facebook
 

Knowledge Center

How to Perform A/B Testing?

For the set goals achievement and successful A/B testing fulfillment, first of all, one should get acquainted with the peculiarities of such checking type and the algorithm of its performance.

Algorithm of the A/B Tests Creation:

  1. Learning the website data.
  2. Monitoring the users’ behavior.
  3. Formulating the hypothesis.
  4. Testing the hypothesis.
  5. Analyzing the test data and formulating the conclusions.
  6. Results reporting.

For the website data analysis, one can use diverse web analytics tools (Google Analytics) in order to define the domain of problems in the conversion funnel. For example, one may set pages with the highest indicator of the so-called “unnecessary pageview” - the user follows the add link but does not perform any actions after and immediately closes it.

In order to analyze the users’ behavior, one may also utilize such tools as Visitor Recordings, Heatmaps, On-page Surveys, and others. They will help to reveal what prevents the conversion enhancing.

What may be changed and how - these are hypotheses.  For instance, will it be better to move the star-shaped button from the very down corner of the page? The page with displaced button is developed and the previous one remains. The analysis of conversion rate results is conducted - this is exactly hypothesis testing and processing the results.

Further, the received and analyzed outcomes should be handed over marketing department, IT, and other specialists for the further work. Also, user interface testing plays an important role in this case.  But usability testing should not be neglected.

Thus, the conversion will grow and the website will be more successful. Web application testing requires the checking of a huge number of system aspects: stress checking, performance control, load checking, functional testing, and etc.