Goodhart's Law
A measure can lose value as an indicator when people are rewarded for maximizing it.
The Mechanism: Goodhart's law is commonly summarized as a measure ceasing to be a good measure once it becomes a target. The precise effect depends on incentives and the relationship between the metric and the outcome people actually care about.
The Philosophical Pivot: Measurement does not merely observe a system; it can reorganize the system around what is counted. People respond to incentives, sometimes improving the number while degrading the underlying purpose.
The Horizon: Institutions need multiple indicators, independent audits, and qualitative judgment to reduce gaming. Yet more metrics can create more opportunities for distortion if goals remain unclear.
Not every target corrupts every measure. The law is a warning to monitor feedback, not a reason to reject measurement entirely.
What gets counted matters, but so does what the act of counting changes.
The moment a number becomes a goal, people begin changing the world around the number.
What outcome am I trying to improve, and how could the metric I chose tempt people to fake the improvement?