What is the Minimum Detectable Effect?

Minimum Detectable Effect (MDE) is the smallest effect size (change in metric) that can be statistically reliably detected in an A/B test at a given significance level and test power.

In simpler terms, MDE indicates how much a metric (for example, conversion rate) must change for us to confidently say that the change is not due to chance.

Knowing the MDE helps to:

• Determine the required sample size for the test.
• Understand which changes are truly significant for the business.

For example, if the MDE is 5%, it means the test can detect an improvement or deterioration in the metric of at least 5% with the specified statistical confidence.