A/B test calculator

LAB

Significance, Bayesian odds and sample size

[ Verdict ]

Enter your data
Fill in visitors and conversions above

[ Frequentist Analysis ]

-
p-value
-
z-score
-
relative uplift
-
statistical power
Control
4.50%
Variant
6.20%

[ Bayesian Probability ]

Chance that each variant is the winner (Monte Carlo, 100k simulations)

30%
70%
Control (A) wins Variant (B) wins
-
expected lift
-
risk of choosing B

[ Sample Size Needed ]

To detect the observed difference at your chosen confidence level with 80% power:

-
visitors per variant