1. Go to Content > Calls-to-Action
From the Dashboard, go to Content > Calls-to-Action.
2. Create an A/B variation test
In the CTA dashboard, click Actions > Create multivariate test next to the CTA you want to test.
3. Edit your variation
In the CTA editor, make the changes that you would like to the variation design. Remember to only change one variable between versions so that you can more easily determine which factor caused more visitors to respond. For a refresher on CTA best practices, review this guide. Once you are done editing, click Next. In the Options screen click Save, then click Finish or Add another variation if you want to test more than 2 versions of this CTA.
You will now see the original CTA labeled as a Multivariate in the dashboard. Click Preview to open up a window on the right side of the screen where you can view all variations of this CTA.
4. Measure the results of your test
Make sure that you let your test run long enough in order to obtain a substantial sample amount of views. At least 100 views each usually makes for an accurate test, especially if you only changed one variable. You can view the results of your test in the Calls-to-Action detail screen. From the Calls-to-Action dashboard, click Actions > View details.
You can see the analytics for your A/B variations on the details screen. You can toggle between Clicks, Views, and Submissions using the dropdown menu in the upper-left corner. Click the down arrow to the left of any CTA to see a detailed breakdown of that variation's performance on specific pieces of HubSpot content, such as landing pages and emails.
5. Take action based on your results
Now that you have determined which CTA variation performs the best, complete your test by disabling the variation that received the lowest Views > Clicks percentage. You can disable a variation in the CTA dashboard by clicking Actions > Edit. Hover over the variation you'd like to disable and click - Turn off. Then click Finish.
To fully remove the variation, click Delete, then click Finish.