What is A/B testing?
In context of the web, an A/B test is an experiment in which you can compare one version of a webpage (A) to another version (B) to determine which one performs better.
An experiment is conducted by splitting your website traffic into groups, showing variant A to one group and B to the other. By measuring the behaviour of both groups, you can determine which variant is better at leading your customers to click, call, purchase, or complete some other desired action.
Why should you test?
Small changes can have a big impact
A different image, headline, or call to action could have significant improvements on your conversion rates. If making small changes like these could translate into more revenue, more leads, and a happier customer, then why wouldn’t you test?
Testing settles arguments
How often have you debated with your colleagues over which banner image or call-to-action message would resonate better with your customers? Most marketing campaigns or website redesigns produce many variations of content and creative, from which a single option is chosen. But how do you know if you’ve made the right choice? Next time, avoid the debate and just run an experiment!
Build a data-driven culture
Testing can help your organization create a data-driven marketing culture, rather than one based on intuition alone. By continuously designing and running new tests, you will generate a regular cycle of learning about your website and your customers.
What to test?
Here are some examples of elements and features of your website you can test.
|Image/Graphic||Different designs, pictures or colors|
|Headline/Copy||Short vs long copy, paragraph vs bullet points|
|Call to Action Placement||Location of button or link on a page|
|Sign-Up/Purchase Form||Form design, number of fields, number of steps|
|Landing Page||Different versions of a landing page|
What you’ll need
First, you need to choose a tool that suits your needs. There are numerous A/B testing tools out there, but each one comes with different features and a different price tag. For those new to testing, Google Optimize is free, easy to install, and comes with all the essential features.
Next, you’ll need someone who can create the new images, copy, or other variations that you wish to test. Although knowledge of HTML and CSS will provide maximum control, many tools have a WYSIWYG (What You See Is What You Get) editor. With this, anyone can make changes to elements on a page without having to write code.
Example of the WYSIWYG editor in Google Optimize
Finally, you’ll need someone to configure the test, monitor the experiment while it’s running, evaluate the results, and recommend next steps.
If in-house resources are not available, you may wish to hire an analytics consultant to get you started with A/B testing. Having expert guidance will help you avoid common pitfalls when testing and help you extract relevant insights from the test results.
Now that you know the what, why, and how of A/B testing, it’s time to start testing. Because if you’re not testing, you’re not learning!