To calculate A/B test sample size, you need: Baseline conversion rate (current conversion rate before testing), Minimum detectable effect (MDE) (smallest relative change worth detecting — typically 5-20%), Statistical significance (usually 95%, meaning 5% false positive rate), Statistical power (usually 80%, meaning 20% false negative rate). Our calculator uses the industry-standard formula and produces the number of visitors needed per variant — you need at least this many visitors in each group before ending the test.