[TUTORIAL VIDEO] Getting Started With A/B Split Testing & Multivariate Testing

ab-testing1The term A/B Split Testing is typically used in website development to compare two completely different webpage designs to determine the effectiveness of one over the other. Marketers also use A/B Split testing for advertising, testing two unique ad designs against one another to see which ad produces more purchases or leads. Conducting this test allows marketers to learn what version (A or B) has the most impact and is most effective.  There is usually a control page (original page) that is tested against a variation page with both receiving the same amount of traffic. The benefit of A/B testing is it confirms the success of a new design before a permanent change is made.

In order to create a successful A/B test, a marketer must first determine the goal they want to measure.  If the goal is to bring in more revenue, a new check out page should be created and tested against the original page.  If the goal is time on site, a landing page variation should be tested against the original landing page.

Multivariate testing is a similar concept to A/B testing, but instead of comparing a control page against a variation page, multivariate testing compares several variables at the same time.  For example, a marketer may want to test two different headlines, two different images and two different calls to actions, for over 20 possible combinations. The objective then is to measure the effectiveness each unique design element combination has on the final goal. Multivariate testing becomes a valuable testing technique when marketers want to focus on redesign efforts and make sure they display the elements that have the most impact and highest conversion rate including the images, colors, fonts and text that tested most effective.

Watch our video on A/B Split Testing and Multivariate Testing to learn more.

2 Responses to “[TUTORIAL VIDEO] Getting Started With A/B Split Testing & Multivariate Testing”

  1. Interested

    How many tests does it take to come to your 95% confidence?

  2. Justin Marerick

    We use 3rd party tools, so the confidence interval is always based upon their 3rd party algorithms and what the goal is.



Leave a Reply