Calculate the required sample size for statistically significant A/B test results
Your current conversion rate before the test
Relative improvement to detect (e.g., 10% means 5.0% → 5.5%)
Confidence that results aren't due to chance (95% is standard)
Probability of detecting a real effect (80% is typical)
Two-tailed is recommended unless you only care about improvements
More variations require larger sample sizes (Bonferroni correction)
Load Preset
Saved Scenarios
No saved scenarios yet. Click "Save Current" to save your configuration.
Sample Size Per Variation
31,244
Total Sample Size
62,488
across 2 variations
Expected Effect
5.0% → 5.5%
+0.5% absolute
With 31,244 visitors per variation, you have a 80% chance of detecting a 10% relative improvement (from 5.0% to 5.5%) with 95% statistical significance.
Estimated Duration
7 days
Est. Completion
Thu, Jan 29, 2026
Based on 10,000 visitors/day allocated to the test
Test allocation is the percentage of traffic included in the experiment. 100% means all visitors are part of the test.