A/B TESTING & PERSONALIZATION

A/B testing generally tests the following elements: Calls-to-action, headlines, images, text length, forms, menu bars. Our experienced experts will be happy to support you in planning and organizing the testing processes.

In addition to websites, A/B testing can also be performed for emails, PPC ads, and CTA buttons.

A/B testing is actually used to examine hypotheses. There are two typical concepts in hypothesis testing: the null hypothesis and the alternative hypothesis. Usually, the null hypothesis indicates that the performance of the two variants A and B are identical, while the alternative hypothesis states that they are not.

There are different opinions regarding this matter, however we recommend continuous testing. You should have a clear goal and enough page visitors, in order to achieve statistical relevance within an acceptable period of time.

The first thing that differentiates your A/B testing requirements is client-side or server-side A/B testing. This aspect is often overlooked. Nevertheless, it should be chosen based on your needs.

• Client-side: commonly used to optimize conversion rates in marketing or funnel, for example by creating page variations directly on the users’ browser.
• Server-side: when you need to test more in depth in relation to the visual changes, such as products (features) or experience for engagement, retention and more.

Many mistakenly think that A/B tests could have a negative impact on SEO. The truth is, that websites rather improve through A/B tests which results in better ranking.

The wrong interpretation of statistical significance is one of the most frequent and serious mistakes committed in A/B testing. Usually the minimum required traffic is calculated using the following key figures:

• The conversion rate of our control variation (variation A)
• Minimum difference between the conversion values of variations
• Confidence level
• Statistical “Power”