A/B testing (also called A/B split testing) is used to test two different versions of a web page: version “A” and version “B.” The data for A/B testing is usually measured via click-through or conversion.
By testing out multiple versions of a company’s web page (e.g., home page vs. product page) or a specific element that might change between A & B (even something as small as whether a form is placed on the left or right hand side of a page), we are able to see which version is most effective and, in turn, which draws the most traffic.
Testing can be conducted several different ways, namely: sequentially or in parallel. With sequential testing, each version takes it turn. Meaning, the various versions are offered within their own specified timeframe.
This is an easier testing method but results can also be skewed depending on time run, duration and location. Whereas in parallel (split) testing, both versions are made available simultaneously, and so the traffic is divided accordingly.