Statistical significance calculator — sample size & power
Work out how many visitors per variant you need before you can call an A/B test. Set your baseline CVR, minimum detectable effect, confidence and statistical power.
Two-tailed z-test for comparing two proportions. Assumes equal traffic split.
Sample size, MDE and power — explained
The smaller the lift you want to detect, the more traffic you need — sample size scales roughly with 1/lift². Most teams over-estimate the lift a test will produce, then call the test early. Set a realistic MDE (5–15% relative) and stick to your pre-calculated sample size.
When to use this vs. the significance checker
Use this calculator before a test to plan sample size. Use the A/B test significance calculator after a test to see whether the observed difference is real.
Ship more winning experiments
Thrivio surfaces the highest-impact experiments to run, drafts the brief and copy variants, and tracks results without spreadsheets.
Ready to unlock your next stage of growth?
Join growth teams who use Thrivio to make smarter decisions and grow faster.