The Devil Is In The Metrics
In many of my posts I have expressed my belief that true performance improvement comes from changing behavior. I’ve also invested many words on my own site on the topic. Accordingly, I’ve mentioned many times that metrics will drive behavior. Our goal is to identify or create metrics that deliberately drive desired behavior.
Unfortunately, many of those assertions are not followed with solid examples of how metrics and behavior work together. So, herein is a discussion of metrics and behavior, and some thoughts for how to manage the phenomenon that occurs between them. The example is so common and widespread I imagine that everyone has witnessed it.
Consider the practice of using Corrective Action Reports (CARs) to identify and track process “escapes” or problems that result in undesired outcomes. Whatever we call our identifiers, most of us with any formal process control or process improvement methodology use something to draw attention to the fact that a process or procedure failed to work or we failed to follow it.
Naturally, we keep metrics on these reports. We want to know how many reports are made every month, where they are coming from, and we want to know that issues or weaknesses in our methods are being addressed. We want to know if more problems are occurring more frequently, and we hope that fewer problems are occurring over time and that we are ultimately controlling or improving our methods.
These are important business and process improvement questions. They are a good idea. Unfortunately, our act of tracking and measuring them, our metrics, do not always work in our favor.