It’s always amazing how ineffective A/B testing is when it’s not done right. We’ve probably all experienced the frustration of testing and testing, only to get minimal results. So where are we going wrong?
For maximum results, improve the UX for A/B testing. Yes, the two are so intricately connected that they might as well just say “I Do” right now.
So why does this work so well? Let’s dive in and see how the two work together to bring harmony (and positive results) to a project.
Why Start With UX
The obvious answer is we need the best UX possible so we’re not getting overwhelmed with poor test results from a horrible user experience. Instead of truly optimizing anything, we’re just running around like a chicken with its head cut off fixing simple UX issues.
The deeper answer is UX give us a solid base for A/B testing. UX design always starts with trying to step inside the mind of the end-user. Through thorough research, we learn where the biggest core issues might lie, such as an older audience needing a larger font for a call-to-action than a younger audience. We learn more about usability preferences and which features actually are the most important.
Once we have this research base, we also have a blueprint for testing.
Common A/B Testing Mistakes
Improving UX for A/B testing helps us to avoid many common A/B testing mistakes that even professional testers make. Making common mistakes too often can and does ruin websites. Two of the biggest mistakes are not understanding exactly what to test and ending tests too early. While UX can’t really help with the length of the test, it does help with knowing which elements should be tested.
Design For Real Users
Ideally, we’re designing for real users and not just a bullet list of features that a client wants. Unless we’re putting the user needs at the forefront, all that A/B testing is going to show is that the project needs some major improvements before users will actually be happy with it.
The more we design for real users, the more we’re testing along the way. This already shows us how users are interacting with the design. It shows preferences, such as preferring the left side of the screen versus the header for a menu or tapping when they need to swipe on an app. All this test data is an invaluable resource when it comes to A/B testing.
Test What’s Most Relevant
We could easily waste hours and money on testing any and everything. However, let’s use a website as an example. Think of how many different elements are on a website. A site owner could test the wording of their CTA for months with only a slight increase in conversions. During the design process, we might have noticed that visitors tended to click the header more than the right side when it came to interactions.
This would mean that it would be more beneficial to test a different placement of the CTA first so it’s in an area where visitors interact more. Suddenly, conversions increase much more with just one simple change.
When improving UX for A/B testing, we’ve already looked at the most common issues and designed with those in mind. This also gives us cues for knowing where users might prefer some tweaks for better results.
Addressing The Root Cause
Obviously, it’s a give and take kind of process. Using UX for A/B testing means first creating the best possible design to start with and then using the research data to focus the A/B testing. Through A/B testing, we discover the root cause of the problem and then can tweak the UX design to improve the project further. It’s through this back and forth that we address the root cause of problems in a project.
It’s only when we combine UX design and A/B testing that we get the most effective results. We’ve already done all the hard work, so why not put it to use to improve A/B testing and thus improve UX design at the same time?
Images:Charles PH, Tim Gouw, Takahiro Sakamoto