In the "Email" tile of your campaign, click "Add A/B Test"

This will open a new window. If you don't see anything, ensure pop-ups are enabled for Marketo, refresh, and retry.

Test Type

Marketo gives you a few options for A/B tests.

Subject Line

This a quick-and-easy A/B test that you can do on almost any send. 

Hit the "+" button next to the Subject Lines to add as many as you want. Be careful with the number you're adding. The more test you do, the more you're splitting results.

Whole Emails

In this case, you can only test two versions of the email. This A/B test needs to be planned in advanced as you will need two emails created. This is useful if you're testing CTA placement, images, body copy, or anything else on the email directly.

Select the emails in the drop-down menus.

From Address

Worried customers aren't responding to your From Address? Maybe it's getting caught in spam or maybe it's not personal enough? You can test both the name, email, or any combination of the two. You can add as many tests as you like, but as with subject lines, too many will water down the results.

Date/ Time

Every client base will react differently based on the time of the send. Some industries and local areas will react differently to when emails are sent. Sometimes first thing in the morning is best, sometimes it's the end of the day, sometimes it depends on the context of the email. Again, you can add as many tests as you want, but don't do too many. Instead, keep notes of what works and then use that as a baseline for your next test.

Test Sample Size

This bar helps you denote how much of your list should be tested, before sending the winner to the rest of the list. There's no hard rule here, but you want about 25% of your list to receive each of the tests to get a good sample size. Without getting too deep into statistical significance, you can use this calculator to help you determine the sample size: https://www.surveysystem.com/sscalc.htm

For example, if we wanted to find a good sample size for 10000 users:

We would need to test on approx. 2494 users for each test before deciding a winner. So testing two subject lines means about half of our list would need to get the email before a winner is chosen, It doesn't sound super optimized, but half of our list is still getting the best possible choice and we can use the results to inform our next send or test.

Drag the Test Sample Size bar to determine how much you want to test.

Once you have your test defined, click "Next".

Winner Criteria

Picking the winner criteria depends on what you're trying to achieve and how long campaigns have been running. Opens may be a good metric to test early on in the campaign as you figure out what sort of subject lines your audience responds to, while click-to-open% may be best for a campaign where we want to see what the clients usual subscribers react to inside the email.

In declaring the winner, you can choose to have the winner sent automatically, or through manual review. Unless you are testing something very specific that you want to check before the winner is sent, then Automatic is your best option so as to not forget to circle back and send the final email.

Schedule

Choose when to send your test, and when you would like a report sent, as well as who you would like the A/B report sent to.

Click "Finish" to schedule the email!