Experiment stats
Compare two proportions (e.g. conversion rates) with a pooled two-proportion z-test. Enter visitors and conversions for control (A) and variant (B).
Uses a normal approximation to the binomial. Best when expected counts are reasonably large and groups are independent. This is not a substitute for a full experiment design or sequential testing. Typical reporting uses α = 0.05 (two-sided) as a rule of thumb only.
