How to run simple A/B split tests

A/B split testing is an awesome method to figure out everything from messaging to colors and layouts of your website. The basic principle of a split test is to show different groups of visitors different pages, and comparing the results. Rinse and repeat – each time, taking the best performing variation for overall performance gains. The great thing about A/B split testing is that the results are measurable and have a direct impact on conversions. So here’s how you can get started with running your own tests.

Planning tests

The most important thing when it comes to running tests is the planning. Given the magnitude of possible variables, it is absolutely necessary to start with what you want to achieve. This could be increasing top of funnel visits to your website, open rates or click-through rates on your sales emails. Once you have these goals, you can then create hypotheses about what affects these outcomes. For example, it could be the email subject or color of the buttons or layout of a web page. Finally, you need to plan for the test variation and control (the version that has no changes) to run concurrently to account for any possible timing factors. Now you’re ready to go.

Setting up an A/B split test

For websites, there are a couple of ways you can set up an A/B split test. The easiest is to use a service such as Optimizely or VWO. They both have free trials, and they allow you to design your variations using a graphical interface. Once you set it up, they will automatically serve different variations to different visitors and track the results. The system will also let you know if the results you gathered are statistically significant. The second way is to create variations from your pages at a different URL and set up Adwords to drive traffic to them. This is a lot more manual but works in a pinch.

You can also split test your sales/email newsletters, all you have to do is to segment out a small portion of your email list to send both variations. Once you have the results, you can send the better performing one to the rest of your page.

Let the results inform your actions

Now that your A/B test is done, you should analyze your results to see you have learned. Can these lessons be applied to other parts of your website? Sometimes the results may come out surprising – and that’s really the point of doing these tests. But do know that if you decide to go ahead and implement a change that you don’t feel comfortable with, you can always change it back or set up another test to find a more successful variation. Good luck!