Help & Support
How Can We Help?
We Think These Might Help
Thank you for your feedback!
cannot be blank
We will respond as soon as humanly possible.

How to Set Up an A/B test with Email

AdRoll Email let’s you run A/B tests on your triggered campaigns. Test everything from subject lines to content.



Create a campaign.

  1. Select a goal for your campaign.
  2. Edit campaign to your taste and launch.

Create another (separate) campaign.

  1. Select the same goal as in step 1.
  2. Edit campaign to your taste and launch. Remember to keep the exact same audience as in campaign 1, if you haven’t changed the targeted audience after  .

Helpful tip: Re-name your campaigns so you know which campaign is which.

Now that you have multiple campaigns targeting the same audience but with different content, your AB test is ready to go. AdRoll Email will then automatically split users at random between these campaigns.

How do you make sure customers won’t get multiple emails?

AdRoll Email has internal logic preventing users from receiving multiple emails at the same time, this means only one campaign email will be sent per user per day. Learn more about how we prevent spam.

What can I test?

Everything! As long as the targeted segments are the same, the campaign will be a part of the A/B test.

A few examples of experiments you can run:

  • Test different subject lines.
  • Test different delays. For instance: send one email after 1 hour and another after 24 hours.
  • Test different offers. For instance: one email with a discount code and another without.

How many options can I test?

As many as you want! Repeat the process and you will have another campaign testing another variation of your campaign.

We do recommended that you are conscious about your volume: you need a reasonable amount of emails being sent, opened and clicked to be able to evaluate and compare the differences between campaigns.

The initial recommendation is to test 2 options at a time and then test the winner against another one, so you have a clean picture of what is working best.

Was this article helpful?
0 out of 0 found this helpful