Proxy capture is what happens when the metric replaces the thing it was meant to measure.
Every organisational metric is a proxy. Revenue is a proxy for value created. Test scores are a proxy for learning. Compliance rates are a proxy for safety. The proxy is useful — it makes the real goal legible and manageable. The problem arises when the proxy becomes the target, at which point the system optimises for the measure and the underlying thing quietly degrades.
Goodhart’s Law states this directly: when a measure becomes a target, it ceases to be a good measure. The mechanism is not dishonesty. It is structural. Incentives, reporting lines, and performance management all orient around what is measured. Behaviour follows. The unmeasured thing — which may be the actual goal — falls outside the frame of attention.
Examples are everywhere once you see the pattern. A hospital optimises for waiting-list numbers and shifts patients between lists to hit targets while clinical outcomes worsen. A sustainability report hits every metric while the underlying ecological trajectory continues unchanged. A consulting firm delivers a rigorous, well-structured engagement that confirms the client’s existing direction because the brief was designed to look for answers inside the current frame. This is the consulting paradox expressed at the level of the metric: the deliverable — the report, the recommendation, the well-structured deck — becomes the proxy for the diagnostic value the engagement was meant to create.
The deeper problem is that proxy capture is self-concealing. The metrics look good. The governance confirms it. The signal that something structural is wrong arrives not as a clear alarm but as a vague sense that the results are not matching the reality on the ground — which is precisely the fifth symptom of the diagnostic pattern.
Related
- Three mechanisms that keep it in place — proxy capture is one of three self-reinforcing mechanisms that hold frame failure in place
- Overoptimisation — the two mechanisms work together: overoptimisation narrows the system to what is measurable; proxy capture means the measure then replaces the goal
- Separated knowledge — the people who can see the gap between metric and reality are typically not the people with authority to act on it
- Productive collapse — the failure mode proxy capture produces over time
- The diagnostic pattern — proxy capture tends to surface as the fifth symptom: results no longer matching the reality on the ground
- Contextual excess — the accumulated weight of metrics, reports, and governance makes the gap between measure and reality harder, not easier, to see