01What’s included

What the experiment includes

Starting fees cover the scope shown for one brand, website and language unless stated otherwise. We agree any additional work and costs before starting.

Format: One-off project
Page experiment setup and analysis
We set up an agreed experiment using supplied page variations and check its measurement.
What you get
An experiment plan, setup checks and a readout showing the evidence, its limits and the recommended next step.
Good to know
Timing depends on eligible traffic and the agreed sample. The fee does not promise a winning version.
from $600/ project
Starting scope
1 A/B or multivariate experiment with up to 4 supplied variations.

02Before you start

Check readiness before choosing a test

Share the page, its traffic and conversion figures, the change you want to test and the available variations. We agree the measure, sample requirements and review rule before launch. Creating variations, fixing tracking and software costs need a separate scope where required.

FAQ

A/B testing questions.

Traffic, page variations and the decisions an experiment can support.

How much traffic does an A/B test need?

There is no single traffic threshold for every test. We consider your current conversion rate, the change you want to detect and how many versions will share the traffic. If the required sample is impractical, we can start with a conversion review instead of promising a quick result.

Start with a conversion audit

Does the price include designing the variations?

The starting experiment offer uses supplied variations, within the limit shown in the price details. We check their readiness and agree what the experiment compares. Writing new copy, designing or developing versions and repairing measurement are separate work when needed. We confirm those dependencies before the experiment starts.

See landing page development

When would you use multivariate testing?

An A/B test compares prepared page versions. A multivariate test examines combinations of changes to several elements, such as a headline and image. More combinations divide the traffic into smaller groups and usually need a larger sample. We choose the method after checking the question, traffic and available variations.

What happens if no version clearly wins?

An inconclusive result is still a result to report. We explain the sample, measurement limits and what the evidence can support. The next step may be to keep the current page, revise the idea or agree further testing. We do not label a small or uncertain difference as a proven gain.

05Next step

Tell us about your project.

Describe what you need help with and which service interests you. You can leave the budget undecided.

Open the project form

What the experiment includes

Page experiment setup and analysis

Check readiness before choosing a test

A/B testing questions.

Related services.

Tell us about your project.