Law of Escalating Commitment
The more publicly someone has backed a project, the harder they'll fight to keep it alive — because killing it would also mean admitting they were wrong.
Why PMs should care
Once an executive has announced a project in an all-hands, written about it in the quarterly board deck, or promised it to a specific customer on a sales call, killing it is no longer just a resource-allocation decision. It's a personal reputational loss for them — and they'll fight to avoid that loss with more energy than they originally put into backing the project.
A smart PM either gets the project right before it becomes publicly owned (narrow audience, revocable commitments, scoping docs that don't leak), or — if the moment to kill has arrived anyway — gives the exec a face-saving way out that reframes the decision as a thoughtful pivot rather than a failure.
The face-saving way out almost always needs new language. 'Pivoting to focus on X based on customer feedback' is a far better frame than 'killing Y'. Same decision. Completely different political cost.
Example in product work
A senior leader has spent six months publicly backing Project Compass — mentioned in three all-hands, on the quarterly investor update, and to a strategic customer who was assured it would be ready by Q3.
The data now clearly shows it should be shelved: adoption in beta is 4%, the technical approach has turned out to be more brittle than expected, and the next two quarters of effort would be better spent elsewhere.
— Amateur move: a blunt memo titled 'Why we should kill Compass'.
— Senior move: work with the exec's chief of staff to draft the framing 'Evolving Compass into a lighter-weight capability within Atlas, following Q2 beta learnings'.
Same outcome for the roadmap. Engineering team gets redeployed. The exec gets to tell a story of responsive leadership rather than a failed bet.
The second version is how good decisions actually get made in the presence of humans.
What to do when you see it
- Once an exec has announced a project in an all-hands or promised it to a customer, killing it stops being a budget decision and becomes a reputational one.
- Get the project right before it becomes public (narrow audience, revocable commitments, scoping docs that don't leak).
- If the moment to kill has arrived anyway, give the exec a face-saving way out that reframes the decision as a thoughtful pivot, not a failure.
- 'Pivoting to focus on X based on customer feedback' is a much better frame than 'killing Y'. Same decision, completely different political cost.
Sources & further reading
- Knee-deep in the Big Muddy — Barry M. Staw, 1976The foundational empirical paper on commitment escalation in decision making.
- Commitment Bias (Escalation of Commitment) — The Decision LabA well-sourced explainer on why decision-makers throw good money after bad.