# Minimum Detectable Effect (MDE)

> The smallest change a test is big enough to notice.

**Field:** General Knowledge  
**Category:** Math & Statistics  
**Also known as:** MDE, Minimum Detectable Lift, Smallest Detectable Effect

## Think of it like

It is like a kitchen scale that does not move for a crumb. You need to add a big spoon before it shows.

## Definition

The minimum detectable effect is the smallest change a test is set up to catch. You pick it before the test starts. A smaller effect needs far more people, because halving it takes about four times the sample. If the real change is smaller than the MDE, the test will usually miss it. So the MDE decides how long a test must run.

## Example

A page converts at 3%. To catch a rise to 3.6% with the usual settings, a test needs about 14,000 visitors in each version.

## Why it matters

It turns a hope into a sample size and a run time. Without it, people stop tests too early.

## Common mistake

Reading a test with no winner as proof the change did nothing. The test may have been too small to see it.

## Three principles

- **Set it first.** Choose it before the test runs.
- **Small costs more.** Half the effect needs four times the people.
- **Sets the run time.** Sample size over daily traffic.

## Related terms

- [Statistical Significance](https://learn.benibix.com/term/general/statistical-significance)
- [Sample (Statistics)](https://learn.benibix.com/term/general/sample)
- [Hypothesis Test](https://learn.benibix.com/term/general/hypothesis-test)
- [P-Value](https://learn.benibix.com/term/general/p-value)
- [A/B Test](https://learn.benibix.com/term/marketing/ab-test)

![Three bars for a page that converts at 3%. To catch a lift to 4.5%, about 2,500 visitors per version. To catch a lift to 3.6%, about 14,000. To catch a lift to 3.3%, about 53,000. The smaller the change, the longer the bar.](https://learn.benibix.com/diagrams/mkt-cro-sample-size.svg)

---

Source: https://learn.benibix.com/term/general/minimum-detectable-effect
Author: Behnia Zarei (a developer in Vienna, Austria, who builds and teaches data, AI and business intelligence at BENIBIX)
Publisher: BENIBIX Learn (https://learn.benibix.com/)
Last updated: 2026-11-20
Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Quote freely with attribution.
