Yeah, terminology is tricky here. I don’t love “face-lift” either — it implies something less comprehensive than what we did. Internally we called it a “refresh.” But I’ll stand by “redesign,” too: if we’d done more usability work I’d probably have used other terminology anyway. (One doesn’t typically do a top-to-bottom usability overhaul since it tends to involve a bunch of disparate things.)

I share your skepticism about simply asking users what they think — better when you can get behavioral data — but I wouldn’t classify surveys as useless. They have their place, as long as one is aware of their limitations. In this case, we wanted to make sure we weren’t missing some reason why users would absolutely hate the change on day one, for which a survey seems reasonable.

As to getting harder behavior numbers: for this project, that would have been prohibitively expensive since we’d need to run the new and old look & feel in parallel for at least a month. If we’d simply tweaked some styles that would be feasible, but we made far more comprehensive front-end changes in this case. Something like that might be more doable in the future, though.