We have benefited from A/B testing. Our latest findings were a test on some messaging for our financing option. We found a bigger conversion rate with one of the variations and implemented that messaging.
We tested changes to our add to cart, messaging, layout, and others, and have found that either, a) one variation saw better conversion rates or b) customer's convert the same with control vs variation, allowing us to put our efforts elsewhere. It helps us prioritize projects, seeing if these changes have value before we implement them.