When should you stop an experiment?

Stop an experiment when its riskiest assumption has been tested sufficiently and customer behaviour does not support it. The decision is made against a metric agreed in advance – not on gut feeling.

Agree the stop criterion upfront

Before starting, write down what you measure, what counts as enough and by when.

Separate signal from noise

• A click is not a purchase decision
• Praise is not commitment
• Contact details are a stronger signal than a like

A decision to stop is as valuable as a decision to scale.