Multivariate tests let you try out multiple variables to see how small changes to your marketing can have a big impact on your engagement. Choose what you want to test, like the subject line or content, and compare results to see what resonates the most with your subscribed contacts.
Unlike A/B tests, Multivariate tests can compare more than one variable type to give information on how multiple variables interact with each other.
When we talk about Multivariate tests, we use terminology that's different from how we talk about other tools and tasks in Mailchimp.
Each version of your test is created from your chosen variables. If you want to test 2 From names and 2 Subject lines, we would create 4 different combinations of your test. Combinations sent in the test phase are called test combinations.
- Test phase
The period of time after the combinations are sent out and we compare the results. Data collected during the test phase can be used to determine the test's winning combination automatically or manually.
The element of your marketing that you want to test. Multivariate tests let you experiment with 4 variables: Subject line, From name, Content, and Send time. Each version of the variable is called a variation.
Winner or winning combination
The combination that performs the best. This may be automatically determined by click rate, open rate, or total revenue, or manually chosen based on the reporting data you find the most valuable.
How multivariate tests work
Set up the Multivariate test
You'll choose up to 3 of 4 variables—Subject line, From name, Content, or Send time—and create up to 8 variations. We'll generate all possible combinations and send them to different sets of subscribed contacts, so no one receives more than 1 combination of your test. The combination that your subscribed contacts receive is chosen at random and tracked solely for the purpose of choosing a winner. This means you won't be able to see which combination a subscribed contact receives.
Choose winner criteria
Send the combinations to all your recipients at once if you have a small audience or if you're testing Send time. With other variables or a large audience, send your test combinations to a percentage of your recipients, and send the winning combination to your remaining recipients.
To choose the winner, use 1 of these options.
- Automatic: open rate, click rate, or total revenue
Use these options to send the winning variation to your remaining recipients after a set amount of time. The winner can be determined by the highest open or click rate, or total revenue if your online store is connected to your account.
- Manual: report statistics
Use this option to choose the winner yourself based on reporting data or other factors that you find to be the most valuable.
Variables you can test
Try different phrasing, sales offers, or emoji to see what gets the most attention.
See if your subscribed contacts are more responsive to emails coming from a person's name or from the name of your company or organization. You'll provide the From name and From email address you want to use for each combination.
Create different versions of your content to see what gets a better response. Use this variable to test small content changes or completely different templates. When you test content, you may want to better understand the efficacy of calls to action, links, or buttons. Use our Link Comparison tool in the campaign report to see how your links performed in each combination.
Learn when your subscribed contacts are most likely to open your emails. Since this option tests specific days and times, you must send your combinations to all of your recipients at once because the winning combination can't be sent at a past time. Instead, use this data to inform when to send or schedule future emails.
Multivariate test ideas
Here are some ways that Multivariate tests can be used to compare engagement among your subscribed contacts.
- Does including your company name in your subject line and from name increase engagement?
- Are subscribed contacts more likely to click a linked image or linked text?
- Should you use your personal name as the from name in combination with a subject line that has an emoji, or use your company name with an all-text subject line?
- Will a different template increase click rate, even if text content is the same in all the variations you test?
- Do your subscribed contacts prefer an email that contains a GIF or one with static images?
- Are people more likely to click a link that is just colored text, or one that is styled as a button?