Tools

MDE Calculator for A/B Tests

In Online Experimentation, you don't choose your traffic. You work with what you have. This calculator tells you the smallest effect your test can detect given your site's traffic, so you can set honest expectations before you run a single experiment.

Two-tailed hypothesis. Detects effects in both directions. Use this as the default unless you have a directional hypothesis.

Statistical & Business Parameters

Significance level (α) *

Probability of a false positive. Typically 0.05.

Type II error (β) *

0.20 = 80% power. Lower beta = more sensitive test.

Baseline conversion rate (%) *

The proportion of visitors who convert today, before any experiment.

Test Design

Week 1 unique visitors *

Total unique visitors entering the test in the first week.

Traffic split *

0.5 = 50/50 split, 0.33 = A/B/n three-way

Max weeks *

Between 1 and 52.

New unique visitors per week (±) *

0 = flat · +1000 = growing · -500 = declining

Common scenarios

Scenario	Traffic	CR	Split	Context
Stable site	2,000	5%	0.5	Consistent traffic, no campaigns planned.
Campaign incoming	5,000	3%	0.5	Low base, big campaign launching from week 2 onwards.
SaaS launch spike	12,000	8%	0.5	Launch spike in week 1, new signups decline week over week.
High-traffic e-commerce	50,000	3%	0.5	Large checkout funnel, stable traffic year-round.
A/B/C three-way test	8,000	5%	0.33	Three variants, traffic split in thirds.