Why Goodhart's Law Isn't All That Useful

We’ve all heard it a thousand times:

“When a measure becomes a target, it ceases to be a good measure.”

Goodhart’s Law gets thrown around constantly in tech, business, policy discussions — anywhere metrics and goals are involved (which is everywhere).

But I’m going to argue that while Goodhart’s Law captures an important truth, it’s often misunderstood and misapplied in ways that limit its usefulness as a practical guide for using metrics effectively.

Here’s why:

1. It’s often interpreted as a blanket condemnation of metrics. But Goodhart’s Law is really about the misuse of metrics, not metrics themselves. All metrics can be gamed, but that doesn’t negate their value when used judiciously. The problem arises when single metrics become sole targets.
2. It assumes bad faith. Not all optimization is harmful gaming. Sometimes people pursue metrics because they believe (rightly or wrongly) that the metric tracks the underlying goal. Goodhart’s Law implicitly dismisses this possibility.
3. It ignores the necessity of metrics in complex systems. In any large organization or system, metrics are essential for coordination, incentive alignment, and performance tracking. Goodhart’s Law offers little guidance on how to use metrics well, only a warning about how they can fail.
4. It’s silent on the alternatives. If using metrics as targets is bad…then what? Never have any quantifiable goals? Rely purely on intuition and vibes? The implication is to avoid metrics, but total rejection of measurement is far more dangerous. Good management requires good measurement.
5. It’s often applied selectively. People are quick to criticize metrics they dislike as Goodhart’s Law violations, while rarely applying the same scrutiny to metrics they favor. The law is used as a rhetorical cudgel rather than a consistent principle.

So what’s the alternative? I believe Goodhart’s Law is a useful cautionary tale, but not a complete philosophy of metrics. A better approach acknowledges that metrics are flawed but indispensable tools for coordination and optimization at scale.

Some key principles:

• Expect metrics to be gamed, and design them with aligned incentives in mind
• Use multiple metrics to balance different goals and limit over-optimization
• Continuously evaluate and evolve metrics to keep pace with changing circumstances
• Combine metrics with qualitative measures and human judgment
• Foster a culture of questioning metrics and rewarding the surfacing of flaws

In short: Goodhart’s Law is a warning, not a prohibition. Metrics are like powerful drugs — dangerous if misused, but also capable of enabling incredible feats of coordination and optimization. The hard work is in developing the wisdom and skill to use them well.

So while Goodhart’s Law is a useful check on naive metric-setting, it’s only the beginning of the conversation around building healthy, high-performing organizations and systems. The real challenge is moving beyond simple maxims to grapple with the irreducible complexity of metrics in practice.

Used wisely, measurement is essential to intelligent action at scale. Goodhart’s Law doesn’t change that — it just raises the bar for how thoughtful and vigilant we must be. Let’s embrace that challenge, learn from our failures, and keep striving to measure what matters.