A/B testing

To put it simply, A/B Testing means that two variants of the user interface is shown on your website or inside your app, and it is random which version a user meets. At the end of the test session, we can determine which solution was more effective compared to the other one.

How does it help

The results of A/B Testing are not assumptions: you will get measured changes. While A/B Testing a part of your website or application won't increase the number of users visiting them, but it can help you make more out of your existing traffic.

What will you get

If it was prepared well and ran correctly, A/B Testing is a powerful optimization tool. The little details on the user interface of your product or service can matter a lot. Next to the test results, you will get improved UX, and increased conversion rates.

While this method sounds really easy to conduct, actually there are many things that can go wrong during an A/B Testing process. To run a proper A/B Test, we need for example:

  • hypothesis to test
  • big enough sample size
  • to run the test for full weeks
  • a version identical with the original in order to see the noise (noise here means that if you have two identical UI solutions, and you split the users into two groups, you will still get some differences because of random factors).

What is the process

Our A/B Testing process consists of the following steps:

  • Looking at the Analytics and other usage data
  • Constructing Hypothesis
  • Creating a variation pers hypothesis
  • Running the A/B Test
  • Analyzing the test results
  • Constructing more tests