Find out whether your A/B test actually won
Enter your results to see whether the difference is real, or plan how many visitors your next test needs. The math runs in this browser and nothing is uploaded.
Check a finished test
Use visitors (or emails delivered, or ad impressions) for each variant, and the number that took the action you’re testing. A is your original and B is the change.
Probability is the chance of seeing a gap this big if A and B really perform the same. A reading under 0.05 (5%) is significant at 95% confidence.
Plan a new test
The lift is relative. A 3% rate and a 20% lift means you’re trying to detect a move from 3.00% to 3.60%. Power is the chance the test catches a real lift of that size.
| Lift to detect | New rate | Visitors per variant |
|---|
Smaller lifts need far more traffic. If the first row is more than your site can deliver in 4 to 6 weeks, test a bigger change.
Want a second set of eyes on your test plan or your results? Let’s look at your numbers together.
Book a working session