Cookie preferences
We use cookies for analytics. Privacy Policy You can accept or decline non-essential tracking.
Practical guide to sample size signup test: formulas, workflow, implementation pitfalls, and a direct execution playbook with Sample Size Calculator.
Go to tool
A/B test sample size calculator (Evan Miller formula).
Signup flows have two unusual properties compared to typical landing pages:
Both properties affect sample size calculation in opposite directions.
The sample size formula depends on p * (1 - p), which is the variance of a Bernoulli trial. This variance is maximized at p = 50% (= 0.25) and drops as p moves toward 0% or 100%.
| Baseline conversion | p * (1 - p) | Relative variance |
|---|---|---|
| 5% | 0.0475 | 19% of max |
| 20% | 0.1600 | 64% of max |
| 35% | 0.2275 | 91% of max |
| 50% | 0.2500 | 100% (max) |
At p = 35%, variance is 91% of maximum — so high conversion rates do NOT dramatically reduce sample size. This surprises many people.
Open Sample Size Calculator and enter:
Result: approximately 900 users per variant, 1,800 total.
If your signup page gets 200 visitors/day, this test takes 9 days. Manageable. But if you want to detect a 2 pp lift (35% to 37%), you need ~5,600 per variant = 56 days. Much harder.
If signup traffic is too low for your target MDE:
If your signup is a 3-step wizard (email, details, confirmation), testing step 2 means your effective traffic is only users who complete step 1. Plan for this:
| Step | Drop-off | Effective traffic (of 1,000/day) |
|---|---|---|
| Step 1: email | — | 1,000 |
| Step 2: details | 30% drop | 700 |
| Step 3: confirm | 20% drop | 560 |
Test on step 2: 700 effective users/day. At ~900 needed per variant, the test takes ~2.6 days. But test on step 3 with MDE = 2 pp: 560/day, ~5,600 needed = 20 days.
Open Sample Size Calculator, enter your signup step conversion rate and the MDE you care about, and verify the test duration fits your sprint cycle.
This article is reviewed by the Tools Hub editorial team for factual accuracy, practical relevance, and consistency with current product workflows.
Last reviewed:
Practical guide to mde ab test: formulas, workflow, implementation pitfalls, and a direct execution playbook with Sample Size Calculator.
Model funnel steps in SQL with deterministic stage logic and CTE-based queries so conversion drops are auditable, comparable, and actionable.
Use Landing Page Copy AI to generate high-converting hero sections, feature blocks, social proof, and full-page copy wireframes for any audience.
Use UTM Builder to keep naming stable across ads, email, influencers, and lifecycle campaigns with an AI-assisted governance checklist.