Knowing when to stop is harder than knowing when to start. A failing initiative rarely announces itself; it just quietly consumes budget, attention, and good people while everyone hopes the next quarter turns it around. This article gives you a clear method to kill a failing initiative at the right time, without the decision reading as blame, and without losing the trust of the team that built it.
Why Stopping Is So Hard
The main obstacle is the sunk cost fallacy: the more you have already invested, the harder it is to walk away, even though past spending cannot be recovered by continuing. Prospect theory, from Kahneman and Tversky, explains part of this. People weigh losses more heavily than equivalent gains, so admitting a loss feels worse than the loss itself.
Two other forces pile on. Identity: leaders who championed the project feel that killing it kills their judgment. Momentum: teams and vendors are already moving, and stopping feels like waste. None of these are reasons to continue. They are reasons the decision feels bad, which is different.
Decide With Exit Criteria, Not Mood
The reliable fix is to set exit criteria before you are emotionally invested, ideally at kickoff. Exit criteria are specific, pre-agreed conditions under which the initiative stops. Because you set them early, they judge the project, not the person.
What Good Exit Criteria Look Like
- Measurable: a number and a threshold, not a feeling.
- Time-bound: checked at a named date or milestone.
- Leading, not just lagging: tied to signals that appear early, like adoption or unit economics, not only final revenue.
- Agreed in advance: written down and acknowledged by the people who can pull the plug.
Continue, Pivot, or Kill: A Simple Comparison
| Signal | Likely call |
| Core assumption confirmed, results improving toward the threshold | Continue |
| Core assumption wrong, but a real adjacent opportunity appeared | Pivot |
| Core assumption wrong, no adjacent opportunity, threshold missed | Kill |
A Real Scenario
A software company built a standalone mobile app to complement its web product. The goal set at kickoff: reach a defined active-use rate within two quarters, because below that the app could not justify its maintenance cost. By the deadline, usage sat far under the line, and interviews showed customers wanted mobile access to the existing product, not a separate app.
Because the threshold was written at the start, the review was calm. The team was not defending careers; they were checking a number everyone had agreed to. They killed the standalone app and redirected the engineers to a responsive version of the main product. Framing mattered: leadership publicly credited the team for producing the evidence that saved a year of misdirected work. Morale held because the story was learning, not failure.
Common Mistakes and How to Fix Them
Mistake: No exit criteria at all
Without them, every stop looks like a subjective attack. Fix: set measurable, time-bound criteria at kickoff for every significant initiative.
Mistake: Framing the kill as someone’s failure
This teaches teams to hide bad news. Fix: credit the team for producing evidence, and separate the person from the project explicitly.
Mistake: Killing silently
Quiet cancellations breed rumor and fear. Fix: state what you learned, what you are doing with the freed resources, and why the call was right.
Mistake: Endless pivots to avoid the word ‘kill’
Serial pivoting can be denial in disguise. Fix: a pivot must point at a real adjacent opportunity, not just a way to keep going.
Action Steps
- Write exit criteria for every major initiative at kickoff.
- Schedule the review date now, not when things look bad.
- At review, compare results to the pre-agreed threshold first, feelings second.
- Choose continue, pivot, or kill using the assumption test above.
- Announce a kill with the lesson learned and the resource redeployment.
- Reassign people to visible, valued work fast to protect morale.
Conclusion
Killing a failing initiative is not admitting defeat; it is refusing to keep paying for a decision the evidence already reversed. Your next step: pick your most uncertain active project and write its exit criteria today, before you need them.
Frequently Asked Questions
When is the best time to set exit criteria?
At kickoff, before anyone is attached to the outcome. Criteria written after problems appear feel like weapons; criteria written early feel like a shared agreement.
How do I stop a project without demoralizing the team?
Separate the people from the project, credit them for the evidence gathered, and move them quickly to work that matters. Morale drops most when a kill feels like blame or when people are left in limbo.
What if leadership is emotionally committed to the initiative?
Point back to the pre-agreed criteria and the data, not opinions. Committing to the numbers in advance is what protects the decision from ego later.
How do I tell a pivot from denial?
A real pivot targets a specific, evidenced adjacent opportunity. Denial keeps changing the plan to avoid stopping, without new evidence that a better path exists.
References
- Daniel Kahneman, Thinking, Fast and Slow — loss aversion and the sunk cost fallacy.