Introduction
Goodhart's Law states, roughly, that when a measure becomes a target, it stops being a good measure. Leaders encounter this constantly, often without naming it: a support team optimizes for faster ticket closure and quality quietly drops; a sales team optimizes for number of calls made and call quality quietly drops. Understanding this dynamic explicitly changes how leaders should think about every metric tied to an incentive.
Why This Happens — It's Not a People Problem
This isn't a flaw in the people being measured, and treating it as one misses the actual mechanism. It's a predictable, rational response to incentive design: when people are evaluated on a specific number, they naturally optimize their behavior to improve that number — sometimes at the expense of the underlying goal the number was originally meant to represent. This happens even among well-intentioned people who aren't deliberately gaming the system; it's simply what focused incentive pressure tends to produce over time.
Common Real-World Examples
- A team measured on ticket closure speed handles issues faster but with less thoroughness, leading to repeat contacts that the closure-speed metric doesn't capture
- A sales team measured on call volume makes more calls, but shorter and less substantive ones, reducing actual conversion even as the primary metric improves
- A content team measured on publishing frequency produces more content, but of declining depth and quality, as the volume metric increasingly dominates.
The Practical Implication for Leaders
Every metric attached to a real incentive — compensation, promotion, visible recognition — deserves more scrutiny than a purely informational metric with no attached incentive. The stronger the incentive, the more likely the metric will eventually be optimized in ways that diverge from its original purpose.
Detecting When a Metric Has Started Producing the Wrong Behavior
- Periodically ask whether the metric still reflects the underlying goal it was designed to represent, rather than assuming a long-standing metric remains valid indefinitely
- Look for signs of narrow optimization — improvement in the specific metric without corresponding improvement in the broader outcome it was meant to indicate
- Ask the people closest to the work whether they've noticed themselves or colleagues adjusting behavior specifically to improve the number, rather than the underlying goal
A Useful Countermeasure: Paired Metrics
A practical way to guard against this dynamic is to pair any metric used for performance evaluation with at least one counterbalancing metric that would catch the most likely way the first metric gets gamed. Pairing ticket closure speed with a customer satisfaction or repeat-contact metric, for instance, makes it much harder to improve the first at the clear expense of the second without the trade-off becoming visible.