single agenda, which is why a programme can be eighteen months late without the organisation ever having chosen to accept eighteen months of lateness. Recognising that the current state was arrived at rather than decided is the first step towards making an actual decision about it.
Count the incentives in the room
- The sponsor's reputation is fused to the programme: killing it reads as a verdict on their judgement, so their optimism is structural rather than personal.
- The systems integrator bills by the month and re-plans for a living: an extension is revenue, a rescue is a change order, and a kill is a lost account.
- The delivery team reports green because amber invites scrutiny and red invites blame: status reporting has become career risk management rather than programme information.
- Steering committee members approve extensions because a write-off would land in someone's current-year numbers, and nobody volunteers to host it.
- Finance flags the variance but does not own the outcome, so its warnings arrive as commentary rather than as decisions.
The question that ignores the past
Everything already spent is gone on both branches of the decision, which makes it irrelevant to the choice, and emotionally central to everyone in the room. The reframing that works is brutally simple: if this programme did not exist, would we start it today, paying only the remaining cost, to get only the remaining value? Put formally on the agenda, that question routinely produces a different answer from the one the status packs imply, because it strips out the two things doing most of the talking: the money that cannot be recovered and the reputations attached to it.
Continue and kill are not the only doors
Binary framing flatters the incumbent plan, because a full kill looks catastrophic next to one more quarter. The options worth pricing sit between the poles: descope to the increment that actually pays for itself, pause and re-contract with the integrator on different terms, replace delivery leadership while keeping the objective, or salvage the completed assets into a smaller programme. Each of these has a cost and a face-saving path attached. A steering committee that sees three priced alternatives behaves differently from one shown a cliff and a treadmill.
Who can answer without flinching
Not the sponsor, the integrator or the delivery team: each of their answers is predictable from their position before they speak. A credible answer needs people with nothing at stake in the current plan: finance building a today-forward case from scratch, an independent delivery review with explicit licence to report red, and operators who have taken comparable write-offs elsewhere. Selected senior operators from the Global Board who have stopped, rescued and rescoped programmes of this scale can say what stopping actually cost them, and what continuing cost the organisations that could not bring themselves to choose.
