The Tale of Useless Nails

In the early years of the Soviet Union, the country faced a desperate shortage of nails. The civil war had ended, and everything needed to be rebuilt: houses, bridges, railways, factories. Nails were essential and nails were scarce. Soviet planners, who believed that any problem could be solved with the right quota, introduced a straightforward incentive: factories would receive bonuses based on the number of nails they produced.

The factories delivered, spectacularly.

They produced enormous quantities of nails. Tiny, delicate, almost microscopic nails. Nails so small they were useless for construction, useless for carpentry, useless for anything except meeting a quota measured in units. The factories had optimized perfectly for the incentive and completely missed the point.

The planners noticed. They adjusted. The new incentive measured output by weight instead of quantity. Surely this would fix the problem. Factories would now have to produce nails of substance.

The factories adapted again. This time, they produced enormous nails. Heavy, oversized, absurd nails that no carpenter could use and no wall could hold. The quota was met. The shortage persisted.

The story of the Soviet nails is often told as a parable about central planning, and it is. But it's also a parable about something more universal: the gap between what you measure and what you want. The planners wanted useful nails. They measured quantity, then weight. At no point did the measurement capture usefulness, so at no point did the factories produce it.

This is Goodhart's Law in its purest form. The metric isn't the mission. It's a proxy for the mission, and the moment people start optimizing for the proxy instead of the thing it was supposed to represent, the proxy becomes meaningless. The Soviet nail factories didn't fail. They succeeded spectacularly at the wrong thing.