Skip to main content

, ,

Tips for Top-Notch AB Testing

on March 9, 2018

AB testing (aka split testing) is the process you use to compare two versions of the same content. Based on your results, you know content resonates best with your target audience. In this post, we’ll answer some of the most FAQ about AB. There’s a lot of info out there. But don’t worry— that’ll be a two-parter, so stay tuned.

Top- Notch AB Testing and Why it is Essential to You

What Goes into AB Testing?

How AB Works

The AB testing process has four main stages:

  1. Make 2 different versions of a webpage.
  2. Direct some segments of your audience to one, some to the other.
  3. Check the analytics.
  4. Link all traffic to the higher-performing page.

Some Quick Notes About AB Testing

ABn testing allows you to test more than one variant against the control. So, instead of showing two groups an original page and one with a variation, you can test multiple pages, with multiple variants, among multiple segments of your audience.It is more complicated, but your analytics (and the high-performing content you create from them) are that much more sophisticated.

Best Practices for How Long to Test

  • At least one (up to 2) full business cycles
  • Until you meet predetermined sample size *and* point of statistical significance
  • In full-week increments

PRO Tip: If you’re not sure how to calculate your sample size, use a tool like this one.

In a nutshell, AB testing helps you maximize your return on investment, and ensures your content s best tailored to your audience. But it’s not always right for everyone. If you’re a low-traffic site or startup, your best feedback might come from asking your customers directly what works best for them content-wise.

But if you can meet the required sample size in two to four weeks, AB testing is definitely the way to go.

How to Figure Out What to Test

AB testing isn’t one-size-fits-all, and your data should be the driving force behind all your AB testing decisions. However, there are some common ways successful brands figure out what to test.

  • Find the problems: How are your pages loading? How do they work on different devices? Do some formats of the page (mobile, desktop, etc.) perform better than others? Looking at your analytics for what’s not working– then design some AB tests that can help you fix it.
  • Surveys: Whether it’s a pop-up after someone makes a purchase, sent via email, or offered as an opt-in for an incentive, surveys are a great way to figure out how and what to test. You can also gather feedback over the phone or in person with some good old-fashioned customer interviews.
  • User Tests: In a user test, you watch someone perform tasks on your site and ask them to describe the experience to you as they go. Noticing what they find confusing, and what they click on first (or last) may give you some good ideas about what to AB test.
  • Session Replays: Session replays let you watch as an actual site visitor (rather than someone you asked to do a user test) explores your site. You can see what real prospects are doing (or not), and structure your AB tests from there.

As you can see, there are tons to learn about AB testing. So we’re making this thing a two-parter. When you’re ready, [jump over to part two of this series], where you’ll learn how to prioritize your AB tests, and get a cheat sheet on AB testing statistics.

See you there!

Not sure where to start?

Find out how much your competitors are spending on digital ads right now.
Take Our Competitor Analysis