I found this article on Gamasutra recently and quite frankly I love it.  It's a very well written article from Laralyn McWilliams all about how metrics aren't meant to just be blindly read and followed.  Primarily she talks about games but really this advice applies within any part of life which she also makes abundantly clear.  In any case though, it's about games that attempt to get players to pay to skip the grind or simply advance in the game.  The people who pay consistently are called "Pay Whales" and though these types of games aren't geared specifically toward the Pay Whales but really how it's geared towards make you a Pay Whale.  This is the point where these companies start to sound sinister but the truth is that more often than not, the developers really are trying to make a great game.  But it gets messed up when people who are just looking at the pure metrics without thoughtful interpretation are making all the decisions.  This isn't about how these games are evil but rather how you can quite frankly make a lot more money by thoughtfully looking at the data rather than blindly following the numbers.  This article actually contains one of my favorite thought's on the matter:

"We don't have to beat them over the head with bricks just so they'll pay us to stop beating them."

Anyway the article can be found here:  http://www.gamasutra.com/view/feature/188197/the_metrics_arent_the_message.php

I would highly recommend it.