What is an A/B test?

A/B testing is a technique you use to compare two different versions (of a landing page, for example) by testing which version delivers the best results.

For example, you create a new landing page that is slightly different from the original page. Then you test both pages for a certain amount of time. In the end, you see which landing page generated the most conversions. The page with the most conversions won the test.

A/B test landing page in HubSpot

With the Pro or Enterprise version of HubSpot, you are able to run very clear A/B tests in an easy way.

By clicking 'Create A/B test' when creating a landing page, you create a second variation of a page. In the variation you make a small change (for example, a change in the image, text, or the button of the form). Then you analyze the statistics of both variations.The version with the highest conversion rate (submissions / views) is the winner of the test.

Click here to learn more about how to run an A/B test with HubSpot.

A/B test call to action (CTA) in HubSpot

In addition to a landing page, it also makes perfect sense to A/B test call to actions (CTA). The more effective a CTA is, the more visitors the CTA brings to landing pages. With a call to action, you don't measure by conversion rate, but by click through rate (CTR).

Setting up an A/B test for a call to action is very easy with HubSpot. You first create a normal CTA. Once you've saved it, you'll see the "Multi variate test" button. Press this button and HubSpot allows you to create a B variant of the CTA. Your A/B test is now complete! You can create even more variants than an A and a B variant. In that case, it is no longer an A/B test, but a multivariate test.

Tip 1: if the page on which you A/B test the call to action has relatively few visitors, do not opt for a multivariate test, but stick to a simple A/B test. The test results will then be more reliable.

Tip 2: Always make only one change per test. So for example, change the copy, an image, or the color of a button, but don't change several things at once. Then you won't know why one CTA, or landing page performs better than another.

Still have questions about A/B testing after this article? Ask your question in the live chat at the bottom right of the screen or let us know on LinkedIn.