Rather, this organization clearly had a mission to give executives results that made decisions they had already made look smarter-apparently so they were better protected when the decision didn't work out. I figured Ballmer would eventually figure this out, since having numbers that say one thing and results that clearly say another should have clued him into the problem, even if he ignored my annoying emails.
This was also about the time I first ran into argumentation theory, which suggests that we are hard-wired to hold in high esteem people who win arguments. What's really screwy about this, when put in the context of human nature, is that we don't seem to care that much if someone is right, only that he prevails. To that end, we'll follow executives that win arguments no matter what.
Now think of Ballmer's position. (I'm picking on Microsoft because I really tried to fix this problem and was catastrophically unsuccessful, and argumentation theory may explain why.) Ballmer's a business guy in a company formed around a brilliant software developer, Bill Gates. This would be like putting a hockey coach in charge of a tennis team. I don't care how long you're there, chances are you won't survive unless you figure out a way to fix the game.
My working theory: Ballmer inadvertently tasked research with a role of making his positions look right after the fact to offset the problem of him running a company of experts in an area where he wasn't an expert himself. The sad thing is that, had he approached analytics a bit differently, he could have made better decisions, held his position of power and made Microsoft a better company. I still hope he'll eventually see this path before his time there is up.
Effective Data Analysis Comes Before a Decision
Perhaps the saddest decision I was ever part of was IBM selling ROLM. I was part of the analysis team-and my report on how to turn the unit around instead convinced Ellen Hancock to sell the unit in the first place, since it showed a number of areas that needed substantial work.