Your traffic constraints
Assumes a 50/50 A/B split, two-sided 95% confidence and 80% statistical power.
What your traffic can detect
-
If you can wait longer...
| Test length | Visitors / variant | Detectable relative lift | Approx. target CVR |
|---|
If you want to detect a specific lift...
| Relative lift | Sample / variant | Total visitors | Estimated days |
|---|
How to use this decision
If the smallest detectable lift is very large, traditional fixed-horizon A/B testing may be impractical for your traffic. That does not mean “do nothing.” It means you should choose higher-impact changes, aggregate more traffic, extend the test window, use a higher-volume KPI, or rely on repeated pre/post experiments with careful context tracking.
This is a planning estimate, not a guarantee. Seasonality, traffic quality, repeated peeking, uneven allocation, novelty effects and measurement errors can all affect a real experiment.
Track the experiments you can actually run
Experiment Impact Lab lets you store up to 100 experiments locally, compare pre/post conversion rates, calculate p-values, export CSV, and generate an AI-ready review prompt.
$12 one-time download. No login, subscription, API key or cloud upload.
Get Experiment Impact Lab — $12