Zynga has defended the idea that designing a game heavily based on metrics input is a winning strategy. Although no one likes the idea of what the output is after that process. The popular trope around the indie game world is a general feeling of distaste for the kind of game that gets produced by putting in more numbers based feedback rather than human based feedback.
So when the company is game design, it makes sense to track as much information as possible. And not use it inside the game. Games have to be fun. Making games mechanically for the “lowest common denominator” among hundreds (hopefully thousands) of players does not make a fun game. This LCD effect of designing for very large audience prunes out errors and issue, quirks, of the game. You get something with pretty art that plays like Farmville.
Metrics are still necessary; especially for a game design start-up company. But I want to argue that they are not important in the same way. Metrics are best used to design the company around the game, instead of the game itself. As the ‘business guy’ I can tell you it’s hard to run a company without metrics. Every company tracks something about itself. Old companies track income and expenses. Newer companies have a whole host of company metrics. Running a company would be impossible without tracking data about how that company is run.
Games should be fun. They should have quirks. They should not be ground out of huge player bases but should be made with devotion by passionate developers. Avoid the “LCD”, lowest common denominator. Go ahead and collect the metrics for the game. Use those metrics to build the company, not to drive game development.