Split testing is often called A/B testing and it’s when you test two slightly different versions of something, one against the other. For example, you may test one buy button against another buy button, leaving everything else on the sales page the same. There are some best practices for split testing that you should be aware of as you embark on this business changing practice.
Know Your Goal
Be very specific about your goals so that it’s something that is very specific and measurable. If you don’t know what your goals are, it’s hard to design an effective split test. You need to know the entire point before you even start.
Test One Variable at a Time
You don’t want to change much. Start with your control, which is what you already have, and then change one variable. For example, change the buy button size or the buy button text but not both at the same time. Or change the colors on your site, but nothing else. That’s the only way you can determine what is really working best.
Spit Your Test Groups Randomly and Equally
When you are split testing, the best way to do it is both randomly and equally. That’s going to produce the best results. Because you want to do it this way, the best thing is to use software such as Google Ads or landing page software that will randomly deliver traffic to various versions of your landing page.
Understand How Large Your Sample Size Needs to Be
This will require doing some math but it’s a very important aspect of conducting A/B testing. You may not even have a large enough mailing list, for example, to do A/B testing in your mailing list. If you’re doing A/B testing with an advertisement on Facebook or Google Ads, you’ll know how many people you need to reach by doing the math. Here is a basic sample size calculator you can use to help.
Use Software and Technology
There is amazing software available to use to reach your audience and conduct A/B testing. Software like HubSpot.com, AWeber.com, Google Analytics and others including Leadpages.net, Instapage.com, ClickFunnels.com and more all offer functionality to make A/B testing easier.
Give It Enough Time but Not Too Much
You need to set a time limit for your testing, not just a traffic limit. If you let the test carry on too long the results may not be accurate. This is very audience-centric. If your audience is, for example, moms of school-age children who play softball, you need to test at the right time based on their schedule.
Test the Right Items
This is based on your goal. If you’re split testing open rates, logically you may want to test a subject line for your email. If you want to test conversions, go to the conversion factors like the buy button or the price.
Don’t Change Your Test Midway Through
Stick to your test once you start it for the duration that you chose to do the test. This is the only way to know for sure if what you thought would happen will happen. Remember that you’re testing your specific hypothesis. You are changing A to B because you think it will affect traffic, or open rates, or conversions. If you change the test midway, you invalidate the results and may end up going the wrong way.
It’s very important to conduct split testing using some form of criteria so that you know your results are valid. If you use the same criteria across campaigns, you can be sure that your results are accurate.