Stop running underpowered tests. Enter your baseline conversion rate (e.g., 3%), the minimum detectable effect you care to find (e.g., +15%), your daily traffic per variant, and pick significance (α) and power (1–β).
We compute:
Required sample size per variant using a standard two-proportion z-test approximation.
Estimated test duration in days given your daily traffic.
Guidance on when to stop and common pitfalls (peeking, uneven traffic, seasonality).
Use this before spinning up a Play like Creative Testing to keep timelines realistic and stakeholders aligned.
