Visitor sample size for detecting a difference

Experiment owners and analysts can estimate the required audience before launching a test intended to detect the target difference.

Plan your test by calculating how many visitors are needed to detect the target difference, creating a measurable basis for the comparison.

How to use

Enter the baseline conversion rate as a percentage.
Type the minimum difference you want to detect.
Run the calculation to get the required sample size.
Use that number to plan the test duration and audience volume.
Review your assumptions before starting the campaign.

Example

With an input such as 5 and 2 on two lines, the tool estimates how many visitors are needed to detect a two-point difference around a five percent baseline. The result helps you decide whether the available traffic is enough before launching the test.

Input

5
2

Limits and formats

The calculation depends on the values you enter and on statistical assumptions; a real test may need a different volume.
A smaller difference usually requires many more visitors, which can greatly extend the test.
Result quality also depends on segmentation, conversion tracking and the absence of outside effects.
Do not confuse sample size with duration: duration depends on the traffic actually eligible.

Where processing happens

Everything happens in your browser. The rates and difference you enter stay on your device and are not sent to any external service. You can test several scenarios without an account or data retention.

Frequently asked questions

Why do I need a sample size?

It helps you see whether your test will have enough visitors to detect the target difference. Without enough volume, a real difference may go unnoticed. The calculation is therefore a planning step before you spend time or traffic.

What does minimum detectable difference mean?

It is the smallest difference you want to be able to spot between two variants. The smaller this difference, the more data the test requires. Choose a realistic value based on the expected impact and the decision you need to make.

Can I run the test with fewer visitors?

You can run it, but the result may be less reliable for detecting the target difference. It is often better to extend collection, revise the goal or accept greater uncertainty. The calculated size remains an estimate to adapt to your context.