Multivariate testing is important in order for you to find out what works well, and what doesn’t. But how do you do it? Read along to find out!
If you haven’t tested your email campaigns yet you might want to start with A/B testing. It’s an essential first step where you test two or more variants to determine which variation performs better for a given conversion goal. Once you’ve gotten the hang of A/B testing and have done it a few times, it might be time to spice things up further
Speaking of spices, you might love to add spices to your cooking. But even if you love cinnamon, cayenne and oregano, they don’t necessarily go together. You have to find the ones that do to make your dish a success, right? And how do you find out? By testing! The very same goes for your email marketing.
Read more: What is retail analytics?
Think of multivariate testing like mixing your spices and trying which work best together. In A/B testing, you probably started with changing one thing, such as a subject line, and let the rest remain constant. In multivariate testing, you add two or three variables and test which combination works best. Sounds simple, doesn’t it?
We recommend starting with two or three variables to change. Remember, it gets time-consuming to analyze the results the more variables you change, but the results can be worth it!
Here’s an example of how you can perform multivariate testing:
- Button color: Red vs. Green
- Position of a button: Under the header image vs. At the bottom of the email
These two variables will produce FOUR different email versions (we told you it’d be time-consuming!)
- 1: Red button – Under the header image
- 2: Red button – At the bottom of the email
- 3: Green button -Under the header image
- 4: Green button – At the bottom of the email
Create sample groups with the same amount of contacts in each (e.g. 5000) and send one version to each group.
After you’ve sent them, analyzing is really simple, which button got the highest percentage of clicks? Send that email to the rest of your contacts and use it going forward. Time to test something else!