What is A/B testing?
A/B testing is a method or approach that helps you make strategic decisions about your online presence. This could be anything from the colour of the buy button on your website, the images you use, or the text you think will entice readers to want to know more.
You can also A/B test ads you use in search engines and social channels, as well as email blasts to existing customers.
In short, it's comparing two versions of something to find out which one works best. Which colour choice leads to more purchases? Which ad generates the most clicks? Which wording in the sales email gets the most people to visit the website?
Design and copy are often subjective. We all have different preferences for what we think is nice or good writing. By collecting data, we can make decisions based on statistics instead of just gut feelings. That way, we can figure out what strategy works for your particular business.
Benefits of A/B testing
The main benefit of A/B testing is that it can lead to better profitability. If you make strategic decisions based on data, you'll end up with a website that converts better and ads that your target audience wants to click on. Working methodically in this way allows you to optimize conversion points, which leads to more conversions.
Regular testing ensures that you keep your website up to date. What worked a year ago won't necessarily work today. Trends change and people change their minds. By regularly testing your marketing activities, you ensure that they are constantly in line with consumer needs and preferences.
Visitors tend to stay longer on a website if there is content they like. By testing out a website with well-written text and images that appeal to your target audience, and using CTAs (Call-to-Actions) in the right place with the right colour and message, you can ensure that visitors stay longer. Something that reduces bounce rates and helps you gain higher visibility in search engines.
How do I do an A/B test?
Step one is to think about what the goal is. You probably want to increase conversions, so think about what it will take to achieve that goal.
For example, do you want to get more people to click on an ad or fill in a contact form? Set as concrete goals as possible and use data to find out how your marketing activities are performing at the moment. By knowing how your business is doing today, you can set reasonable goals. For example, if you have a conversion rate of 1%, you could set a goal to increase it to 3%.
There are two ways to conduct an A/B test:
One way is to develop an alternative (B) that you compare with what already exists (A). For example, to develop a new heading for the website (B) that you compare with your old heading (A). This will tell you straight away whether it is worth keeping the old heading or changing to a new one.
The other way is to produce two versions of something and compare them. In order to know whether A or B is really right for you, you first need to know what the results look like today. For example, if you develop two brand new headlines and find out that A converts better than B, it doesn't matter if the old headline still converts better.
You don't just want to see which of two options performs better. You also want to see that your changes lead to an improvement from how things were before.
Whether it's on the website or in another channel, you can review the colours, text, images and placement of CTAs. Is the content easy to read and do people quickly understand what you offer? Is there a prompt to move on or click to buy? What fields are you using in the form?
Find out which elements you can try and then think about what you think will have the biggest impact.
If you change everything at the same time and run two completely different ads in parallel, you can't know exactly what it was you changed that made one perform better. So start by making small adjustments, such as changing the headline. If it's a contact form, try removing a field to see if that gets more visitors to actually fill in the form.
Example of an A/B test
A concrete example of an A/B test is to make two versions of a sales email. Send it out so that half of the people in the mailing list receive each version and then see which version performed better. You can measure how many people opened the email, how long they stayed in the email and how many clicked through.
Once you have the answer in front of you, you will know which version to use in the future.
Disadvantages of A/B testing
Just because you can A/B test just about anything doesn't mean you should. A/B testing is time-consuming and it's easy to get bogged down in the details. Instead, focus on the big strategic moves and which changes you think will have the biggest impact.
Changing the colour of the buy button can lead to increased conversions. Changing the sales pitch on the first page can make visitors stop and click through. The choice of image for your display ad can make or break whether someone clicks on it.
You also need to be patient when it comes to A/B testing. It can't be done in a day and usually not even in a week. To get reliable results, you need to collect enough data and the time it takes therefore depends on how much traffic you have on your website. An A/B test takes at least 2-3 weeks, but often a little longer than that.
Checklist for A/B testing
Here is a checklist you can use when planning to conduct an A/B test:
- What is the goal of an A/B test?
- Decide what you want to test
- Create two versions (or a new version) of an ad, landing page, image or similar.
- Set a time limit for how long the test should run
- Choose the tools you are testing by
- Go for it!
- When the test is done, look at the results
- Adapt and redesign as needed
Tools for A/B testing
There are many tools on the market and one alternative is Google Optimize, which is free.
Want to know more? Read more about A/B testing your website with Google Optimize