One of the brighter patches during this gut-wrenching recession has been e-commerce. comScore, a leader in measuring the digital world, recently reported that retail e-commerce spending for the entirety of the November–December 2009 holiday season increased 4 percent over the same period last year, for a total of $29.1 billion spent online. Online retailers continue to hire web designers and developers, e-commerce analysts, and site merchandising managers, who have a plethora of tools and techniques they can use to optimize conversion rates and resulting revenue.
One increasingly common online technique is multivariate testing. While A/B testing allows you to test just one factor at a time, multivariate testing enables you to test many changes simultaneously. Evaluating the impact of combinations of factors and variations often reveals significant interaction effects that can have a dramatic impact on your conversion goals.
A/B and multivariate testing were around long before the invention of e-commerce. In my career, prior to the advent of the World Wide Web, I used A/B and multivariate testing to figure out which version of my company's direct mail campaign had the best conversion rate, and we'd roll out that version to the bulk of the list. It was important to test different creative combinations because invariably, some top executive wanted to weigh in on what they thought was the best creative. And invariably, they'd be wrong. But it would have been career suicide to tell them their ideas were wacky without backing it up with data. So, instead of saying, "I don't agree with you that a yellow headline is going to convert better than a purple one," I could just say, "We tested a yellow headline and it converted 82 percent less than the purple headline."
There's an industry acronym for that top executive: HiPPO, or "highest-paid person's opinion." I can't remember who coined this term, but it's perfectly descriptive of the politics that happen whenever you are creating something new. What testing does is reduce and sometimes eliminate the HiPPO's influence altogether, thereby allowing you to understand what your customers they really want from you.
We didn't have an easy way to execute online testing when the Web first came into being. I remember being in conference rooms with multiple VPs (there's the HiPPO in the room again) expounding on what our website should look like, what it should feature, and how it should function, without any way of knowing whether those ideas were effective. Now, there are many different ways to test and track user behavior and understand what combination of content influences that behavior. This data should drive the decisions you make about your website.
+ + + + + + +
About the Author: Kim Ann King is Chief Marketing Officer of SiteSpect Inc. Contact King at kking@sitespect.com