The Cobra Effect
When the reward is the problem in disguise — a field study in incentives that bite back.
Delhi, under British rule, had a snake problem. The administration’s solution was the kind that looks airtight on a whiteboard: pay a bounty for every dead cobra, and the cobras will dwindle. For a while the ledgers agreed. Carcasses came in, coins went out, and someone, somewhere, marked the initiative a success.
Then the curve bent the wrong way. Enterprising residents had begun breeding cobras — not to clear them, but to kill them for the bounty. When the government caught on and scrapped the program, the breeders released their now-worthless stock. The city ended with more cobras than it started with. The cure had quietly become the disease.
Every incentive is a hypothesis about human behavior. Most of them are wrong in ways you only discover after you’ve paid for the answer.
The trouble was never the snakes. It was the measurement. A bounty rewards dead cobras presented, not fewer cobras living. Those two things look identical until someone notices the gap between them — and steps into it.
The shape of the mistake
This pattern has a signature, and once you’ve seen it you start seeing it everywhere. A support team measured on tickets closed learns to close fast, not to solve. A sales floor paid on bookings learns to book deals that never renew. A newsroom paid on clicks learns which lies travel fastest. In each case the metric is a faithful servant of exactly the wrong master.
The three-pipe move is to ask, before you reward anything: what is the cheapest way for a clever person to satisfy this number without delivering the thing I care about? If you can’t answer, you haven’t finished designing the incentive — you’ve just scheduled your own cobra problem.
So name the outcome, not the proxy. Watch the behavior the number produces, not the number itself. And when the carcasses keep arriving but the streets feel no safer, believe the streets.
