A/B website testing is a controlled experiment in which at least two different versions of a page are shown to other visitors to find out which version achieves a defined goal more often.


A/B testing consists in creating two or more versions of the website that differ in selected elements (e.g., different CTA size/color, different placement of elements on the product page, etc.).

Web visitors are thus divided between at least two pages: the original A version and the new B version, which are somewhat different. In doing so, the degree of achievement of the desired goal is measured, e.g., purchase, completed contact form, etc.

The version with a statistically significant higher conversion rate is considered more successful and becomes the main version.

A/B testing can be repeated practically endlessly and constantly improve the website’s performance.

For A/B testing, it is recommended that the site under test achieves at least around 1000 conversions per month. Otherwise, it can be time-consuming to ensure statistically significant results.

Read more about A/B testing in Jara Uram’s article, or try A/B testing directly on the web with the help of our specialists.