Is my A/B test result significant?

Two-sided, two-proportion z-test at 95%, with the relative lift, its confidence interval, and an automatic check on the traffic split.

Control visitors

Control conversions

Variant visitors

Variant conversions

Configured traffic split (for the SRM check)

How many visitors does my A/B test need?

Required visitors per variant at 95% confidence, plus how long that takes at your weekly traffic.

Industry preset

Baseline conversion rate (%) Your current rate, in percent. 4 means 4%.

Minimum detectable effect, relative (%) A 10 means a 10% relative lift (5.0% to 5.5%), not 10 percentage points.

Statistical power 80% 90%

Visitors per week entering the test (optional)

What lift can my traffic detect?

The inverse of the sample size question: with a fixed audience, what is the smallest relative lift you could catch?

Baseline conversion rate (%)

Available visitors per variant

Statistical power 80% 90%

How to calculate revenue per visitor

Total revenue divided by everyone who entered the test, with a 95% interval on each variant and on the difference. No p-value: for a mean, the signal is whether that difference interval excludes zero.

Control visitors

Variant visitors

Control order values One value per line, or separated by commas. No thousands separators: write 1299.00, not 1,299.00.

Variant order values Use one currency across both boxes. Results come back in that same unit, unnamed.

Outlier cap, maximum order value (optional) An order above the cap is excluded outright, from the revenue and from the order count. Leave it empty to keep every order.