Sunk cost fallacy
The economic principle: past investment is irrelevant to forward decisions. The behavioral reality: we feel like quitting wastes the prior investment. The discipline: ask "If we hadn't spent the last six months, would we start this project today?" If the honest answer is no, the previous six months don't change the future six months' calculation.
Escalation of commitment
Doubling down to recover the original investment. The team adds people, the schedule extends, scope expands — each expansion compounding the original mistake. McConnell calls this one of the worst classic failures. The fix is hard: recognize the dynamic, surface it explicitly, force a decision rather than letting the project drift further.
↳ in the wild
The discipline most senior PMs eventually learn: it's better to kill a project early and ship a postmortem than to limp it across the finish line and ship a half-product nobody uses. The career-protective move (don't-have-a-failure-on-record) is usually the career-damaging move (have-a-half-product-on-record).
How to actually kill it
- Surface the analysis. Cost-to-complete vs. value-on-completion. Honest numbers.
- Name the alternatives. What else could the team / budget do that's higher value?
- Recommend. Don't leave the call to leadership without a recommendation.
- Run a closure. Per Track 01 lesson 8 — even cancelled projects deserve closure docs.
- Communicate. Sponsors, team, stakeholders. Frame as redirected effort, not failure.
When NOT to kill
- The struggle is teaching the team something they'll need on the next project.
- External constraints (regulatory, contractual) make completion mandatory regardless of value.
- The metric driving the kill recommendation is itself unreliable (e.g. you don't actually know the value-on-completion).