A/B Testing (or Split Testing) is a method of testing which of two options is better. An easy example when you create one ad campaign but have two different versions of ads that randomly show in the first phase of the campaign. This type of A/B test is easy to conduct in AdWords, and you can quickly see which version (version A or B) performs better in terms of click through rate (CTR).
One mistake that people often make with A/B testing is that, at the beginning of their campaign, they assume they are starting off with a “good A” and “good B” and so they just need to see which is better. But if you are at the start of your campaign, you really have no idea which is better, or if either is any good. So when you are starting out, you need to first find the right “A” before you can really think about fine tuning the messaging.