The Measurement Paradox: How Good Metrics Corrupt When Chased Alone

Faced with the failures traced in these essays, a natural instinct is to measure more carefully. If the organization is learning the wrong lessons, or optimizing the wrong things, surely the answer is better instruments: sharper metrics, cleaner dashboards, more rigorous key results. Measurement feels like the responsible, hard-headed response—the opposite of vague good intentions. And it is genuinely necessary. It is also, pursued alone, one of the more reliable ways to make the underlying problem worse.

The mechanism was named long before the dashboard era. In 1956 Ridgway catalogued the dysfunctional consequences of performance measurement, observing that even well-chosen measures distort behavior once people are held to them. Decades later the pattern was compressed into what we now call Goodhart's law, in Marilyn Strathern's crisp formulation: when a measure becomes a target, it ceases to be a good measure. The number and the reality it was meant to represent begin to drift apart the moment the number acquires consequences, and the harder the number is chased, the wider the drift.

The examples are wearily familiar because they are everywhere. Support teams close tickets briskly and resolve problems slowly. Sales hits the quarterly figure by pulling deals forward and mortgaging the next quarter. Teachers teach to the test; hospitals manage waiting-list definitions; researchers optimize citation counts. In each case the metric improves while the thing it was supposed to track quietly deteriorates, and for a while no one notices, because the metric is what everyone is watching.

What turns this from a familiar annoyance into a genuine paradox is the response it provokes. When a metric is gamed, the usual remedy is more measurement—add a counter-metric to catch the gaming, then a third to catch the gaming of the second. The dashboard proliferates. But each new measure is subject to exactly the same law, and the organization now optimizes a larger bundle of proxies, none of which it has stepped back to question. Jerry Muller's account of metric fixation describes the endpoint: an institution expending ever more energy measuring, and ever less achieving the ends the measurement was meant to serve.

The reason more measurement cannot cure the disease of measurement is that the disease does not live in any single metric. It lives in the isolation of measurement from two things it needs and rarely has. The first is a standing willingness to question the criteria themselves—to ask not merely whether the target was hit but whether it remains the right target, a discipline most organizations lack any occasion to exercise. The second is an honest relationship to being wrong, so that a metric can report bad news without triggering the concealment that corrupts data at the source. A measure surrounded by neither of these will decay no matter how well it is designed.

The isolation also distorts what an organization attends to, not merely how it scores. Whatever is measured draws effort toward itself and away from everything that is not, so a metric introduced to illuminate one corner quietly darkens the rest. Important work that happens to be hard to quantify—mentoring, maintenance, the patient repair of relationships—loses ground to whatever the dashboard counts, regardless of which matters more. Adding metrics to cover the neglected corners only widens the field of distortion, since each new measure recruits its own share of effort and invites its own gaming. The organization ends with a richer scoreboard and a thinner grasp of its own priorities, mistaking the precision of its numbers for an understanding of its situation.

So measurement is a genuine good that turns toxic in isolation. On its own it produces the confident illusion of control—every indicator green—while the organization drives harder toward proxies that have quietly detached from the things that matter. Better metrics are not the mistake; treating better metrics as a self-sufficient answer is. The instrument only tells the truth when it sits inside something larger that keeps asking whether the instrument still points at reality, and that makes it safe to say when it does not. Measurement, in other words, is not a remedy you can install by itself. It is a component that misfires unless it is wired to the rest of a system that most organizations have never actually built.