It’s a simple fact that running A/B tests will boost your conversions- right? Unfortunately, not always. In fact, A/B Testing is often a frustrating experience where conversion rates either don’t improve, or they actually get worse. In fact, a study conducted by VWO showed that only 1 out of every 7 A/B tests was a winning test. That means that 6 out of every 7 tests were bad! It is easy to get discouraged and think that A/B testing is simply a waste of time, and these initial stats might not be helping persuade you, but the truth is A/B testing can work for you and your website. The difference is, you really have to take the time to study your results and study your audience. Here are our tips for how to get the most out of your A/B testing.
Do research ahead of time
Every other aspect of your marketing takes time, so your A/B testing should be no different. If you decide to blindly go in and make changes to your website or landing page, there’s a pretty good chance you won’t see positive results. Instead, do your research on determining your audience and what you think will best work to persuade them to convert. While it is true that A/B testing is largely comprised of trial and error, there should be information that you gather that helps you determine what your “trials” should be.
Look at other landing pages from your competition and analyze what they are doing, for example. If you are A/B testing a landing page, do your research on what good landing pages should look like (here is a good infographic on what not to do). While nothing is set in stone, the more information you compile and research you do ahead of time, the better prepared you will be to set up solid A/B tests that have a better chance of yielding positive results.
Set clear goals
This is important when it comes to measuring your results. While the true goal of anything related to your ecommerce business is to increase sales, you should be more specific when setting up an A/B test. For example, if you are testing a landing page, what does a user do when they convert? Are they signing up for a newsletter, downloading an eBook, making a purchase, or something else? In order to get any value from your testing results, you need to first understand what it is you are testing, and how much you want to improve by. Do you want to increase newsletter signups by 20%? Gain 1,000 new Twitter followers? Set a clear goal and keep it in mind throughout the testing process, and don’t forget to mark your starting point as well.
Make note of all your changes
It is easy to be over ambitious when setting up A/B tests, and you can easily get carried away switching CTAs, colors, layouts, copy, and more. What you need to be doing, however, is making clear notes about everything that you change and when you change it so that you are able to track your results over time. It is likely that your final result will be a combination of several changes you make over the course of your testing period. Some things you try will work, some won’t, some things work better when combined together; there are a lot of variables. Keep track of each change you make, noting what you changed, as well as the results from those changes. Some things you change might be super simple, but those simple small changes can sometimes make a big difference. Don’t run the risk of forgetting what you did.
Understand what you are testing
This tip goes hand in hand with the previous one. You will likely make a lot of changes and run a lot of different tests. Each time you run a new test, make sure it is clear what you are testing. Are you testing the location of your CTA button? Are you testing a new color scheme? For example, the images above show the one simple change that was made to the NuFACE home page. The first image has no line about free shipping, the second example does. In this case study, the site listing the free shipping performed much better.
When these results come in, you want to make sure you know exactly what was tested so that if you get a positive result, you know why. While you may test multiple things in one test, it is usually better to test one thing at a time. If you change multiple elements and get a positive result, how will you know what it was that caused the increase in conversions?
A good rule of thumb to follow is to understand what point you are starting at. If you are currently at a really low conversion rate, you may want to consider switching multiple elements in an effort to get rates to improve. However, if you are in a pretty good position, small tweaks (as in making one change at a time) will be a better solution for you. Either way, make sure you are aware what you are testing each time you start a new test.
Have a reasonable timeline
Finally, the most important thing to remember when A/B testing is to be patient. You cannot produce results overnight, no matter how badly you want to. If you look at your results after only a few hours, or even after 1 day, you will not have accurate results. You need to give the webpages enough time to accumulate enough traffic to get an accurate reading of their performances. You should run your tests for a minimum of one week, possibly more if needed. If you are getting great results right away, it can be tempting to want to stop the test and declare victory. On the flip side, if you start seeing negative results, you might want to call it quits immediately. However, it is always best to determine a timeline and stick with it.