Doing an A/B test
We like to test on real users. That's why we often do usability testing. What we also regularly do more data-driven testing. For example, in the form of an A/B test. We then make 2 or more variants of the same feature or flow, which we put live. Using analysis software, we can find out which variant performs better; which converts better, which generates more users, more interaction or longer page visits.
Turning data into UX improvements
You can use live user data in this way to effectively improve your product. Often A/B testing is part of a larger strategy to continuously improve digital products data-driven, along with campaigns to attract visitors to certain parts of your product and UX'ers who turn quantitative insights into concrete design suggestions.