A/B testing is similar to feature switching, but we usually use it to test the effectiveness of different web page designs and how they affect the conversion funnel analytics. We normally use it for digital marketing rather than for software development because it's less about what is correct and more about what the customers prefer. However, the underlying technical principals are comparable to feature switching and experimentation.
In a simple example, you serve half your visitors the old version of a page and the other half, a new version designed to increase engagement. You then record how many visitors click through or perform some action on each variant. If the new version performs better, then you keep it and roll it out ...