MDE Calculator
The minimum detectable effect (MDE) is the smallest A/B test uplift you can reliably detect from your traffic. To calculate MDE, use: MDE = (Zα/2 + Zβ) × √(2p(1−p)/n) ÷ p, where p is your baseline conversion rate and n is sample size per variant.
Example: at a 3.5% baseline CVR with 70,000 visitors per variant (95% confidence, 80% power), the minimum detectable effect is about 6.4% relative — so any real uplift smaller than that will read as inconclusive. Know this before you run a test, not after.
Or calculate from traffic + duration:
Enter your sample size and baseline CVR to see your test sensitivity.
What is MDE and why does it matter?
MDE (Minimum Detectable Effect) tells you the smallest improvement your A/B test can reliably detect given your sample size, baseline CVR, and statistical requirements. It's the "sensitivity" of your test.
A high MDE means your test is "blunt" – it can only detect large changes, and smaller real improvements will appear inconclusive. For low-traffic sites, accepting a higher MDE (20–30%) is often necessary. For high-traffic pages, you can detect smaller changes (5–10%) reliably.
If your MDE is higher than the improvement you expect from your hypothesis, your test is underpowered before it starts. Either increase test duration, focus on a higher-traffic page, or recalibrate your hypothesis.
MDE is the flip side of sample size: fix the effect you want to detect and use our sample size calculator to find the traffic you need, or our A/B test duration calculator to see how many days that takes. For more on planning underpowered tests, read how much traffic you need for A/B testing and A/B testing statistical significance explained.
Not enough traffic to test everything?
We prioritize the highest-ROI experiments for your actual traffic volume – so you stop running underpowered tests that waste months.