Split / A-B Testing

When trying to optimize your site for specific goals, a very handy process to engage in is A-B testing on specific pages. Essentially, A-B testing is creating variations on a page (e.g., Version A and Version B), and each time a new customer arrives at that page, the site automatically serves up either version A or version B at random. You then measure the success of each variation based on a predefined goal (e.g., did version A result in 50% of your users signing up for your free site versus version b on resulting in 10%). Based on the analysis, you select version A or B (in this case clearly A), and then start the process all over again, each time improving the effectiveness of your page. For example, google presents a helpful walkthrough on A-B testing in order to optimize the revenue you can generate by placing AdSense banners on your site (http://adsense.blogspot.com/2006/08/abcs-of-ab-testing.html).