A
B
Result
Confidence level
—
Your inputs are stored in this browser, so everything is still here next time. Nothing is sent to a server.
This calculation is for informational purposes only and does not replace advice from a qualified professional. Formulas and rates may not fit your exact situation — double-check the figures before making decisions.
How it is calculated
The calculator uses a two-proportion z-test — the standard statistical method for comparing shares (conversions) between two independent groups. The null hypothesis is that A and B have the same true conversion rate, and any observed difference is just sampling noise.
The confidence level shows how confidently that hypothesis can be rejected: a value of 95% means that if there were truly no difference, a result this extreme (or more extreme) would happen by chance only 5% of the time. More visitors in each group and a bigger conversion gap both push the confidence level up.
95% is a common significance threshold, not a universal law: some fields use a stricter bar (99%), others a looser one. Statistical significance on its own doesn't mean an uplift is worth shipping — a small but reliably measured gain can be worth less than the cost of implementing the change.
One important methodological rule: don't stop a test and check significance every day, reacting to the first moment it crosses the threshold — that sharply inflates the odds of a false positive. Sample size and test duration should be planned ahead of time, before data collection starts, and not cut short early.