Compression: It Has Its place and HR Isn't It
April 1, 2019
Congrats! You're a 3
The late nights and weekends, the big deals closed, the mentoring, the treats you baked for the team...how did all that become a 3?
In complex systems theory there is a concept called compressibility - a measure of how easy it is to summarize something. Some things in business are very compressible; financial performance can be easily summarized into metrics like profit or earnings growth.
Some things in business are not compressible however. When you try to summarize something as complex as human endeavor you have to be very careful that you don’t end up just removing the very information you’re trying to summarize.
Think of It This Way
You could choose between two divisions to invest in by using just a handful of financial metrics (because financial performance is compressible).
Now do this one: You know two candidates for a role both scored 3s, so who do you promote?
In so many cases a desire to build metrics in HR actually just ends up removing information from the organization (and demoralizing people along the way).
HR Does Silly Things with Metrics
- OVERSIMPLIFICATION. Trying to compress something that can’t be easily summarized (like giving someone a "promotability" rating out of 5)
- SUBSTITUTION. Reporting on something that isn’t easily measured by substituting something that can be measured (like using absenteeism rates as a proxy for staff engagement).
- OBFUSCATION. Failing to produce any real metrics useful for decision making, hiding behind the excuse "we are dealing with people". After all, everyone is special...
- NORMALIZATION. Using parlor statistics to create a false air of respectability. Like trying to normalize staff performance under a bell curve - eliminating information and variability and also ignoring the fact that things change over time.
This last one seems to be peculiar to HR. Nobody ever says: "What profit did we make this year? Same as last year - once we normalized it!"
Why do we do these things? Because in order to make predictions, HR has been forced to try and report the way finance does - using simple topline metrics. But finance can create these metrics because most of their system is compressible. HR’s isn’t.
Should We All Give Up and Go Home Then?
It’s not that bad, really. Why do we want metrics? So we can make better decisions. Which means that it’s not about metrics at all; it’s about modeling and predictability.
The weather is a complex system, but it can be predicted (to an extent!) Just as the weather has recurring patterns that emerge out of the chaos, so too does the behavior of people in an organization.
And the good news is there is a model for this in business already. Marketing. Go and talk to those folks - they’re also dealing with human complexity and they’re pretty good at finding predictable patterns in a chaotic market (they don’t get paid otherwise!)
Two Simple Things You Can Do
- Make sure that you record all the good stuff - interactions and context - as part of your performance reviews
- Roll all of that information up in a way that you can use to identify patterns over time and make predictions. You’ll need a killer performance management system for that!
After all, even if a system is complex, you can still make predictions.
Photo: Creative Commons