Kill criteria
A type of decision criterion for shutting down.
Prodability definition
Kill criteria are the conditions defined in advance that, if not met after a defined number of tests, lead to shutting down the idea. Defining them before the test is essential because, once a project is under way, the emotional pressure on those who have already invested time and resources pushes them to read negative data as exceptions or as signals to adjust. Setting out in advance which outcome corresponds to which decision — continue, modify, abandon — keeps you from confusing courage with stubbornness.
Understand the concept in depth
A guided exercise in three phases: your own definition, exploring the sources, and comparison with the parts that hold the concept together.
Kill criteriaNot to be confused with
- Test success threshold
The success threshold of a single test defines pass/fail for one metric; a kill criterion aggregates the outcome of several tests and sets the condition for shutting down the idea, not the single experiment.
- Pivot
A pivot is a change of course that keeps the project alive; a kill criterion identifies the threshold beyond which further pivots become mere indecision and the idea should be abandoned.
Other perspectives on the concept
Authoritative definitions from different angles, to compare in the “Help me understand” exercise.
Kill criteria are stop rules declared up front: combinations of metrics and thresholds whose simultaneous failure signals that the underlying assumption cannot be rescued with further pivots.
Sources
- Osterwalder, A., Pigneur, Y., Bernarda, G., Smith, A. (2019), Testing Business Ideas — A Field Guide for Rapid Experimentation — source
