I was working with a client today whose Strategy Execution framework I'm helping to build and automate. When we were talking about setting good targets for some key performance measures, I was reminded of an important lesson I had learned years ago about effectively measuring service levels: hitting average service level targets is not enough.
Several years ago, I worked for a CRM company that managed the laptops of very large companies' sales forces. If a sales rep had a problem, he or she simply overnighted the laptop to our company, we fixed it, and overnighted it back. We targeted a 4-day turnaround. In reality, we almost always turned around repairs and shipped the unit back in one -- or at most two -- days. Great, right?
Well, sometimes there were problems. Sometimes an incoming unit was put in the wrong pile. Or the rep didn't put the right RMA number on it. Or it was shipped out without a label. Not to worry, though. These kinds of thing were REALLY rare and we were always within our targeted 4-day average.