How to do A/B Testing
A/B testing, put simply, is testing two different webpages (or different elements of a webpage) against the other and seeing which one performs best. Half of your visitors are directed to webpage A and the other half to webpage B, whichever page gets a better conversion rate is the one you use. This process can be done in different ways and some people prefer to use multiple webpages which is called multivariate testing. A/B testing allows you to see which campaign is more appealing to clients and develop a strategy that is the most attractive to your target audience.
Step One: What are you testing?
Before you can begin, you need to decide what it is that you want to test, what you want to find out, and what you want to improve. There are a number of things that can be tested from the layout to the images used, the amount of text on a landing page, to the call to action button and many more. You need to be clear exactly what it is that you want to test and make sure that this is the only thing you change.
Step Two: What tool will you use?
Once you have decided what you want to test, it is time to select a tool that can do the job. If you are comfortable hardcoding or you have your own software then you can skip this step but if not, then here are some sites you can use: Optimizely, Visual Website Optimizer or Unbounce.
Step Three: A Controller and a Challenger
Once you have established the tool you will use to run the test, you then need to create an alternative to your existing variant. If you want to test whether you have too much text on a webpage then you need to create a new webpage with less text or a variant with less text that you can then swop in on the same webpage. The existing webpage would be the controller and the redesigned webpage or text variant would then become the challenger.
Step Four: Replace or Redirect
There are one of two ways you can do A/B testing. For example, let’s say that you have chosen to test the call to action. One way of doing A/B testing will replace the call to action button randomly on the same webpage so that some people see, for example, a smaller sign up button and other people will see a larger sign up button on your same webpage. If you want to see which theme works better for example, then the other way of doing it is to create a different webpage. Some people will be directed to the existing one and others will be redirected to the new one.
Step Five: How significant do you want your results to be?
Before you start the test, you need to determine how significant the results need to be for you to choose one variant over the other. The higher your percentage level the better. Usually, a 95% minimum confidence level is used. However, this percentage is most often for time strenuous tests. In other tests that are not as demanding, the confidence level can be lowered.
Step Six: Conversion
After conducting the test for a substantial amount of time and once you have achieved your confidence level, you can determine which button, webpage, theme or call to action led to more conversions than the other. Now, you can take action based on the results of your test and increase your sales.
Some things to consider when A/B testing:
- You must always test each sample at the same time. Testing them at different times may not mean that one is more successful than the other, it might just mean that there were more sales during that specific time. It is imperative for the accuracy of these tests that they are done at the same time to produce valid results.
- It is important that you leave the tests up for a considerable amount of time. If one is instantly doing better than the other, it does not necessarily mean that it is better it might just mean that it was a good day for sales on that day. It is best to run your test for a full week – Monday to Monday and if the results are not significant enough, run it for another seven days until they are. Pay attention to holidays and external factors that may have affected your A/B test that might make it less reliable.
- Only run one test at a time, running multiple tests on different things will cause unreliable results and you will not know whether it was the button that made users choose to buy or the webpage or the amount of text. One test, on one thing at a time.
We post weekly articles on sales calls, LinkedIn for marketing and the latest digital strategies. If you want to keep informed on the latest trends, then click subscribe. For faster guaranteed leads email email@example.com to see how we can help increase your sales.