Lesson
Metrics Change Behavior
A metric can be a useful signal until people are rewarded for moving it; then the signal often gets gamed.
By the Math Says Yes editorial team
Human-reviewed under our source and correction standards.
How we review content
A metric is a proxy
Most important goals are hard to measure directly. Health, learning, safety, quality, and trust are complex, so organizations choose simpler signals: test scores, wait times, ticket counts, conversion rates, readmission rates, or star ratings. A proxy can be useful when it moves with the real goal. The danger begins when people forget it is a proxy and start treating the number as the goal itself.
Targets create incentives
Once rewards, penalties, rankings, or public pressure attach to the proxy, behavior changes. People learn the scoring rule and optimize it. A support team measured only by tickets closed may close easy tickets first. A school rewarded only for test scores may narrow teaching to the test. A hospital judged by length of stay may discharge too early. The measured number improves while the underlying goal may not.
Signs the metric is breaking
A metric is in trouble when it improves faster than the lived experience it is supposed to represent. Watch for sudden jumps after a target is announced, teams arguing over definitions, cases being excluded, easy work replacing important work, or other quality indicators moving the wrong way. These are not just moral failures. They are predictable responses to an incentive system that has made the proxy cheaper to optimize than the outcome.
How to use it
Use metrics as instruments, not masters. Pair each target with counter-metrics that catch obvious gaming, review real examples behind the numbers, and rotate or revise metrics when behavior adapts. Reward judgment as well as movement in the dashboard. Most importantly, keep asking what the metric is supposed to stand for. If the number improves while the real goal worsens, the number has stopped doing its job.
FAQ
Does Goodhart's law mean metrics are useless?
No. Metrics are useful signals when interpreted with context. The warning is that a metric becomes less reliable when people are strongly rewarded or punished for moving it.
How can a team reduce metric gaming?
Use several balancing metrics, inspect individual cases, keep qualitative review in the loop, and change incentives when the metric starts improving without the real-world outcome improving too.
Core Concepts
Framing Effects
The same true numbers can give very different impressions depending on how they are presented: axis ranges, baselines, absolute versus relative, or wording. Framing changes perception without changing the facts, so check how a number is shown, not just what it is.
Goodhart's Law
Goodhart's law says that when a measure becomes a target, it tends to stop being a good measure. People optimize what is rewarded, so the metric can improve while the thing it was meant to represent gets worse.
Quick Check
Why can a useful metric become worse after it becomes a target?
A
People optimize the number itself, not the goal behind it.
B
Targets make people stop caring entirely.
C
Metrics can never summarize anything useful.