- research methods
A/B testing, also known as split testing or bucket testing, is a research method to gain statistical insight into user preferences. In an A/B test, you create two versions of a user interface and run a randomized test. The performance of each version is based on well-defined goals such as the percentage of people who click on a link, trigger an event, or purchase a product. A/B tests allow you to compare 2 versions of something to learn which is more effective. Simply put, do your users like version A or version B?
A/B testing is a core practice of conversion rate optimization (CRO). CRO is the systematic process of increasing the percentage of website visitors who take a desired action — be that filling out a form, becoming customers, or otherwise. A/B Testing helps you get more conversions out of your traffic by allowing you to test different versions of your website against each other. The differences in the variants can be subtle but should kept to a minimum to know what caused the better (or worse) performance. It is also important that the experiments' samples are statistically valid.