What is A/B testing? Ab Testing
A/B testing, also known as split testing, refers to a randomized experimentation process wherein two or more versions of a variable (web page, page element, etc.) are shown to different segments of website visitors at the same time to determine which version leaves the maximum impact and drive business metrics.
In the 1920s statistician and biologist Ronald Fisher discovered the most important principles behind A/B testing and randomized controlled experiments in general. “He wasn’t the first to run an experiment like this, but he was the first to figure out the basic principles and mathematics and make them a science,” Fung says.
Fisher ran agricultural experiments, asking questions such as, What happens if I put more fertilizer on this land? The principles persisted and in the early 1950s scientists started running clinical trials in medicine. In the 1960s and 1970s the concept was adapted by marketers to evaluate direct response campaigns (e.g., would a postcard or a letter to target customers result in more sales?).
A/B testing, in its current form, came into existence in the 1990s. Fung says that throughout the past century the math behind the tests hasn’t changed. “It’s the same core concepts, but now you’re doing it online, in a real-time environment, and on a different scale in terms of number of participants and number of experiments.”
Split Testing vs AB Testing: What Are the Types of Tests?
The terms “split testing” and “A/B testing” are often used interchangeably. They’re actually two different types of tests.
A/B testing involves comparing two versions of your marketing asset based on changing one element, such as the CTA text or image on a landing page.
Split testing involves comparing two distinct designs.
I prefer A/B testing because I want to know which elements actually contribute to the differences in data. For instance, if I compare two completely different versions of the same page, how do I know whether more people converted based on the color, the image, or the text?
A/B Testing Process
The following is an A/B testing framework you can use to start running tests:
Collect Data: Your analytics will often provide insight into where you can begin optimizing. It helps to begin with high traffic areas of your site or app, as that will allow you to gather data faster. Look for pages with low conversion rates or high drop-off rates that can be improved.
Identify Goals: Your conversion goals are the metrics that you are using to determine whether or not the variation is more successful than the original version. Goals can be anything from clicking a button or link to product purchases and e-mail signups.
Generate Hypothesis: Once you’ve identified a goal you can begin generating A/B testing ideas and hypotheses for why you think they will be better than the current version. Once you have a list of ideas, prioritize them in terms of expected impact and difficulty of implementation.
Create Variations: Using your A/B testing software (like Optimizely), make the desired changes to an element of your website or mobile app experience. This might be changing the color of a button, swapping the order of elements on the page, hiding navigation elements, or something entirely custom. Many leading A/B testing tools have a visual editor that will make these changes easy. Make sure to QA your experiment to make sure it works as expected.
Run Experiment: Kick off your experiment and wait for visitors to participate! At this point, visitors to your site or app will be randomly assigned to either the control or variation of your experience. Their interaction with each experience is measured, counted, and compared to determine how each performs.
Analyze Results: Once your experiment is complete, it’s time to analyze the results. Your A/B testing software will present the data from the experiment and show you the difference between how the two versions of your page performed, and whether there is a statistically significant difference.
If your variation is a winner, congratulations! See if you can apply learnings from the experiment on other pages of your site and continue iterating on the experiment to improve your results. If your experiment generates a negative result or no result, don’t fret. Use the experiment as a learning experience and generate new hypothesis that you can test.
A/B Testing Statistics: What Are Champions, Challengers and Variations?
The statistics or data you gather from A/B testing come from champions, challengers, and variations. Each version of a marketing asset provides you with information about your website visitors.
Your champion is a marketing asset — whether it’s a web page, email, Facebook Ad, or something else entirely — that you suspect will perform well or that has performed well in the past. You test it against a challenger, which is a variation on the champion with one element changed.
After your A/B test, you either have a new champion or discover that the first variation performed best. You then create new variations to test against your champion.
How Do You Interpret the Results of an A/B Test?
Chances are that your company will use software that handles the calculations, and it may even employ a statistician who can interpret those results for you. But it’s helpful to have a basic understanding of how to make sense of the output and decide whether to move forward with the test variation (the new button in the example above).
Fung says that most software programs report two conversion rates for A/B testing: one for users who saw the control version, and the other for users who saw the test version. “The conversion rate may measure clicks, or other actions taken by users,” he says. The report might look like this: “Control: 15% (+/- 2.1%) Variation 18% (+/- 2.3%).” This means that 18% of your users clicked through on the new variation (perhaps your larger blue button) with a margin of error of 2.3%. You might be tempted to interpret this as the actual conversion rate falling between 15.7% and 20.3%, but that wouldn’t be technically correct. “The real interpretation is that if you ran your A/B test multiple times, 95% of the ranges will capture the true conversion rate — in other words, the conversion rate falls outside the margin of error 5% of the time (or whatever level of statistical significance you’ve set),” Fung explains.
If this is hard to wrap your head around, join the club. What’s important to know is that the 18% conversion rate isn’t a guarantee. This is where your judgment comes in. An 18% conversation rate is certainly better than a 15% one, even allowing for the margin of error (12.9%–17.1% versus 15.7%–20.3%). You might hear people talk about this as a “3% lift” (lift is simply the percentage difference in conversion rate between your control version and a successful test treatment). In this case, it’s most likely a good decision to switch to your new version, but that will depend on the costs of implementing the new version. If they’re low, you might try out the switch and see what happens in actuality (as opposed to in tests). One of the big advantages to testing in the online world is that you can usually revert back to your original pretty easily.
How Do Companies Use A/B Testing?
Fung says that the popularity of the methodology has risen as companies have realized that the online environment is well suited to help managers, especially marketers, answer questions like, “What is most likely to make people click? Or buy our product? Or register with our site?” A/B testing is now used to evaluate everything from website design to online offers to headlines to product descriptions. (In fact, last week I looked at the results of A/B testing on the language we use to market a new product here at HBR.)
Most of these experiments run without the subjects even knowing. “As a user, we’re part of these tests all the time and don’t know it,” Fung says.
And it’s not just websites. You can test marketing emails or ads as well. For example, you might send two versions of an email to your customer list (randomizing the list first, of course) and figure out which one generates more sales. Then you can just send out the winning version next time. Or you might test two versions of ad copy and see which one converts visitors more often. Then you know to spend more getting the most successful one out there.