When building a webpage, design and content are always major concerns. Decisions over these matters are often difficult to take: everyone has an opinion based on their personal taste and on what they think will work best. But if you are focused on driving performance, decisions should not rely on these subjective views – your options and hypothesis need to be constantly tested.
A/B testing is the way to do so: a truly simple and effective tool to improve your conversion rate. It provides you with the best answer on which content elements work best, allowing you to make efficient, data-driven business decisions. This practice, also known as “split testing”, basically consists in taking two or more variants of the same piece of content (like a landing page or call-to-action button) and showing them to part of your audience. The one that ends up with a better conversion rate – provided that it was seen by a statistically significant audience – wins!
A/B Testing should be used every time you want to make sure that any proposed changes favour your conversion goals. But they should also be used proactively: if you constantly test new content, you will be constantly improving conversion. The moment one test ends, another one should begin!
3 A/B testing tips
You can test almost everything, from a simple call to action button, to an entire landing page. There are however some things you should be aware of in order to have more effective results and better analysis.
We chose three tips that we think that sum up what you should pay attention when testing something on your website.
- Test continuously – By testing your content you significantly improve ROI. Therefore, the end one test should be the beginning of a another one!
- You can run several tests at the same time, but it is wise to restrict each test to one particular piece of content in your page. If you create a test where you confront your original page with one that is totally different (i.e. offers a totally different UX – structure and design), you will never know which particular changes might be influencing your result.
- When developing an a/b test strategy, you should start by focusing on testing the main components – colour, copy, format, position – of your most important call-to-actions (eg. a “finish purchase” button). Before creating your variants, develop a matrix with the combinations you want to test. Say you would like to test three different colours and four different copies in one button. This would be your matrix:
This means that besides your control (buy it – black), you should create 11 other variables in order to test all possible combinations and find the winner among them.
At the end of the test, your winner will be found, and you are ready to take some objective conclusions about what works best on your website and what works best for your audience. By constantly testing you will keep improving content and results!