Agreed, which is why user studies are (or should still be) a thing. You can't just add telemetry and hope A/B testing will surface the real struggles people hav

Agreed, which is why user studies are (or should still be) a thing. You can't just add telemetry and hope A/B testing will surface the real struggles people have with your products. Likewise, you can't remove the 1% case of "restore backup after critical failure" feature because it's used rarely. It'd be like removing seatbelts from cars or fire extinguishers from kitchens.