A/B Testing

What is A/B testing?

A/B testing (or split-run testing) is a user experience research methodology of comparing two versions of a webpage or app (their components) against each other to decide which one performs better.

In a nutshell, AB testing is an experiment when two or more variants of the same entity are displayed to the audience to gather statistical data. This data defines the winning variant and it goes live.

How A/B testing works

Let’s take the website page A/B testing for example.

Using special tools or services you create two versions of the same page with one difference and run the test. The algorithm splits the incoming traffic (usually it’s 50/50) and shows one variant to one audience share and another one to the other part.

The visitors perform actions and the system collects data. The page variant that lets visitors complete the goal more times (or has a higher conversion) wins.
When you have one component optimized you can test another one.

Why should I do A/B testing?

To optimize and improve the performance of digital assets. To raise conversion, to make visitors reach the goals you want them to, to earn more, and decrease the spendings.

Web elements perform differently for different audiences, messages, products, etc. You need to test yours to find those that work best for your case.

Things you can test with A/B testing

• Buttons — test the color, size, positioning, fonts, etc.
• Visuals — use faces, abstractions, change fonts, text positioning, or size, test the image size.
• Headlines — experiment with wording, punctuation, or capitalization, use emojis.
• Call to action (CTA) — make it short or long, change the person (1st, 2nd).
• Page content (or structure) — test the layout, the types of blocks, the colors, etc.
• Design — experiment with the colors, typography, and the customer journey, test the navigation, etc.