A/B Test Calculator

A/B

A/B Test Calculator

Experiment planning & significance analysis

Sample size & duration

How many visitors per variant you need before you can trust the result. Computed before launching the test.

Baseline conversion rate · current rate of variant A

Minimum detectable effect (MDE)

Smallest uplift worth detecting. 10% relative on a 2% baseline = target 2.20%.

Statistical significance · 1 − α

Statistical power · 1 − β

Hypothesis

Daily visitors · total across both variants (optional)

/day

—

visitors needed per variant

Total (2 variants)—

Baseline → target rate—

Absolute effect—

Estimated duration—

Significance & confidence

Enter the results of a finished test to see whether the difference is real or noise.

Variant A (control)

vis

Variant B (test)

vis

Confidence level

• Enter data to see the verdict.

Conversion A / B—

Absolute uplift—

Relative uplift—

p-value (two-tailed)—

95% CI of difference—

CI conv. A (Wilson)—

CI conv. B (Wilson)—

Two-proportion z-test · normal approximation · Wilson score intervals for single rates.
All calculations run locally in your browser. No data leaves this page.