The danger with data is it can keep you trapped in a local maxima. Also there's the whole thing with a lot of users seeming to be more and more annoyed with telemetry.
THIS. So many people don't realize that true innovation cannot come from existing data.
Maybe, instead of either just declaring what is the "right" way or observing users like some kind of behavioural scientist, we could just go back to, I don't know, asking them.
Innovation comes from experimentation. It's useful to have hypotheses and quantitative/qualitative data to back the assumptions up or disprove them.