If you want a classic example of the Folly of the Quants, look no further than Microsoft and Windows 8.
For those of you who don't know the story, Microsoft ditched the classic Start Menu from Windows 8 and replaced it with a full-screen menu that makes people want to put their fists through their monitors. Me included. This was bad enough, but they then tried to justify this nonsensical decision by citing user-behavior telemetry that allegedly showed the Start Menu just wasn't used all that much.
Lies, damned lies, and statistics, said I. Sure, you could say that nobody uses the Start Menu, therefore let's ditch it — which sounds to me more like a justification in hindsight than anything else. But what about the few people who do use it? I know I get tons of use out of it, pinning to the Taskbar notwithstanding. I'm reminded of similar stats about readers: not a lot of people actually read, but those that do read, read voraciously.
I've written more neutrally about this issue elsewhere, but this is my personal opinion: Using all this to justify changing over to a UI that is predominantly organized for touch was a bad idea. Touch UIs are not solely the problem; lousy motives are. (And Windows 8.1 goes some distance towards ameliorating these issues.)
This brings me, in turn, to the way stats and quants are used to justify idiotic decision-making in entertainment. The most common form of this is focus-test post-injection — where the studio demands that a dog be put into the film, because test audiences like movies with dogs! For a truly epic example of such a horror show, go look up the shoulder-surfing performed by Warner Brothers during the making of Tim Burton's Superman, which was bad enough to make Burton walk off the project without looking back.
A complete lack of a sense of context, of why things are put into something in the first place, is required to be this boneheaded. Ingredients are not themselves the finished product, but people who don't actually sit there and agonize over the significance of the details themselves never understand any of this. Would you hire a chef who insisted on adding paprika to an egg cream, because it worked so well on his deviled eggs?
Data is not human, and that's why worshipping it like a god is a mistake. What's more, when we place trust in data, we are never placing our trust in the data itself, but rather in the people who have packaged it up for us and derived conclusions from it. Nothing stops them from being fallible or self-deluding.
Other Lives Of The Mind