March 23, 2010

How to Set up A/B Testing for an Online Campaign


Hostway Team

By Monique Martin

Even if your product or service actually is the best thing since sliced bread, you can kill your conversions with a poorly designed landing page. If your users bail out before they even get going, you’re doomed. Changing your landing page is simple enough, but knowing what to change and what not to change can be a bit more challenging. That’s where A/B testing comes in.

What Is A/B Testing?

A/B or split testing is a type of marketing testing that lets you compare a control page (your baseline unchanged page) with various new versions of the page. This will help you determine which elements are the most successful, and which need to be changed or removed.

How to Conduct an A/B Test

Once you’ve decided on the page you want to test and what your goal for conversion is, you need to create alternate versions of the page. Be sure to change only one element on each new version. If you change more than one, you won’t be able to properly evaluate what affected user behavior. Here are some elements you might vary in your tests:

  • Call to action wording or placement
  • Color schemes
  • Headings
  • Banners
  • Image locations

Be sure that your landing page doesn’t have any extraneous links including top level navigation. Do not make any changes to the original page until testing is complete.

Now that you’ve created your experimental pages, you need to conduct the test. Ideally, you should serve up the landing pages randomly and to an even number of visitors. You can’t just make one version available one day and another available the next. There are too many variables that can affect your sampling. To get true results, you need to randomize the pages and have the same number of visitors to each.

In order to randomly serve the pages and control the number of visitors, you’ll need to install small bits of JavaScript. Tools like Google’s Website Optimizer offer Webmasters a quick and easy way to implement and track experiments.

Tip: Make sure your sample size is large enough and that you run the experiment for a sufficient duration to gather good data. Anyone who’s ever flipped a coin a dozen times knows that you might get 12 heads in a row, but if you keep flipping it, the odds will even out. Don’t be wooed by small samplings.

Once you’ve run your A/B tests, you’ll have invaluable information that will allow you to design the most effective landing page possible and maximize your conversion rate. A/B testing is a great way to refine your Web pages. And, it works for email marketing too. Don’t be afraid to experiment. The more you know about your visitors' behavior, the stronger your business will be.

About the Author

Monique Martin served as chief operating officer for a successful online insurance marketing firm for five years.

Stay in the Loop

Join Our Newsletter

Stay ahead of the pack with the latest news, web design advice, and digital insights, delivered straight to your inbox.
This field is for validation purposes and should be left unchanged.
© Copyright 2021 Hostway. All rights reserved.