حاسبة دلالة اختبار A/B.

Two-proportion z-test for conversion rates; not for revenue per visitor or other averages. Do not stop a test early because it looks significant. Everything is calculated in your browser; nothing you type is uploaded.

z = (p_B − p_A) ÷ √(p(1 − p)(1/n_A + 1/n_B)) with the pooled rate p. The interval uses unpooled standard errors. A chi-square test flags splits that are unlikely under 50/50.

500 conversions from 10,000 visitors against 580 from 10,000 is a +16 % lift with p = 0.0123: significant at 95 %.