Split testing (also referred to as A/B testing) is one of the most necessary practices in eCommerce website optimization. When you split test your website you are conducting controlled, random experiments to better understand visitors behavior in order to learn how to most effectively meet your metrics. Split testing works best when you do it frequently and often, and there are an infinite number of experiments that you can run for any number of goals you need to meet. Whether you are an eCommerce site looking to boost visitors shopping carts or a SaaS app working to get more conversions from trial to paid visitors, split testing will help drastically improve your chances of meeting your objectives.
Almost anything on your website can be tested. Whether it’s your headline, call to action button, links, images, or placement of awards and badges, you can A/B test it. And like any other experiment, there is a framework to follow that we have broken down into three categories: defining objectives, testing, and best practices.
First step is to clarify why your site exists. After you have determined that you built your site to sell a product, or maybe it was built to get people to subscribe to a service, the next step is to examine how you measure it’s performance. Are people leaving your site after just a few seconds, or are you seeing a higher number of shopping carts being abandoned, or maybe there are certain pages getting more views than others. Analytics can show some clearly problematic pages, like landing pages with high bounce rates, although some other areas may not be so obvious.
Once you determine what your pain points are, you can now look at value and cost of testing. How important is it to test each page? Would it be easier to test one page over another? Then ask yourself how much potential for a conversion rate increase does this testing opportunity have? Once you have thoroughly examined value, cost, and priority it’s time to test your current design (version A) against an alternate variation (version B).
When running a split test, incoming traffic to your website is distributed between the original design (version A) and the variation (version B). It is important to note that while the test is running, visitors to the site are not made aware that they are participating in an experiment. Whether you want to test your headline, banner images, call to action text, colors, or you are interested in overall engagement, bounce rate, page views, sign-ups, or click goals; by using effective goal tracking, any element or user interaction can be a source of data.
Once you’ve determined what you are testing against, calculate the number of visitors and days you will need to run a conclusive test. The amount of time required for a reliable test will depend on your conversion rates and how much traffic your website gets. Always keep in mind that not all split tests will be successful. Split testing uses a process of elimination and you will likely begin to see website performance improvements after implementing several different variations.
There are several universal best practices when running most A/B tests. The most important are as follows:
- Test one variable at a time – When testing simple variables like headlines, button text, or images, use one variation at a time. This way your results will be clear and you can be more confident that they are conclusive.
- Don’t run your experiments too long - Once you are sure of the statistical significance of the test, update your site with the winning variation asap.
- Use rel=”canonical” - Use rel=canonical instead of using a noindex meta tag on your variation page. If you are testing two different variations of your homepage, you don’t want the search engines to not index them. You just need the search engines to understand that your variations are what they are – variations of the original URL.
- Use 302 instead of 301 - Use a 302 temporary redirection method instead of the permanent 301 version when you need to redirect a visitor to one of you variations so that Google will keep the original URL in its index.
Today’s online market is extremely competitive and if you are not constantly testing and optimizing your site, you are going to find yourself behind the competition. But, Aydus can help. Using our custom split testing tools we will move, resize, display, or hide any element on your page in order to collect the most accurate data that will deliver results. Our split testing tools allow us to target specific groups of visitors for a more tailored experience and you will have access to live reports on how each variation is performing. If you’re ready to make your site more effective contact us today.