Every Wednesday at 12 PM Eastern time, Joel Comm’s latest installment of The Joel Comm Show airs on Webmaster Radio. Last week’s guest was Mark Widawer, talking about the ins and outs of split testing. Here are some of the details. (To hear the whole broadcast, click here.)
Split testing can be defined as the organized process of having two different versions of something, for example your landing page, and then sending similar amounts of traffic to both to see which one gets you closer to your desired result, like more leads, more buys, etc. Mark emphasizes that this is not necessarily rocket science. “If you can count on your fingers, you can split test.” He says that it only makes sense to test the things that we do and that what we measure, we can improve on.
The reason more people don’t split test is that people are generally afraid of it, even though they don’t have to be. The things that you test are called factors and they can include headlines, landing pages, sales pages, etc., and then variations of each factor. There are various tools you can use to help you in the split testing process (Google Website Optimizer is one), and you will need some traffic so that you can test the different results. The biggest problem that Mark sites for people new to split testing is that their variations on headlines and pages aren’t varied enough. You don’t want to test things that are too similar or else all of your hard work and effort will be wasted.
He cautions against narrow thinking in split testing, which is basically when all of your ideas for variations are coming from your one brain, which means that the ideas will all be tied to the same set of assumptions about people and how they think. Get some outside opinions about possibilities and what might work, and then test them. Diverse ideas are what are wanted in split testing. While its true that they may prompt wild failures, they may also give you wild successes, and that is what you are looking for.