A/B testing
A/B testing, also called split testing, is an experiment in which two versions of the same page or element are shown to different visitors at the same time. Version A is the control, version B is the variation. After enough traffic, you compare a single metric, such as sign-ups or clicks, to see which version wins.
The value of the method is that it replaces opinions with evidence. Instead of debating which headline is better, you let real visitors decide.
How an A/B test works
A test follows a simple loop. Visitors are split at random between the two versions, each visitor always sees the same version, and the results are recorded against one goal.
- Hypothesis: "If we shorten the form, more visitors will submit it."
- Variation: change one meaningful element, not ten.
- Split: typically 50/50 of the traffic.
- Measurement: one primary KPI defined before the test starts.
- Decision: keep the winner or learn from a flat result.
What to test first
Start where a small gain has the biggest impact: pages with high traffic and a clear goal. A landing page, a pricing page or a checkout step is a better candidate than a rarely visited page.
- Headlines and value propositions
- Call-to-action wording, colour and position
- Form length and field order
- Page layout and social proof
- Pricing display and offers
A/B testing, multivariate and split URL: the differences
| Method | What changes | Traffic needed |
|---|---|---|
| A/B test | One element, two versions | Moderate |
| Multivariate test | Several elements combined | High |
| Split URL test | Two different pages on two URLs | Moderate |
Common A/B testing mistakes
Most failed tests share the same causes. Stopping a test as soon as one version looks ahead produces false winners: wait for the planned sample size and a full business cycle, usually at least two weeks. Testing several changes at once makes it impossible to know what worked. Running a test on a page with too little traffic gives results that are mostly noise.
Also remember that a statistically significant result is not always a meaningful one. Check that the gain matters for your business before rolling it out.
A/B testing at BeBranded
We help teams choose what deserves a test, define the KPI and read the results with the right level of caution. It fits into our consulting service, where we turn analytics and user feedback into prioritised experiments.
FAQ
What is A/B testing?
How long should an A/B test run?
How much traffic do I need?
What is the difference between A/B and multivariate testing?
Does A/B testing affect SEO?
What should I test first?
Linked tools
