Concept
Kill Criteria
Kill criteria are conditions, set before a commitment begins, that trigger quitting or at least a serious review if they are met. In Quit, Annie Duke sums up a good criterion as two parts: a state and a date.
Once we are inside a project, our judgement about it is compromised: we have invested, people are watching, and the evidence arrives in pieces that are each easy to explain away. Kill criteria move the decision to the start, before there are sunk costs or an audience. The practice is familiar in fields such as drug development and venture investing. In Duke's formula the state is a condition you can observe (fewer than twenty paying customers, a unit cost above a set level, a monkey still unsolved) and the date is when you will check. 'If we have not reached this by then, we stop' can be checked; 'we will see how it goes' cannot. Criteria also change what quitting means: nobody confesses failure, because the agreed condition has simply been met. Two cautions matter. The urge to move a criterion is strongest exactly when it fires, so criteria should be written down, shared, and changed only with a recorded reason. And a criterion does not replace judgement; it triggers a review.
A real-life example
The owner of a furniture workshop in Kayseri gives her accountant the criterion for her online shop: if orders are below fifteen a month or returns above ten percent by the end of September, paid ads stop. In September there are nine orders; the ads stop and the budget goes to a trade fair for hotel buyers.
How to use it
- 1For each serious commitment, write one sentence with a state and a date: if this has not happened by then, we stop.
- 2Give the sentence to someone who will read it back to you on the date.
- 3If you want to move a criterion, write down what new information justifies the change.