Use this when
A campaign or growth test has reached its decision date or produced enough evidence for a review.
Process
- Restate the original hypothesis, primary metric, baseline, threshold, audience, and time window.
- Check tracking quality, sample composition, external changes, and missing data before interpreting results.
- Compare the primary outcome with the threshold and use supporting metrics to diagnose why.
- Separate observations from explanations and state the confidence of each explanation.
- Decide to stop, iterate, scale, or gather more data.
- Define the next action, owner, cost cap, and what would change the decision.
Output
Include decision, evidence, data-quality notes, interpretation, confidence, lessons, and the next experiment or rollout step.
Guardrails
Do not cherry-pick supporting metrics, hide negative outcomes, or imply causation from correlation. Avoid false precision when sample size or tracking is weak.