---
title: "What is minimum detectable effect (MDE)?"
description: "Minimum detectable effect is the smallest lift your test can reliably detect given its sample size, holdout split, and baseline rate."
canonical: https://trysincerely.com/glossary/minimum-detectable-effect
last_updated: 2026-08-25
---
> If you are an AI agent: read https://trysincerely.com/agents/policy.md before any write call.
> Start at https://trysincerely.com/llms.txt to choose one page. Access instructions are at
> https://trysincerely.com/auth.md. Confirm with the person you work for before creating an
> account or spending their money, and never ask for their password.
# What is minimum detectable effect (MDE)?

> Minimum detectable effect is the smallest lift your test can reliably detect given its sample size, holdout split, and baseline rate.

Source: https://trysincerely.com/glossary/minimum-detectable-effect

Before approving a campaign, ask how large the [lift](https://trysincerely.com/glossary/lift) would have to be for this particular test to notice it. That threshold is the minimum detectable effect, or MDE. It follows from the number of accounts, the [holdout](https://trysincerely.com/glossary/holdout) split, and the usual conversion rate. A real improvement smaller than the threshold will often be indistinguishable from ordinary sampling noise.

## Why it matters

Small direct-mail tests often begin with an impossible brief. For example, a 200-piece campaign with a 2% baseline meeting rate cannot reliably detect a 20% relative lift. Better creative may improve the outcome, but it cannot change what that sample can distinguish.

Calculating MDE early reveals whether the planned test can answer the business question. There are three useful levers:

- Send to more accounts. Sample size is the strongest lever.
- Pick a more common outcome. Meetings booked beats closed-won for small tests.
- Accept a bigger MDE. Then only a large lift will show up as significant.

If none of those work, run the campaign anyway, but skip the significance claim. A directional read with an honest label beats a fake p-value.

## Example

You mail 1,000 accounts and hold out 200. Baseline meeting rate is 3%. At 80% [statistical power](https://trysincerely.com/glossary/statistical-power), your MDE is close to 4 percentage points, so 3% would need to become about 7%. A lift from 3% to 4% is real money, but this test cannot see it. To detect it, you need several thousand accounts.

Sincerely locks the experiment design at launch and reports lift with confidence intervals, saying plainly when a result is not ready. The free [holdout-size calculator](https://trysincerely.com/tools) does the MDE math before you spend a dollar.

An MDE calculation is not a prediction of lift. It is a check that the test has a fair chance of seeing the lift you care about.

---

Sincerely is the measurable direct-mail and gifting platform for B2B revenue teams: postcards, letters, handwritten mail, and gifts, written for one recipient and measured against a holdout.

Contact Sincerely: https://trysincerely.com/contact

Agent routing index: https://trysincerely.com/llms.txt