Ignoring the unnecessarily editorialized title, this article is an example of "lie with statistics". The are basically arguing "look at the data", but every single example I've seen about arguing a data-driven approach to WFH is cherry picking things to fit their desired outcome, and that's on both sides of the argument.

Yes, it's fairly straightforward to look at productivity for rote, repeatable tasks. "How many calls did a CSR handle?" or "How many insurance claims did this person process?" are easily measurable. But that ignores the fears that business leaders have that, over time, the kind of innovation, motivation and culture that in-person interactions help foster will fade, leading to a sort of slow, long term decline. And the past 25 years have shown that companies that can foster and harness innovation well do much better than companies that simply focus on "repeatable process improvement".

I'm certainly not arguing either side is "right", but I would say anyone arguing "look at the data" on this topic is just using data to match their desired outcome, as there is simply not enough time to assess the long term "innovation" impact of WFH.