My Web Content is Better Than Your Web Content

By Robert Bergquist (@cmswire) Sep 14, 2007 Comments

1 comment

We’ve all experienced meetings that go like this: “OK, time to choose which landing page to launch with our new search campaign. Which page version do you guys like best?”

In the meeting are the copywriters, designers, product managers, SEM managers, and more. Around the table are some pretty smart, experienced marketing professionals. Everyone weighs in.

You’re not surprised each has a different, and often very strong, opinion. “The shorter quote is strongest.” “The headline should be blue.” “This image is best.” If that wasn’t enough, the division president drops in and shares her opinion too. What do you do?

Predictions are Often Wrong

The task is set up to fail. Trying to predict the best-performing page design is like predicting next month’s stock market close. Experienced people may get close, but accurate predictions are extremely rare.

Why? For two reasons: First, there are many thousands of potential page content combinations, so the odds of simply picking the best-performer are extremely low. Multiple versions and combinations of images, offer copy, color, calls-to-action, screenshots and customer quotes stack the odds against even the most experienced marketer. Which version of each content factor is most effective? Which combination of factors works best together?

Try it yourself. Guess which call-to-action button below performs best at enticing search visitors to sign-up for a free trial of an online project collaboration solution? This is based on a recent case study. Read to the end for the answer.

Figure 1. Call-to-Action Button Choices

Firgure 1

Secondly, people are fickle — their actions not easy to consistently predict. Sally may react better to a blue button and John to a red button. How do you know this beforehand? People react in different ways to different versions of content in different situations, for many complex and often unconscious reasons.

The reason why is not as important as knowing what resonates most strongly with your target audience. As marketers responsible for developing effective Web pages, our job is to figure out what will maximize conversion rates and deliver the highest returns on ad spend. Leave the psychoanalysis to others.

Multivariate Testing Offers a Better Approach

The good news: Innovative companies are taking advantage of the emerging field of digital content optimization using multivariate testing and predictive analytics. “Multivariate testing” is a mouthful, but the concept is simple and powerful. Multivariate testing enables marketers to modify key content areas of a web page, experimenting with multiple versions of an offer, image and copy, for example, to determine the best-performing combination under actual user conditions.

The operative word here is “actual.” Without interfering with the user experience, your actual site visitors determine for you the page that best inspires them to act. The guesswork is taken out of the final equation; you present the proven optimal page for your target customers. Multivariate testing with your actual user traffic will answer the question “Which page version is best?”.

Identify and Focus on Leading Factors

Another powerful attribute of multivariate testing is the ability to identify the relative importance of each tested content factor, allowing you to focus on those that have the greatest influence on driving conversions. For example, in the case study mentioned above, the influence on conversion provided by the sign-up button was more than six times the influence of another factor. Pick the wrong version and overall page performance will suffer.