I work in a analytics centered branch of the gaming business, which is free-to-play. Mobile games constantly collect data on usage. The expression data-driven design is widely used, but I am not sure that is the right way of doing things.
People more expert than me say that there are basically 3 kind of design approaches towards data: data inspired, data informed and data driven.
Data-driven design, as defined here is when customer data makes decisions and shows what to design next. In my experience, this always reduces to a basic behavior: follow the biased interpretation of data made by your game lead(s).
Data is intended as information here and not as pure data. To me the best approach is the data-informed one. Players with their behavior should help you improve your game, but you should never forget about your inspiration and intuition. Games are artistic artifacts, at least good games are that.
Most of indie games using data prefer the data-inspired design, but that is very risky because often leads to not set goals properly. In order to set a good goal you should also set up key performance indicators to make periodic sanity checks.
Don’t let just your intuition decide, don’t let just the data. A good game is the perfect balance of both!