You'd think measuring things would make them better. Count the surgeries, track the test scores, tally the arrests—surely numbers bring accountability. Yet something peculiar happens when we start measuring public services. The measurements themselves become the goal, and the actual service quietly slips away.

This is one of government's great paradoxes: the harder we try to measure success, the more we distort what success actually means. Those impressive statistics you see in agency reports? They might tell you everything about the measurement system and nothing about whether citizens are actually being served.

In 1979, social scientist Donald Campbell articulated something government workers had known for decades: "The more any quantitative social indicator is used for social decision-making, the more subject it will be to corruption pressures and the more apt it will be to distort and corrupt the social processes it is intended to monitor." That's a mouthful, but the idea is elegant. Once you decide to judge someone by a number, they'll start optimizing for that number—even if it means abandoning the work the number was supposed to represent.

Consider hospital emergency rooms measured by wait times. The target seems reasonable: patients shouldn't languish in waiting rooms. But watch what happens. Some hospitals began stopping the clock by having nurses do brief initial assessments, technically "starting treatment" while patients still waited hours to see doctors. Others created express lanes for minor complaints, improving average wait times while serious cases sat longer. The measurement improved. The actual patient experience didn't.

This isn't cheating, exactly. It's rational behavior when your funding, your reputation, and your job depend on hitting specific numbers. Campbell's Law suggests the corruption isn't a bug—it's the inevitable consequence of high-stakes measurement. The people gaming the system aren't necessarily bad actors. They're responding logically to the incentives we've created.

Takeaway

When a measure becomes a target, it ceases to be a good measure. The act of measuring changes the behavior being measured.

Here's a quiet scandal in workforce development programs: agencies judged by job placement rates naturally gravitate toward clients who are easiest to place. The single mother with outdated skills and unreliable childcare? She's risky. The recently laid-off professional who just needs resume help? That's your ticket to good numbers. This phenomenon—called "creaming" in policy circles—means the people who need services most become the people least likely to receive them.

The same dynamic plays out everywhere performance is measured. Schools in systems with high-stakes testing have been caught discouraging struggling students from showing up on test days. Police departments measured by crime statistics may pressure officers to downgrade offenses—that robbery becomes a theft, that assault becomes a verbal dispute. Hospitals evaluated on mortality rates may transfer the sickest patients elsewhere. Everyone's statistics improve. The hardest problems remain unsolved.

What makes this particularly insidious is its invisibility. No one announces they're cherry-picking. The selection happens in a thousand small decisions: which clients get callbacks first, which cases get extra attention, which problems get coded as priorities. The people turned away don't show up in the success statistics, so their absence doesn't trigger concern. The measurement system creates its own blind spots.

Takeaway

Performance metrics often reward agencies for avoiding difficult cases rather than solving difficult problems.

The British rail system once set a target: 90% of trains should arrive within ten minutes of scheduled time. Reasonable goal, right? But railway operators found an easier solution than improving service. They extended scheduled journey times. Trains that once officially took 45 minutes were rescheduled to take 55. The trains didn't get faster—the definition of "on time" got more generous. Target met, passengers still late.

This is goal displacement in its purest form: the metric succeeds while the mission fails. It happens when organizations lose sight of why they're measuring something in the first place. The Veterans Administration scandal of 2014 revealed that some facilities maintained two appointment lists—one showing veterans waiting months for care, another showing wait times that met federal standards. Staff weren't hiding problems maliciously; they were responding to a system that punished missed targets more than poor care.

The deeper problem is that measurements always simplify reality. A school's mission includes developing critical thinking, fostering curiosity, building character, and preparing citizens. But we can only easily measure test scores. So test scores become the mission, and everything else—field trips, arts programs, classroom discussions that don't "teach to the test"—gets squeezed out. The metric colonizes the institution, reshaping it in the measurement's image.

Takeaway

When we mistake the metric for the mission, we optimize for the scorecard while the actual work withers.

None of this means we should stop measuring. Flying blind is worse than flying with imperfect instruments. But we might hold our metrics more loosely, remembering they're proxies for what we care about, not the thing itself. The best measurement systems use multiple indicators, rotate metrics to prevent gaming, and pair quantitative data with qualitative judgment.

Perhaps the most important insight is also the simplest: whenever you see a performance target, ask what behavior it might accidentally encourage. The numbers never lie—but they rarely tell the whole truth either.