Have you ever gotten into a discussion (or perhaps even an argument) with a colleague or client over the effectiveness of one particular feature of your web application? Goodness knows I have. The good news is that if you can both agree that the end goal is to serve your site's users, then there's a solution: A/B split testing.
A/B split testing is a technique that will let you and that colleague or client stop arguing about which of you is right, and instead let you both try out your own idea and measure how users react to it. For instance, if you're arguing about the title sentence on your homepage or the placement of the submit button in your registration form, you can do the following: set up your site so that it serves each alternative to half of your site's users on a random basis. You'll want to track which user sees which alternative for two reasons: first, you want to show each user the same feature consistently, and second, you'll want to record how they react. Does submit button placement get more users to register than the other? Does one homepage slogan get users to request more pages than another? The possibilities for testing are endless. Just remember to test only one change at a time, whether it's copy, color, placement, etc. If you test more than one change, you won't be sure which change accounts for any differences you see in user behavior.
So stop arguing and get testing. Let everyone put their alternative out on your site and let your users decide which is the best.
Post a comment