Goodhart’s Law
Goodhart’s Law shows that a measure and a target are not the same thing, and treating them as the same thing is what breaks the measure.
What Is It?
Goodhart’s Law states that when a measure becomes a target, it ceases to be a good measure. As long as a statistic is simply observed, it reflects the reality it was tracking. The moment people are rewarded or punished for moving that statistic, they start optimizing for the number instead of the underlying thing it was supposed to represent, and the number stops meaning what it used to mean.
British economist Charles Goodhart introduced the underlying idea in 1975, in a paper on UK monetary policy, originally phrased as: “any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes.” Anthropologist Marilyn Strathern later condensed it into the form now commonly quoted.
Why Does It Matter?
Almost every organization has a version of this: a support metric that gets hit by closing tickets faster instead of resolving them, a sales target that gets hit by pulling deals forward instead of creating new ones, a safety statistic that improves because incidents stop getting reported rather than because they stop happening.
None of this requires bad faith. It’s the predictable result of attaching consequences to a number that was only ever a proxy for the thing leaders actually cared about. The number was never the goal. It just used to correlate with the goal, before anyone was managing to it directly.
What Changes Once You See It?
You stop assuming a metric that’s improving means the underlying reality is improving.
You start asking what behavior this specific number rewards, separate from what it was originally designed to reflect.
You start treating metrics as flashlights that help you see, rather than hammers you use to drive behavior directly.
Common Misunderstandings
- It is not an argument against using metrics at all. It’s an argument against using a single metric as a direct control lever without expecting it to get gamed.
- It doesn’t require intentional gaming. People often shift behavior to hit a target without consciously deciding to game anything; the incentive does the work on its own.
- It is not limited to economics, where it originated. It applies anywhere a proxy measure gets turned into a target with consequences attached.
- It doesn’t mean the original measure was a bad one. Often the measure was a perfectly good indicator right up until it became a target.
Diagnostic Question
Is this number still telling us something true, or have we been managing directly to the number long enough that it’s stopped correlating with what we actually care about?
Explore Further
Field Notes
None yet.
Origin
Charles Goodhart introduced the underlying principle in a 1975 paper on UK monetary policy. The now-familiar phrasing, “when a measure becomes a target, it ceases to be a good measure,” was popularized by anthropologist Marilyn Strathern in 1997.