A/B testing is a type of split testing. It helps to measure and improve specific metrics.
The audience is split into two groups, and both get almost the same item to test but for the only difference, e.g. group A gets a web page with a red CTA button and group B - with a yellow one. The number of clicks is measured to see which button is more attractive.
A/B testing can be used to test images, colors, layouts, or text.