Multivariate Testing, 1 2 3…


mic_smallADOTAS – Many online marketers use web analytics tools to track and understand what’s happening on their websites. Key factors commonly tracked include number of visits, bounce rates, who clicks on what content and where traffic comes from.

The most successful online marketers, however, integrate their web analytics data with multivariate testing tools, which enable them to test many changes on their websites simultaneously without having to increase total sample sizes. There is a very natural relationship between testing platforms and analytics applications—one is fundamentally designed to drive improvements on your site and the other is designed to help quantify the value of those improvements.

Thus, the combination of measurement and testing should support both optimization and an incremental learning process about your visitors and customers. Assuming you have a robust measurement program already in place, the integration of testing into those efforts is often trivial and requires little more than patience.

The upside from taking the time to integrate these systems correctly includes the ability to evaluate testing efforts over multiple criteria and the ability to evaluate test participant behavior over multiple sessions.

These online marketers typically use a “test-learn-repeat” pattern that helps convert more browsers into buyers. Expert online marketers constantly test to learn what users prefer in terms of content, offers, design and navigation.

They also test to learn what users don’t click on and why. It is quite common for seemingly dramatic design changes to have no significant impact when examined using simple measures such as click-through and conversion rate. An increasing number of companies have started leveraging more complex measures such as “return visitation rate,” “lifetime customer value,” and using more qualitative measurement systems to develop a more holistic view of test impact.

By integrating testing and web analytics, marketers learn what works and what doesn’t. They discover new correlations between web site content and visitor behavior. Marketers also learn that sometimes subtle changes, such as the color, size, placement or call to action of the buy button, can positively or negatively impact sales.

Here are some additional examples of site factors to test and track:


  • Long vs. short;
  • Style and tone, such as chatty vs. formal;
  • Positioning, such as which value propositions, features and benefits work best; and
  • Call-to-action text.

Font, color and size

  • Think about your target audience. For example, if your audience is older, test larger fonts.


  • Color;
  • Location; and
  • Copy.


  • Sequence of items;
  • Labels;
  • In-line text links; and
  • Link style, such as underline or bold.


  • Content of image, such as including a man vs. a woman vs. a group;
  • Size; and
  • Location on page.


  • Location and size of areas or boxes on the page;
  • Attention focus, such as which layouts help focus vs. create distraction; and
  • Buttons and links, and where to place them.


  • Three-page checkout vs. five-page checkout;
  • Requiring form fields vs. making them optional; and
  • Adding bells and whistles vs. turning them off.

Testing with analytics in mind is a process that can start off small, such as changing a few things at a time and then moving on to further Web site modifications in an effort to constantly tweak the user experience. It can also start out large by testing many things at once and looking at interaction effects. Either way, the smart marketers who use both testing and web analytics will understand more quickly and easily what’s needed to optimize their Web sites.

The integration of measurement and testing is designed to help you better quantify test performance and the resulting impact on the business. Integrated systems, when properly used, support a wide range of metrics and measures and support the analysis of both short- and long-term impact of tests. The long-term view towards testing is one that few companies currently take today, but this view most often reveals a great deal of insights about visitor behavior.


  1. Carl – multivariate testing from firms like SiteSpect and others (Maxymiser, Optimost, Omniture etc) takes things a step further, not only do they show you where the issues are but create the solutions too. Knowing there is a problem with conversion is only stage one, creating the solution is where multivariate testing comes in.


Please enter your comment!
Please enter your name here