I'm not sure if it's clear from your comments above, but was the one year "trial" set up as an experiment, with controls and measurements? In other words, "in 2017, we were not remote, and we measured xx results. in 2018, we were fully remote and we measured yy results." Or literally, let's "pow wow" the upper management and ask them about their feelings?

I think I know the answer. I know that this is the ongoing core problem - measuring "results" is really hard. Running a business means "experiments" are expensive, and really difficult to control in a way that you can get some kind of scientific learning out of them.

I think about the Jim Collins books which were written based on decades-long studies. Certainly imperfect and vulnerable to survivorship-bias and various other issues, but still perhaps better information than any single company is able to come up with for themselves on what tactics would benefit their company more or less than others. (i.e. they may use successful tactics and profit, without knowing what the opportunity cost may be.)