Tuesday, March 3, 2009

ACTION: Go identify your multiple. Now.

Two posts ago, I buried a key idea in a lengthy list.

Here's the gist. When the managers come around trying to figure out who to "re-engineer", would it be helpful to be able to say, "Guys, the firm invested in our team about $1 million last year...and got about $2 million back NPV"...might that allow you to relax if you could say that? (NPV means Net Present Value...your Product Owner should know this stuff.)

By multiple, I mean the relationship between NPV and cost. Usually the multiple is between 3 and 20 (or should be).

OK. Action item. For today. Go talk to your Product Owner and estimate the NPV of your last project as a team or your current set of work (say, next 2 releases). Annualize it (managers think more easily in those round numbers).

Much better if the NPV estimate is based on actual results rather than someone's dream of what the benefits "ought" to be.

Yes, there are lots of issues. Yes, life is difficult to measure. Get the best approximation you can.

Example: "We have a risk project. How can you measure NPV for risk?" Umm, who is really good at putting a price on risk? Auto insurance anyone? Go talk to an actuary and let 'em help you figure it out.

Yes, there are lots of problems with these numbers. Heck, there are lots of problems with helmets in the NFL, but I would not recommend playing without one.

No comments: