A/B test significance and sample size calculator

Is B really better than A: p-value, confidence interval and the sample needed for the lift you care about.

A/B test and sample sizeIs B really better than A, and how many visitors does it take?

VerdictB differs from A at 95% confidence

Conversion A3%

Conversion B3.5%

B against A16.5%

p-value0.0299

95% interval of the difference0.05 → 0.94 points

Sample per group to detect 10%53,211

Do not stop the moment it looks significant: peeking raises false positives from 5% to about 30%.

Visitors A

Conversions A

Visitors B

Conversions B

Confidence

Power

Smallest lift worth detecting (relative)%

What you enter is kept in this browser; a share link carries exactly these numbers.

Run it on your own figures

Upload your orders or report to DataMimi and ask in plain English — this calculator opens prefilled with the prices, volumes and costs from your own file, and every figure shows which rows it came from.

Open DataMimi

Related calculators

Profit sensitivity analysis (tornado chart)Move each driver ±10% and rank which one moves profit the most.Monte Carlo profit simulationThousands of scenarios at once: P5, P50 and P95 profit and the chance of a loss.