Bayesian A/B test calculator: probability B beats A, credible intervals, expected uplift and risk.

Control (A)

Visitors Conversions

Variant (B)

Visitors Conversions

Credible level

Probability B beats A

92.4%

B is likely better, but not yet conclusive.

Control (A) Variant (B)

Conversion rate — A (control)

5.09%

95% credible interval — A

3.82% – 6.53%

Conversion rate — B (variant)

6.59%

95% credible interval — B

5.14% – 8.20%

P(B > A)

92.4%

P(A > B)

7.6%

Expected uplift (B vs A)

+29.41% relative · +1.50% absolute

Expected loss if you ship B

0.04%