Basic Benefits of Split Tests in Marketing Automation - dummies

Basic Benefits of Split Tests in Marketing Automation

By Mathew Sweezey

Split testing, also known as A/B testing as well as multivariate testing, means to conduct a test in marketing automation by holding a factor constant and testing different outcomes from various scenarios. An example is to have a single paid search ad driving people to two separate landing pages. Testing which landing page has a higher conversion from the same advertisement allows you to maximize the effectiveness of your campaign.

Most of the mainstream marketing automation tools allow for split testing, so you should easily be able to perform this type of test inside your application. Consider split testing your emails, landing pages, and content first. After you have mastered the testing process, you can dive into deeper split testing and test smaller items such as email subject lines, dynamic calls to action (CTAs) on landing pages, and so forth.

Split testing goes better if you follow some basics steps.

  1. Define your variable.

    It is suggested you test only one element at a time. This makes for the clearest test and removes the subjectivity of most people’s opinions as to which email is better. Trying to test more than one element at a time is not good practice and doesn’t give you a clear result. Good variables to test are the following:

    • Call to action language

    • Form length

    • Type of content (video vs. white paper)

    • Email subject line

    • Offers

  2. Create the content you want to test.

    For example, to test landing page copy, you need multiple variations of your landing page, with each one having a different version of the variable you want to test. The following shows two versions of the same landing page with different call-to-action language in order to test which language converts best.


  3. Create your test.

    How you create your split test depends on the item you’re testing

  4. Run the test.

    It’s time to go live with multiple scenarios to test which one is the best. It is suggested you test this with emails with a small segment of your total segmentation, with landing pages only for a short while, and with content for a limited time as well.

  5. Review it.

    If you have a clear winner, first determine why and then end your split test. Next, replace the test group with the single landing page, email, or content that had the highest engagement.

Just because one asset has a higher conversion rate doesn’t always mean it’s the best one. If you can, look at the leads that converted during your initial split a few months down the road.

If one of the lower-performing options has converted more leads into opportunities, you may have found what you really were looking for — that is, not just higher conversion rates on an asset, but more deals at the end of the day. Just remember to take a look back and double-check that the higher-converting asset is converting the leads you really want.

You can split test emails a few different ways, mostly depending on the tool you use. Here are the two main ways to do a split test:

  • Inline split test: An inline split test is a split test built into your email-creation process. This means that as you are building your email, you can build a few different options simultaneously, letting you streamline your testing and making it easier to do so. Note that this is a very advanced option, so a tool with this feature usually comes with a higher price tag.

  • List sampling: Sampling has always been a good way to split test. The sampling technique is how Gallup predicts the presidential elections. The sampling method is as follows: Take a few random samples of people from your list, about 10 percent of your total list.

    From the sample, divide the list into the number of variations you’d like to test. Send a separate email to everyone on each small list. Then send whichever email has the highest engagement to the remaining 90 percent.