Compare two conversion rates with a two-sided test and confidence interval.
No matching tools.
Uses a pooled two-proportion normal z-test at a fixed 5% significance level and an unpooled 95% interval for the absolute difference. Small success or failure counts trigger an insufficient-sample message. Assume independent randomized visitors, one conversion per visitor and a planned stopping point; repeated peeking or multiple comparisons are not corrected.
50 conversions from 1000 visitors is 5%; 80 from 1000 is 8%. The difference is 3 percentage points; the two-sided p-value is about 0.0065.