What Is Split Testing and How Do I Do It?
Split testing is also called A/B testing. It refers to a controlled experiment that helps you determine whether your stated goals and the actions you choose to reach them are useful.
Phew! It’s just testing one thing against the other to find out which one works more.
Split testing can help you increase website traffic, get higher conversion rates for all your CTAs, and even lower your bounce rate and cart abandonment. You can do this easily when you know how, and you’ll be glad to do it because it will improve your results.
Determine What You Want to Test![split testing]()
Test only one variable at a time. For example, if your goal is to improve your conversions, you may want to work on trying out different calls to action words or even CTA locations.
Only do one thing at a time.
If you’re going to test the CTA, keep each the same except for the one thing you want to test.
Know What Your Goal Is
It helps you to choose your goals before you get started because the goal informs the test. Some potential goals may be, get more traffic, boost conversion rates, reduce bouncing, stop cart abandonment, get more newsletter sign-ups. Any goal you want to reach can be the basis of an A/B test.
Create Your Test
One test is the control, and the other is the challenger. If you are making changes to an existing site, then that change is the challenge, and the original is the control. If you are setting it up brand new to test something, you can call one A and one B or one the control and one the challenger whichever you want to do.
Example: Let’s say you want to test your CTA button size and how that affects your audience’s willingness to convert. You’d set up two sales pages that look the same except that one has a bigger button than the other. Keep the words the same, the colors the same, and everything the same except the size. Each page you’re testing needs to look the same other than the factor you’re testing for.
Split Sample Groups Equally and Randomly
This is where Google Analytics technology and other paid technology that helps you with split testing can come in handy, but you can do this manually too. You want the new page to appear randomly on an equal basis to anyone who clicks through to your site and that landing page through your advertisements or promotions.
Ensure You Have a Good Sample Size
This is a problem if you’re new to websites, and you don’t have a lot of traffic yet. If your traffic is still low stop doing this and focus on traffic building. When you have enough traffic to get a good sample size, then you’ll be able to use A/B testing more effectively. Most systems will require at least 1000 hits before you get enough data. Plus, you really need to get that information within a couple of weeks for the test to be worthwhile.
Conduct the Test![conduct split testing]()
Once you have everything in place, conduct the test. Remember to check the metrics before, during, and after the test period, which should be short to be more accurate. The longer you conduct a test, the less accurate the results will be statistically.
Test only one thing at a time to ensure the best results. Focus on one metric at a time for improvement so that you can take the time to immerse yourself in the research about that one issue. For example, learning as much as you can about writing effective calls to action is a critical component of improving your CTA results regardless of conducting testing or not.
