Experiment best practices

Interactive simulators that show the most common ways tests go wrong, and how to avoid them.

Significance Testing

See how statistical power, effect size, sample size and significance level relate to each other.

Understand how checking results too often inflates false positive rates.

Set manual or statistical guardrails to stop tests when key metrics fall below acceptable levels.

Explore how uneven traffic splits affect statistical power and sample size requirements.

Learn how variability differs for proportion and continuous metrics, and how it affects the minimum detectable effect.