The field of UX wasn’t born the moment someone wrote the first telemetry library.
If anything, people seem to have much more difficulty understanding user pain points right now.
So while telemetry might show that moving an item from one group to another (just making something up) takes > 1s, fixing this will not bring in $.
So when we then do Sprint Planning all of that gets pushed to the ice box.
The question whether it can tell you the value of anything at all is a hard one that needs plenty of context, and nobody seems interested on answering. But your reasoning doesn't need this answer.
But it can allow you to extrapolate from the value of things already done and usage patterns around them?
This is why companies like Microsoft cram it down our throats.
Telemetry has been the default for networked applications since longer than I've been alive. Think of a terminal connecting to a mainframe, how much telemetry it has access to, all of it, of course.
Huh? No, they had access to basically nothing unless the user did something that triggered a network request. What did you type in that form, but delete before submitting? No visibility. Which parts of the page did you linger on the longest? Which parts of the text did you highlight? No visibility.
They could see when you requested/submitted stuff, but that was about it. Pages couldn't sit there looking over your shoulder while you were using the page.
Well maybe I deleted it because I thought twice about what I wanted to send you.