A Beginner’s Guide to A/B Testing

It sounds uber-technical, but A/B testing is something that anyone can do. Here’s how…

Sometimes your website can seem like an enigma. You’ve followed all the best practices and it’s still not creating the conversions you were counting on. If you’ve ever wondered why, you need to know about A/B testing. This is a method for picking your website’s brains. It can improve conversion rates by up to 90 percent in some cases.

What is A/B Testing

In a nutshell, A/B testing is an experiment on your website, which compares two or more variants. Using certain tools, you can send two or more different versions of your website out to different users to see which one works best according to a goal. These goals can be everything from conversions to click-throughs.

How do I start A/B testing?

There was a time when A/B testing was the realm of web designers who knew HTML. That time has passed. There’s now a range of software available that makes A/B testing easy. At UK2, we use Visual Website Optimizer. It’s a paid-for tool, and the price you pay depends on the number of visitors your site attracts.

Tips for Using Visual Website Optimizer

You can plan an A/B test in minutes using VWO. Before you start, you have to give VMO permission to access your website pages. You do this by embedding a script on your website. This is actually less difficult than it sounds. If you have a WordPress site, you can use a plug-in to give permission. If not, you should ask your web developer to help.

From here, starting an A/B test is as simple as clicking on a create button and entering the URL of the page you want to test.

Then you can use drag and drop functionality to change elements of that web page compared to a control. You can do everything from changing text, the sizes of page elements, colours and style.

To send the two different versions live to your users, you simply need to click the ‘start now’ button.

What Should I A/B Test?

There are a few common A/B tests that can often be instant wins with regards to goals like click through rates and conversions. These include…

1. Calls to action: Test the wording, colour and size of your calls to action such as Buy Now or Find Out More buttons. Some companies have found that the colour red works well for call to action buttons.
2. Image size: Certain studies have shown that bigger images can improve bounce rates of a website.
3. Navigation bar: Experiment with the order of your navigation bar and see which order leads to the best engagement.
4. Headlines: Play around with the structure of your headlines to see which ones gain the most attention.
5. Page layouts: Experiment with introducing your call to actions in different places.
6. New functionality: See how new widgets are responded to.

Golden Rules of A/B Testing

Timing

To get validated results, the control (current version) and the variant (new version) need to be running simultaneously. If you run variant 1 during week 1 and variant 2 during week 2, the results cannot be statistically significant.

Sample Size and Duration

They key rules here are don’t rush your test and make sure you experiment on a large enough sample.

Focus on Fresh Meat

Try not to run A/B tests on frequent visitors – best practice suggests experimenting on new users only. A loyal user will have a different reaction to a new environment then a person who never saw your site before and that will impact the resulting data.

Which Other A/B Testing Tools Are Available to Beginners?

Optimizely

Unbounce

Wingify

Convert

Google Website Optimizer – This is free but requires a bit more technical knowledge to use.

For more information about running a website, visit the UK2 Website Builder page.