How to kill cobras, without breeding snakes
In the late 1800s, Delhi tried to solve its deadly cobra infestation with a simple incentive: pay citizens for every dead snake delivered. It worked - until it didn't. People began breeding cobras to harvest the bounty. The first practical antidote to Goodhart's Law, and a competitive advantage no CEO can afford to ignore.
In the late 1800s, Delhi tried to solve its deadly cobra infestation with a simple incentive: pay citizens for every dead snake delivered. It worked - until it didn't. People began breeding cobras to harvest the bounty. The first practical antidote to Goodhart's Law, and a competitive advantage no CEO can afford to ignore.
In the late 1800s, Delhi had a problem most cities only deal with in movies.
Cobras.
Cobras slipped through courtyard cracks, coiled beneath carts, and appeared where people least expected them - kitchens, bathhouses, government offices.
In a world before antivenom, a cobra strike wasn't a medical emergency. It was a death sentence.
So the British colonial administration did what many well-meaning organizations still do today: they created a metric and paid people to hit it.
Cash. For every dead cobra delivered to the authorities. Rational. Simple. Measurable. And for a while, it worked. Dead cobras piled up. Officials congratulated themselves. Undoubtedly someone wrote a glowing report.
Then came the twist.
The numbers kept rising, and rising, and rising.
Delhi wasn't running out of cobras. It was getting more of them.
Investigators soon discovered the truth: people had started breeding cobras - entire cottage industries of carefully raised, soon-to-be-"hunted" snakes.
And even worse, the moment the government discovered the scheme and canceled the bounty, breeders simply released their now-worthless cobras into the city.
The snake population exploded.
Goodhart's Law: The Metric that Eats Itself
Charles Goodhart wasn't thinking about snakes at all.
As an economist at the Bank of England in the 1970s, he watched a different kind of creature misbehave: financial markets.
Policymakers had identified tidy correlations between money supply, inflation, lending, and growth and turned those correlations into hard targets. But the moment a metric became a target, banks and traders changed their behavior to hit the number rather than reflect underlying reality.
The dynamics broke, the models collapsed, the policies failed.
Goodhart articulated the pattern:
When people optimize for the metric, the metric stops measuring what mattered in the first place.
This is Goodhart's Law. And it's everywhere.