Write the hypothesis card before building: tests with a pre-registered metric ship a decision 2x more often.

Write the hypothesis card before building: tests with a pre-registered metric ship a decision 2x more often.
The one-page template that prevents 'we're not sure what happened.'
— Because we observed [data, e.g. 30% phone-field drop]…
— We believe [change] will cause [primary metric] to move by [≥ MDE]…
— For [segment], measured over [runtime/sample]…
— We're wrong if [guardrail breaches or effect < MDE].
No card, no launch. An untestable hypothesis produces an unactionable result every time.
Decide the decision rule before you see the data. Read the number, not the story. [1 card per test]