A/B Test Significance Calculator
Enter visitor and conversion counts for control and test variants and evaluate a two-sided significance result locally.
A QUICK WALKTHROUGH
How to use this tool
- Enter visitors and conversions for both variants.
- Choose 90%, 95% or 99% confidence.
- Review rates, relative uplift, p-value and verdict.
Statistical method
This uses a two-sided pooled two-proportion z-test for two independent groups. It is a planning aid, not a guarantee of causal validity or experiment quality.
Interpretation
A smaller p-value is stronger evidence against equal conversion rates. The confidence choice changes only the verdict threshold; it does not change the underlying data.
GOOD TO KNOW
Common questions
What test is used?
A pooled two-proportion z-test.
What confidence levels are available?
90%, 95% and 99%.
Does significance prove the change caused the result?
No. Randomization, sample quality, stopping rules and practical impact still matter.