Three Steps to Split Testing Your Landing Pages

Split testing is like flossing your teeth — you know it’s a really good idea, but you never quite get around to doing it. Let’s face it, neither activity is exactly loads of fun.

There’s an even more important similarity, though: Not doing it is ultimately going to cost you money. Split testing can help you develop the most effective landing page possible, with each element of the page proven to yield maximum conversions.

There are three steps to split testing (also known as A/B testing or bucket testing):

Create two different versions of your landing page. To make the test meaningful, there will be only one element that is different per test — otherwise you won’t be sure which change is the one that worked.

Commonly tested elements include

  • Headline
  • Wording of the call to action (“buy now” vs. “add to cart”)
  • Size, color, placement of the “click here” button
  • Product photo (test both different photos and the size and placement of the photo — in different tests, of course)
  • Description of the product’s benefit (“low price” vs. “easy to use”)
  • Length of copy (shorter is often, but not always, better)
  • Video vs. text
  • Different kinds of visuals (videos, stills, screen shots)
  • Incentive (freebie vs. upgrade or other bonus)

Once you’ve set up the two versions, have half of your incoming traffic land on one version and half on the other. After that, you’ll:

Measure the results. You need to actually calculate clicks, and not just rely on your feeling that one is doing better than the other. You should also let the test run until you’ve got enough numbers to:

Calculate the difference in effectiveness with 95 percent confidence that your numbers are statistically significant.

The smaller your pool of results, the more dramatic the difference needs to be before you have that confidence. (And if you’d rather have a root canal than do all that math, there’s a free tool that can help you crunch the numbers.)

In addition to the various elements of your landing page, you can run split tests on other pages within your site, email subject lines, banner ads, and so on.

(By the way, if you were wondering if there are any significant differences between split testing and flossing — failure to split test will probably not end with a couple of latex-gloved professionals poking around in your mouth with sharp objects.)

« Prev Article
Next Article »