You've just launched your landing page which looks great, concisely communicates the benefits of your latest promotion, and has a straightforward call to action. Now, you could simply review the reports to see how many impressions, click-throughs and conversions it generated. Or, perhaps you can take it one step further and create a different landing page to compare the effectiveness of the design, call to action or price points.

By applying A/B testing you can find out how people are responding to both landing pages, and where you can improve to generate a higher click-through rate and more conversions.

What is A/B testing?

Landing pages can be designed for email subscription or lead generation. But how do you know what works and what doesn't? A/B split-testing can be used to refine elements of your landing page for usability, conversions and most importantly, your return on investment. When you create two different versions of a page you can conduct an A/B test to see which is performing best. This is becoming a very popular practice because you can easily show different versions of your landing page to people at random (i.e. using your online marketing software, Google's Website Optimizer tool, etc). 

How do you conduct an A/B test?

Typically, you start by dividing your audience into two groups. One group is exposed to the original version of the landing page and the other group views an alternate version. The other version may have a single element changed which could include the headline, banner, or call to action. Then you can begin to track the results between the two slightly different versions.

For example, suppose you want to determine the best call to action to use for your landing page. If you have a form, you can insert the CTA directly into the submit button name. Instead of telling your visitors to "submit" their information, you can tell them what they will receive upon completing the form. For instance, the button could read, "Download our brochure." By creating two identical landing pages, with the exception of the "submit" button versus the test call to action, you can track the success of each page to determine the winner.

What should you be testing?

You can test a variety of elements on your landing page including the content, offer, headings and banners. Colors can also affect the success of your page, as well as the location of the images, form and call to action.

Only one element should be tested at a time throughout the A/B split testing period. In other words, do not try to test content and page design at the same time. It’s also important to avoid assuming what elements will perform best. Thorough testing will give you measurable results. What you expect to perform best might actually be the opposite.

NOTE: To ensure accurate test results, make sure your landing page is hidden from organic searches and the main navigation is removed. Typically, landing pages are hidden to ensure that people are only coming from your intended email marketing or PPC campaign. Furthermore, usually there is no top level navigation or unnecessary links so that people don't get swayed to another part of your Web site before completing the desired goal or conversion.

How do you track A/B test results?

There are two ways that you can track your landing pages. You can set whether A and B will be directly competing against each other by displaying version A to half of your visitors, and version B to the other half. Alternatively, you can have a dominant design that is seen as the standard to most of your visitors, then have a second design that’s viewed by a smaller group until the new design has proven itself.

Now you can compare the click-through and conversion rates between the two groups, and when you have your winner, it becomes your standard or benchmark landing page. Continue to test it against other page elements until you’re convinced that you have the best landing page for your online marketing campaign.

How do you know if your results are accurate?

The length of your test is determined by the strength of your data and a significant sample size. For instance, if your sample size is too small, results may be skewed because you will not have enough data to provide statistical significance. Make sure your time span and metrics are set to give you enough performance data to validate your landing page testing. For a comprehensive explanation of "The Math of Tuning" refer to Chapter 7 of the Tim Ash book, Landing Page Optimization: The Definitive Guide to Testing and Tuning for Conversions. If the results of your tests have not met your expectations, you’ll need to test longer or make further changes to each of your landing pages.

Remember, once you have decided on the look and feel of the page, it’s important to maintain the design until testing is complete. If changes are made throughout the test period, you will never know what was successful and what was not. Even the slightest change can affect your page results drastically.

When you have completed your test, don’t hesitate to do it again. Why? Because what appeals to customers will change over time. Thus, testing should be an on-going part of the marketing process to help you optimize and get the best results from your online marketing campaigns. Once you have the results from one test, examine them to determine how to improve the element tested, and then continue on to test the next element on the landing page.

A/B split testing is a cost-effective way to help you gauge how well your landing page is performing and what elements need to be modified or improved. In the end, you’ll know exactly what to focus on to generate greater results and higher conversions.

Additional Resources:

• Putting A/B Testing in Its Place
• The ABCs of A/B Testing
• Introduction to A/B Split-Testing for Your Webpages

###

About the Author

Susan Bell is a training specialist at Lyris, Inc. She is responsible for creating and delivering training and materials for Lyris HQ Landing Pages. She also regularly delivers training classes on Web Analytics and Search Marketing Fundamentals.

Comments (2)

Subscribe to this comment's feed

Name

Email

Title

Comment

smaller | bigger

Add Comment