Kill criteria

Projects rarely die cleanly. Without a pre-agreed line, they drag on, absorbing time and money because nobody named the point at which enough is enough. The longer something runs, the harder it is to stop, since everything already spent feels like a reason to continue. That is the sunk-cost trap. Setting the line before starting means the decision is made by a calm earlier self, not by a person who has spent months on the idea.

• Choose a number that can be measured without argument: revenue, activation rate, paying customers or months of runway.
• Add the date. The number means nothing without a deadline.
• Write it down before the work starts and share it with whoever is affected, so it cannot be quietly moved.
• Put the metric where it triggers a real decision, such as a calendar review, not a chart nobody reopens.
• When the date arrives, decide. Stop, or change the plan on purpose and write a new criterion for the new plan.

Kill criteria are a written line: if this happens by this date, the project stops. They have three parts. A number that can be watched, a date by which it must be reached, and a named person who makes the call. "If fewer than one in five signups has finished setup after sixty days, the product wedge is dropped" is a kill criterion. "Stop if it is not going well" is not.

• Setting the line after a bad week, when it is already shaped by disappointment.
• Moving the goalposts when the date arrives. A criterion that changes when it bites was only a suggestion.
• Setting only a stop condition. Writing the success condition too makes the decision balanced.