Minimum Detectable Effect (MDE) Calculator
The smallest conversion-rate lift an A/B test can reliably detect from baseline rate, visitors per variant, alpha and power — absolute and relative.
How it works
MDE = (z₁₋α/s + z_power) · √[2 p(1−p) ÷ n]; relative MDE = MDE ÷ p
Before running an A/B test, the MDE tells you which effects are even detectable with the traffic you have. Small baselines and small samples push it up quickly: a 4% conversion rate with 20,000 visitors per arm can only see lifts of roughly 14% relative. If the lift you realistically expect is smaller than the MDE, the test will most likely end inconclusive, however long it feels.
Worked example
Baseline 4%, 20,000 per variant, α = 5% two-sided, 80% power: MDE = 0.55 points (13.7% relative). Seeing a 5% relative lift would need about 150,000 visitors per variant.
Frequently asked questions
Should I use one- or two-sided?
Two-sided is the safe default; one-sided is defensible only if you would never act on a decrease.
Does this account for peeking?
No — fixed-horizon tests assume one look at the end. Sequential methods need larger samples.
Related calculators
For education — check assumptions (sample size, distribution, independence) before using a test result in a real decision.