Ben E. C. Boyter's Blog

A/B Testing (822 words)

A/B testing is testing the comparison of two outputs where a single unit has changed. It is commonly used when when trying to increase conversion rates for online websites. Also known as split testing an example would be trying to increase user clicks on a specific button on a website. You may have a theory that red buttons work better then green. You would try out both against real users and see which one performs better.

You should use A/B testing when you have some existing traffic to a website (or mailing list), know what outcomes you are trying to achieve and most importantly are able to track what actions resulted in the outcome. Without these three things split testing will not achieve the outcomes you wish.

Numerous things are able to be A/B split tested and are limited to what you have to pivot on and your imagination. A few examples include,

Things to keep in mind when A/B testing is that running your tests for a long time can result in SEO penalties from Google and other search engines. Quoted from the Google Webmaster Central Blog on Website Testing,

“If we discover a site running an experiment for an unnecessarily long time, we may interpret this as an attempt to deceive search engines and take action accordingly. This is especially true if you’re serving one content variant to a large percentage of your users.”

It is highly recommended to read and understand the post mentioned in order to ensure you are following best practices. The consequences can be dire indeed including being black listed by Google and other search engines as the worst possible result.

A/B testing can be implemented in a variety of ways. Perhaps the best known is using using Google Analytics. However there are other free and paid for solutions. Visual Website Optimizer is one example of a paid for service and if you are using Ruby on Rails there are many libraries to help you out.

A few things to keep in mind when doing A/B testing.

A real genius in this space is Patrick McKenzie and a few very worthwhile articles to read about it are A-B Testing Made Me a Small Fortune and A-B Testing. Other articles worth reading include, Practical Guide to Controlled Experiments on the Web by Microsoft Research (PDF), Writing Decisions: Headline Tests on the Highrise Sign-Up Page], “You Should Follow Me on Twitter Here”, How We Increased our Conversion Rate by 72%, Human Photos Double your Conversion Rate