A/B testing is comparing two webpage versions to see which one will give a better result. As the name suggest, you should have a version A and version B. The versions should have some slight changes such as image location, button design, and headline. However A/B testing should not be confused with split testing: testing two completely different versions of a webpage. A/B testing can give you insight in the behavior of your visitors and shows you what you can change to increase your conversion rate.
What Do You Test?
It is possible to test all sorts of elements on your webpage. The most common elements are headlines, sub-headlines, text, call-to-action, buttons, images, social media buttons, testimonials, and links. To know what you are going to test with your A/B testing depends of the goal of your webpage. Keep in mind that with A/B testing you test one thing at the time. If you would like to test several elements ate at the time, it is called multivariate testing and the process for this slightly differs.
Lets say your goal of your page is to have people sign up. There are different possibilities to test. You can test the length form, the display of the form, or types of fields. Perhaps your visitors do not like the length of your sign up form, maybe it is too long or they do not feel that the information you’re asking is of relevance. They might wonder what you are going to do with (all) information, is there any privacy attached? It could also be the headline of your sign up or something ‘simple’ like the design of your button that does not appeal to your visitors. A/B testing can give you the answers to these questions.
The Process of Creating Your A/B Testing
Increasing your conversion rate is a continuous process of improvement. These 5 simple test will help you to create your A/B test.
1. Determine your goal
It has no use to A/B test if you don’t know what you want to test. It all comes back to the goal of your webpage. If you are unsure on what your goal is, you can start with the top of your funnel and work your way down once to have more experience with A/B testing. You then start with your lead generation page and end with your check out page or confirmation page.
2. Decide what to test
Once you have set the goal, you can decide on the possible elements you want to test. As mentioned before, there are many possible elements to test. For some pages changing the button color was enough to see improvement in their conversions. For others it can be the headline that gives the visitor more clarification on what they are doing on the webpage. It is always possible to test different elements later on.
3. Create your test
You don’t know before hand if another headline is going to increase your conversions, therefore you are A/B testing. You are not always sure that the thing you are testing is actually better. The data from the A/B test will reveal which version is better.
Besides deciding on what element you want to test, you need the technology to do it. However we have some modern technology nowadays that can make this very easy. There are several website that offers software for A/B testing, and with some of them you can start with a free trial. Some examples are, optimizely.com, unbounce.com, mailchimp.com, performable.com or Google Website Optimizer.
Once your test is up and running, the only thing you can do is wait for the results. According to unbounce there are a few criteria for running your test.
- Run the test for at least one week. This is to cover your daily fluctuations in access to your page
- On average you will want to have about 500 unique visitors see each variant in your test
- Don’t stop the test until the statistical significance has surpassed the 90% mark. Statistical significance refers to the probability that the conversion rate of your test page different from the original page for reasons other than chance alone.
5. Determine your winner
In the last step you should determine which version is a winner. For this you can use an easy calculator online that helps you to determine the winner. You have the fill the amount of visitors and goals for both version. Although it can be that the amount of conversions did not increase much or not at all. In this case the element you tested is not of much significance for your visitors and perhaps you want to test another element.
Some Do’s and Don’ts
Before you start with your A/B test, here are some do’s and don’ts to help you get along.
- Do not separate your test variations. Always test both version simultaneously to minimize external factors such as amount of visitors.
- Do not conclude too early. Wait until you have your statistical significance before you conclude which version is better.
- Do not surprise your regular visitors. You want to avoid shocking your regular visitors if you are testing the core part of your website. In that case, include only new visitors in your test.
- Don’t let your personal opinion get in your way. If your A/B test result show that version A is better despite the ugly color or weird headline, version A is probably better because that appeals to your visitors.
- Set a goal on how long to run the test. You don’t want to stop the test too early without getting the significant results.
- Use the option in your tool to show your repeat visitors the same version as always. This is to avoid blunders like offering different promotions.
- Be consistent with your A/B test across the whole website. If you are testing a particular button design, make sure that it is the same everywhere. Not doing this can affect your results.
- Do many A/B tests. It doesn’t matter if one test will not give much results, there are always other things to test. The more you test, the more likely your conversion will increase over time.