← back to tools

A/B Test Calculator

Two-proportion z-test with pooled standard error. Drop in conversions & visitors for two variants, pick a confidence level, and get p-value, lift, and a CI on the difference. Same math I used to compare markdown pricing approaches at a NA retailer.

Variant A (control)

Visitors

Conversions

Variant B (treatment)

Visitors

Conversions

Confidence level 95%

Hypothesis

Rate A

—

Rate B

—

Relative lift

—

z-statistic

—

p-value

—

CI on diff

—

Null distribution Rejection region Observed z

Under the hood: pooled-variance two-proportion z-test. Pooled rate p̂ = (cA + cB) / (nA + nB), SE = √(p̂(1−p̂)(1/nA + 1/nB)), z = (pB − pA) / SE. p-value from the standard normal CDF. Confidence interval uses the unpooled SE. Don't peek at the p-value mid-experiment — you'll inflate Type I error.