A paperclip maximizer would at least have a supportable goal. I don't think that's the case here. A lot of the rethoric around "data-driven UX" just pretends t

A paperclip maximizer would at least have a supportable goal. I don't think that's the case here. A lot of the rethoric around "data-driven UX" just pretends that the company's goals are always perfectly aligned with the customers' - when of course in reality, they are often completely at odds. This makes it easy to talk about "improvements" and "optimisations" and make it seem like those changes would be in the interest of the users - when in reality, the changes are primarily improvements for the company. E.g. "retention" is really a poor proxy for user satisfaction. Yes, a user can stay because they like the service. They can also stay because they haven't yet found what they were looking for, because they don't know how to unsubscribe or because they are addicted. Retention doesn't care. However, what retention is a very good indicator of is how well a user can be monetized. I think, the fact that after so many years, retention is still one of the primary KPIs shows pretty well how serious the talk about "user satisfaction" really is.