It should be unsurprising that we don't really have a science of what company policies work best. There are a lot of variables at play and they likely interact

It should be unsurprising that we don't really have a science of what company policies work best. There are a lot of variables at play and they likely interact with each other in non-linear ways. Outcomes are hard to measure and may look different at different time horizons. For example, if policy A results in more productivity but makes retaining effective employees more difficult you may need several months or even a few years of observations to truly understand the effects. Companies are quite private about their information and they want control over their own company policy. Experiments are costly to perform. How would you perform one? Randomly select a bunch of companies and try to randomly get them to adopt different sorts of policies? I think this is why you see a lot of herd behavior amongst companies. Some business leader who feels they just have an intuitive sense of what's best just tries something. Are they right? Who knows? But if the company does OK it is evidence the policy isn't a total disaster so more companies follow suit.