Experiment Lab

    A/B testing essentialsfor product managers

    Foundations, statistics, and a launch readiness checklist. Free reference material for the math and judgment behind every experiment.

    Launch Readiness Checklist

    Verify each requirement before launching your experiment. Every item represents an industry-standard practice that helps ensure valid, reliable results.

    0 of 12 completed

    Question & metrics

    Hypothesis is falsifiable and documented
    Exactly one primary metric defined
    1–3 secondary metrics documented
    At least 2 guardrail metrics with degradation thresholds set

    Sizing & duration

    Target population and exclusions defined
    Minimum Detectable Effect (MDE) justified by business impact
    Sample size calculated and reviewed
    Duration covers at least 2 full weeks

    Decision & launch

    Decision criteria pre-registered (ship, iterate, or revert)
    A/A test passed this quarter
    Rollout ramp-up plan documented
    Experiment design peer-reviewed before launch

    Recommended Statistical Defaults

    ParameterDefaultRange
    Significance level (α): false positive tolerance0.050.01 to 0.10
    Statistical power (1 - β): detection probability0.800.80 to 0.95
    Minimum Detectable Effect (MDE)Varies by metric and business contextDetermined per experiment
    Minimum experiment duration14 daysNon-negotiable minimum
    Maximum experiment duration56 days (8 weeks)Extend only with documented justification
    Test directionalityTwo-tailedOne-tailed requires pre-registered justification

    Beyond the theory

    If you've got the theory down, see how it plays out in the simulator.

    See the simulator