Understanding what works and having the data to support those buying decisions are one of the most powerful assets a business owner possesses. Knowing how your marketing strategy performs and how to leverage this information is critical to your bottom line. So what happens when you send something out into the the world you think is a winner and it is unsuccessful? Or, what if you want to try something completely new, but you’re scared to pull the trigger in case you upset your current customers?
It is in these instances that testing different marketing messages becomes the appropriate solution. Testing different variables is one of the most effective ways to really see what your audience responds to and how you can get the best response possible.
But for those who haven’t run a test before, the entire process can seem really daunting. After all, we all originally learn a long, detailed process that includes introductory research, a formal hypothesis, strong results (sometimes with statistical or other mathematical backing), and a firm conclusion. That’s a lot of work to wrap your brain around!
Fear no more: this post will help you understand how you can begin running simple tests on your marketing efforts.
A/B testing is one of the most basic and often-used methods of testing by marketers, but it’s also the most simple to understand. At its core, A/B testing is creating two versions of the same marketing campaign but changing one factor.
Because A/B tests isolate one single factor (the subject line, an email design, or paragraph copy within an email, for example), the version with the highest conversion rate wins. And you know exactly why it wins, which can help you replicate success in the future.
A/B tests are also very simply to run, making them worthwhile time investments especially for informing your email marketing strategy. They will also help you learn about your audience and what it reacts best to.
Identifying Variables for your A/B Test:
If you’re a newbie to testing, begin with a single factor you want to test. Maybe it’s a subject line or perhaps you want to test how a different call to action performs (button vs. text link). Remember to start small; you can always run follow-up tests on a second variable.
Next, set up your “test” and your “control.” What exactly does this mean? In essence, the control is the message you are testing against. So perhaps your strategy is to create an email campaign with a very standard, informative subject line. You want to test this against a more fun subject line that references something from pop-culture. The basic subject line email would be your control, and the email with the pop-culture subject line would be the test.
Once your two segments are created you will want to create the test groups. To do this, split your subscriber list completely and randomly. It’s also important to determine your sample size (if applicable). One common test is a straight 50/50 split. This means that 50 percent of your list will receive the control and the remaining 50 percent will receive the test.
In no way are 50/50 list splits your only option. You could test a smaller segment of your list and then, after a set period of time, the winning email can be sent to your remaining email subscribers.
As an example, you could send 15 percent of your list the control and another 15 percent of your list the test. You can then send the email with the higher conversion rate to the remaining 70 percent of your email list, allowing you to capitalize on the higher conversion rate and capture that at the time of the test. Not all programs allow this, but it is certainly one way to instantly improve your results.
Wrapping up the test:
You’ve created your two segments and you’ve sent those segments to your list. So now the golden question: how long do you test? Well, that answer varies dramatically depending on what you’re testing. If it’s an email, you can often see results in minutes. For landing pages, the results may take several weeks.
One golden rule is that no test should last no longer than two months. Of course, if you test an email, the time requirement is as little as a few hours. In all instances be sure to clearly set dates and track your website traffic during period.
When the time comes to end your test, track your results using the reporting features available to you. If you are running a test on a landing page, it may be important to consider any seasonal factors that could play a role.
More Complex Testing
Once you’ve completed your first test, it’s pretty common to want to keep testing. And that’s totally OK! Simply start the process over with a new variable.
Don’t feel limited to two options, either; you can easily perform an A/B/C test and try three different elements in your email. Or an A/B/C/D test to test four elements. The possibilities are endless.
For the purpose of simplicity, I’ve focused on an A/B test with only two versions that isolates a single variable. But do know that there is something to be said for testing multiple variables, or completely different designs, at once. There is also a second type of testing called multivariate testing, which looks at multiple variables. This helpful article explores multivariate testing if you’d like to learn more.
What other questions do you have about A/B testing and how it can improve your digital marketing strategy? What is the first A/B test you’ll run for your company? Leave your questions and thoughts below!