split-testing

I’ve mentioned several times before on this blog that landing page testing is one of the most important keys to success. Almost no other changes you can make have as much impact on the bottom line as landing page split testing. I thought I would expand on that a bit in this post.

Why split test:

Small improvements in conversion rates can have a big multiplier effect on your net profits. Increasing your conversion rates from landing pages can take a struggling campaign and make it a superstar. If you look at top affiliates in almost every niche, one thing you can see in common is they are converting at a much higher rate than the average, and that is reflected in their higher eCPCs. Much of those gains are from relentless split testing of landing pages.

Unlike some other factors in your campaigns, you have complete control over your landing pages. You can make any changes you want at any time on the fly.

What to test:

Absolutely everything. You can test all of your ad copy, logos, images, colors, buttons, call to actions, authority images, relative position of elements, page structure, and about a million other things. Literally everything can and should be split tested. The only limit is your time and patience. The interesting thing about landing pages is that you just never know what will convert better. There is not a single formula or template that you can use for every campaign. I have been shocked so many times that a page I thought would be a complete winner was beaten out in a split test by some crazy variation.

Methods of split testing:

There are 2 main methods: A/B split testing and multivariate testing. A/B testing is generally when you test 2 completely different versions of a page and is a good choice for lower traffic sites. Multivariate testing is when you are testing multiple content sections on a page with higher traffic. Both can work well depending on the site, traffic levels, and your goals. As a science major I tend to geek out on setting up complex multivariate experiments, but it can be a really simple process.

There are many free server based php scripts you can install to manage your split testing. Most of these will rotate your traffic to the tested pages and gather your stats quite well. One of the easiest ways is to use Google’s Website Optimizer. Its very simple to set up and produces great stats using their analytics package. It will even help you determine the length of time to run your tests in order to gather statistically valid results. This great tool predicts how long you need to run your experiments based on different variables.

How long to test:

Always. Landing page testing is not a step in a process, its a process that never ends. You should always be split testing a new page against your currently most successful page. The quickest way to end a good campaign is to set it and forget it. Since you never know what will work better and there is always room for improvement, testing should never end. Just like most things in life, its only a matter of how much time and effort you want to put into achieving your goals.