Increasing your response rate on a web site is one of the most important elements of User Experience. From better click-thru ratios on offers to greater satisfaction in site use, knowing what elements generate the best responses and reactions is key to the user experience. As a result, one of the essential tools for improving response is A/B testing.

By narrowing down elements and presenting different choices of a single variable, the tester can determine which choice produces the desired result.

What To Test?

Anything and everything! Certain ad copy, placement of a graphic, even the color of a Buy button can have a huge impact on a final purchase or use decision. By presenting choices of a single variable and measuring the responses, greater response results can be had, with a real time testing group who are already users of the site. As well, we don’t have to limit ourselves to simply two choices in our testing. We can present several options, in real time, and let people create the data that shows us the best results.

Likewise, A/B testing doesn’t necessarily need to present an equal split between the old and the new variable(s). While showing new features for an internal company information site won’t cause employees to flee their jobs, you certainly won’t make a radical change in an element on an outward-facing site and present it to half your users. This shock could result in a significant drop-off in users and sales. Rather, you can present the new feature being tested to a certain percentage of users; perhaps have the website show the new feature to a certain range of IP addresses.

Combined with tracking software such as Google Analytics, the choices present can be quantified and the results that give the best response implemented immediately. Rather than using an opinion from a meeting, trying to learn a ‘lesson’ from what is presented or make a ‘best guess’, we can let actual proven data guide to the best choice.