Plan and analyze your A/B test
Need help planning your next A/B test? Book a free consultation
Your traffic
Hypothesis
What you can detect
Smallest lift you can detect in 4 weeks
–relative lift
| Test length | Visitors per variant | Smallest detectable lift | As a rate |
|---|
How to read this
Each row shows how long you would run the test and the smallest improvement you could reliably see in that time. If the real improvement is smaller than that number, the test will look like it did nothing. Before you build anything, check that the idea is big enough to clear the row you plan to run.
Your results
Hypothesis
What it means
How to read this
Significant means a gap this big would rarely happen by chance alone. At 95% confidence, chance alone would produce it less than 5% of the time. The interval tells you how big the true lift could plausibly be.
Two-proportion z-test. Multiple variants use a Bonferroni-adjusted alpha so adding arms never makes significance easier.
Try the Tested newsletter for conversion tips
Conversion, messaging and copywriting for B2B SaaS.