AB Testing: The Full Guide

Have you been struggling with choosing the best subject line for your emails? the best creative for your ads? the best product picture to display? The solution to all of these problems is AB testing.

The process of improvement takes trial and error because no one can perfect anything from the first try. Be it an email template, a paid ad campaign, ...etc, it requires trying different options and analyzing the results to find the one that serves our purpose the best. And AB testing is the perfect approach for that process.

We will discuss in this article the questions of what is ab testing?, and how does AB testing work? and step-by-step instructions on how is it used in different areas of digital marketing.

What is AB testing?

AB testing, also known as split testing, is the process of trying multiple versions of a certain variable (an email template, an ad creative, a landing page, ...etc), to see which one achieves the best results. 

It is called AB testing as if we are testing version A and version B to see which one is better, but the name is just used to give an idea because AB testing is not limited to testing between 2 versions, it can include an unlimited number of versions to test between. 

How does AB testing work?

First, we choose the variable that we want to choose between multiple versions of (a landing page for example). Then we create multiple versions of that variable and test each version several times. After that, we analyze the results we got from the different versions of that variable so we can choose the one that performed the best. That simple. And many useful tools help us make the technical part of AB testing extremely easy.

AB testing in email marketing.

Email marketers widely depend on AB testing to get the best results from their email campaigns. Improving emails require skills and creativity, but the use of AB testing makes it more of a science that can be learned than it is an art. Now we will see, step by step, how is AB testing used in email marketing.

Step 1: Chose the variable that you want to improve.

AB testing is used in email marketing to perfect one small part of the email at a time, and not the whole email at once. For example, the variable that we want to improve can be the subject line, so we would use the same email but change the version of the subject line with each test.

Choosing the variable in email marketing depends on what metric you want to improve in your email campaign, so if you want to improve the email open rate, you would choose the variable to be the subject line because it plays the most important role in how often to people open the emails (The open rate). Read our full guide on the different metrics in email marketing and what variables correlated with those metrics.

Step 2: Create multiple versions of the variable you chose.

This step is simple but very critical. After choosing the variable to improve, it is now time to decide on the versions that you want to test. For the example of the subject line being the variable, you might decide for one version to be straight to the point and affirmative, while you choose another version to be more mysterious and vague.

It is a good number to have between 3 to 5 versions to test, but you can have as little as 2, or as high as you need.

Step 3: Split the email list.

Now that you have chosen the variable and created different versions of it, it is now time to split your email list for testing. 

You should split the email list into groups that are the same number of versions that you have created for the variable that you want to improve. For example, if you have an email list of 1000 email addresses, and have created 5 versions of the variable, you can split your email list into 5 groups each containing 200 email addresses.

Step 4: Execute.

Now that everything is set. You should assign each group of the email list to a different version of the variable. You send the emails to the group with each email having the version of the variable that corresponds to the group it's sent to.

Step 5: Analyze the results and choose the best version.

After you have executed the test, you should now gather the relevant results for analysis. The results you compare should be metrics that are correlated with the variable that you want to improve.

For example, if the variable is the subject line, you should analyze the email open rate metrics because it is directly correlated with the performance of the subject line, and you shouldn't study the click rate for example (unless the results vary a lot), because the subject line doesn't directly affect that metric. Read our full guide on the different metrics of email marketing campaigns that you should track and what metrics do they correlate with.

Then you simply choose the version of the variable that performs the best to use it in your future email campaigns to enjoy better results.

AB testing for digital marketing in general.

Whatever the field of digital marketing you are in, be it paid add campaigns, SEO, ...etc, the process is very much the same as email marketing. You choose the variable to be improved, create multiple versions of that variable, and test each version to see which one performs best to use it in the future.

But the splitting part can be a little more complex if done by hand. For example, how do you split the viewers in paid ad campaigns? the solution is already done for you by most paid ad campaign providers. In Facebook ads, for example, they let you include multiple versions for each variable of your ad campaign, so you can include multiple creatives (the variable is the creative of the ad campaign), and Facebook will automatically split the viewers into groups, the members of each group see a different version of the creative, and they show you the performance of each creative you have included.

Conclusion.

AB testing is a skill that can be acquired very easily and quickly by anyone, but it still has tremendous benefits on performance.

And considering how useful and how easily learned AB testing is, I find it to be a MUST for every digital marketer to learn and benefit from, as it's implemented the easiest in the field of digital marketing and shows the results the fastest.