Metrics: The Measurements that Control the Masses

Metrics

Photo by Stuck in Customs

We are being controlled.  Of our own free will of course, but nonetheless, our thinking is completely manipulated by society’s chosen metrics.  If society defines inflation as the CPI Index, everyone defines inflation that way.  If society defines growth as an increase in GDP, everyone defines growth that way.  We are all puppet’s to society’s metrics.  And if we try to escape it, who will understand what we’re talking about, anyway?

Of course, there are occasionally people who create their own metrics – people who take the time to figure out what drives the results they want to achieve – people who know the system isn’t quite right.  And those people are usually rewarded. 

Very often, what people in life do is satisfice.  If Herbert Simon were still around, he would tell us all about this.  If a metric meets an invisible “good enough” threshold, it’s packaged for consumption by everyone – this is satisficing. 

But satisficing will not give you the metrics you want.

What you really want to do in creating appropriate metrics is to first use signal detection theorySignal detection theory is all about separating the wheat from the chaff – figuring out what’s important and what’s not important – otherwise known as “Signal vs. Noise.”  This is easier said than done.  I know of no definitive way to do this.  Andy Grove discusses this exact issue in Only the Paranoid Survive, and concludes that there’s no readily available way to deciper signal from noise.  

But just because there isn’t a definitive way to accomplish this, doesn’t mean there aren’t methods that are more useful than others.  I’ve found that separating signal from noise is usually best achieved through a combination of statistical and natural inference.  By that, I mean combining quantitative data with qualitative observations to form a theory, and then using that theory as the basis for establishing metrics.

Metrics are usually created in the hopes of gaining better control over behavior - something appears important to measure, a metric is established for it, and then all efforts are focused on using that metric to control behavior.  This is not necessarily a good thing.  The opposite of control is chaos, and chaos is far more likely to be a reality than control.  There will always be chaos – chaos in life, chaos in business.  There will always be more moving parts than you can comfortably monitor.  So, it’s better to just accept this and move on.  As Andy Grove says, “Let chaos reign.  And then reign in chaos.”

Metrics should be designed with the implicit understanding that they are merely assumptions.  The more assumptions you rely on, the less likely you are to have a true understanding of what’s actually taking place.  Instead of trying to control a situation, try to understand the underlying behaviors that create the situation.  Instead of 30 metrics, use 3.  Eliminate any metric that can’t be gathered within 24 hours of the completion of the action you’re monitoring.  Figure out what matters, and focus intently on that.  Forget about everything else.

Metrics are situational, like much of life.  I can’t give you a set of universally applicable metrics, simply because one doesn’t exist.  Metrics are unique to companies, individuals, environments, etc.  And figuring out the right ones to use requires work.