What is AB test and multivariate testing

Multivariate testing

In addition to A / B testing, multivariate testing is another method for optimizing the website with a view to maximizing the conversion rate. The aim is to find the most user-friendly version of a website possible. While A / B testing only compares two different website variants, multivariate testing compares numerous modifications to the website.

How it works [edit]

Multivariate testing looks for reasons for dissatisfaction with the user, reveals weaknesses of the website and determines from a large number of slightly modified websites the one with the highest degree of user-friendliness. The test is carried out by splitting the incoming traffic on the modified pages. The test pages can contain different texts, use different colors, or have different font sizes. After evaluating the test results, it is clear with which variant the greatest success can be achieved. Finding the easiest page to use is important because you don't want users to just click the page for a moment. The website operator wants to increase the length of stay, turn the visitor into a buyer and thus increase the conversion rate.

Limitation of test pages by incoming traffic [edit]

With multivariate testing, there is basically no limit to the number of test pages that can be compared. However, it should be ensured that a result becomes less meaningful as soon as the number of test persons per test page is insufficient. Extending the duration of the testing proves to be unfavorable in most cases, as test results are no longer up-to-date due to the fast pace of online markets. In order for the evaluation to be representative, a realistic number of website variants should be determined based on the incoming traffic before the start of the test.

Differentiation between multivariate testing and A / B testing

Since these two analysis methods hardly differ in their areas of application, webmasters are often faced with the question of which method should be preferred. Since in most cases an optimization consists of several successive steps, it is not uncommon for a combination of both optimization methods. A / B testing is used to weigh up major changes. The results here are clear, delivered quickly, and involve less effort. A particularly high traffic is not necessary for this test start. If a decision was made regarding the basic structure of a website with the help of A / B testing, detailed questions such as the color, size and position of buttons can now be clarified using multivariate testing. As the number of test pages increases, so does the evaluation effort. A correspondingly high traffic is the prerequisite.

Web links [edit]