From its beginnings up to now, Economics has concerned itself almost entirely with Finance, so much so most people think of Economics as "the study of money". But it's not; it's the study of incentives. Up until now, money has been the best way to look at how people respond to incentives: what they buy, if they're willing to save, what they spend their entertainment dollars on. Given enough credit card receipts, you could piece together a personality, watch the eras in a life. But tallying credit card receipts from the garbage is the business of a decidedly "dismal science".

I'm excited to watch this change. The change probably started way back in the Chicago School and picked up speed with Gary Becker,but it's the Internet that's made it easy to see like-minded people in action. The group that first caught my eye was one of the first things I found on the web, baseball statisticians. Until I found Rob Neyer's (once-free) column at ESPN, I had no idea Bill James or the handful of other nerds existed. Now there are thousands of people running regressions and creating new metrics online. But there are still only a handful of people making a living at it. Which means the others all have day jobs. It would be nice (for us, not them) to think they're all actuaries carefully setting insurance premiums, but if my non-salad days are any guide, some percentage of them are working in crap customer service jobs dying of boredom or working in a field that doesn't require even a little bit of number crunching.

Pair that with something else I noticed over 8 years working as a consultant: irrespective of size, a lot of companies run on Excel spreadsheets.Some of them are well-designed affairs that link in live data from an Access database (bonus points if the Access database is just a reporting front-end to a more expensive product), some are of the "signle spreadsheet out on a network share on a machine we can't even find anymore" and the other 95% are spreadsheets "emailed around the team the first Friday of the month so we can keep trackof stuff (in an extroadinarily poor way)". Mom and pop shops and Fortune 500 companies make business decisions based on the data in these. But I always wondered how. It seemed like so many seances: cast the bones and read the pivot chart The Elders left us in the mists of time.

An example: one place I used to work at had reams of historical data gathering dust. A database with a record of everything anyone ever did on any project in the past decade and how much time it took. Even though our two biggest challenges were allocating resources and budget overages on projects, no one ever looked at that data. Instead the company paid for a senior manager to get an MBA so he could better crunch numbers. The only thing the MBA did was make him less open to input from others. Along with myself, there was anothe developer with a stastics/ economics background. Every couple of years, we'd get dragged into a meeting to look at a poorly conceived set of metrics in Excel that, theoretically, was going to start getting updated on a weekly basis. We'd explain better ways to measure what mattered rather and how to capture the data without relying on human intervention. And then we'd be patted on the heads and happily ignored. It's not that we could have revolutionized how business was done, but given a couple of weeks and a bit of digging, the company could have had an aid in future budgeting: "on projects of this size with this client, we tend to go over budget by x%" isn't a crystal ball that will fix all budget estimation problems, but it's a heck of an improvement.

I mention this not to grind an ax, however pleasurable, but as a caution the next time someone wants to roll a bunch of measurements into a "dashboard" on your web site. A dashboard is a great concept, but if your car reported the status of the windows, car locks and the current time in a great big analog clock rather than current speed, gas stank status and engine temperature, you'd be driving to a hell of an accident,  regardless of where you meant to go. Make sure you know what you're measuring, why you're measuring it and what it means. If you don't know how to filter out noise or when correlation means something, there might be someone around the office who does. Just don't immediately look for a suit and tie.