There was an apocryphal tale of the landscape architect being asked "These foot paths are perfect, how did you come up with that design?" And she answers, "We planted grass everywhere and then replaced where the grass had been flattened by people walking on it with foot paths."

I had a professor once who explained how this can be problematic: What about the handicapped or elderly? Will they necessarily take the same paths? How does the ideal foot path change in winter months when there is snow? Is it just based on the first person who plowed ahead and started packing the snow down? In the spring, would the path be different if it rained often and you could walk underneath a tree?

This is not to say that the story is without utility, but that it can't be a universal maxim.

This is a perfect illustration of the how data is an important tool, and yet blindly relying on observable statistics can have problematic results.