The 'Cobra Effect' and when government incentives backfire

The British in India once tried to clear out cobras by paying a bounty for every dead snake. It’s the kind of logic that sounds great over a beer but falls apart the second it hits the street.

Locals started breeding cobras in their sheds like a backyard DIY project just to cash in. When the government scrapped the payments, the breeders simply opened the cages and let them go.

The city ended up with more snakes than before. That’s the Cobra Effect: when a reward accidentally makes the problem you're fighting much worse.

Hang on, did they seriously think people wouldn't just farm the snakes?

Look, when you set a target, people will find the shortest path to hit it, even if it wrecks the original plan. It’s like telling a builder you’ll pay him per nail he drives in; you’re going to end up with a house that’s 90% metal and 10% wood.

The British fell into the trap of measuring the wrong thing. They measured "dead snakes collected" instead of "snakes remaining in the wild." Once a metric becomes a reward, it stops being a good metric because everyone starts gaming the system.

It’s basically human nature. If there’s a buck to be made by taking a shortcut, someone’s going to grab a hammer and start swinging.

Is there actually a way to set targets that people can't game?

You’ve got to stop looking at just one dial on the dashboard. If you only watch the speedometer, you’ll hit the speed limit but blow the engine because you ignored the temperature gauge.

The trick is using 'balancing metrics.' If you pay a crew for how fast they lay bricks, you also tell them they don't get a cent if the wall isn't perfectly level. You pit their speed against the quality of the work.

It’s about measuring the actual result, not just the busywork. You stop counting the number of shovels moving and start looking at whether the hole is actually getting dug.

But won't people just find a way to cheat both metrics at once?

Spot on. Humans are basically Olympic-level athletes when it comes to finding loopholes. If you give someone two rules, they won't just follow them; they'll look for the exact spot where they can do the least work while technically satisfying both.

It’s a classic cat-and-mouse game. There's a famous idea that once a specific measurement becomes the goal, it stops being a useful way to track progress. You can keep adding rules until your spreadsheet looks like a flight manual, but eventually, the cost of policing the system becomes more expensive than the actual job.

At some point, you have to stop tweaking the dashboard and look at the culture. If the only incentive is hitting a number, people will treat your metrics like a puzzle to solve rather than a job to do.

What does 'fixing the culture' actually look like on a job site?

It’s about moving from 'compliance' to 'ownership.' In a small family business, you don't need a 50-page KPI manual because everyone knows if the roof leaks, the whole family gets wet. They have 'skin in the game.'

If a mechanic cares about the car, they’ll fix the engine right because they don't want it to explode, not because a sensor is tracking their wrench turns.

When people care about the finished house, not just the nail count, they start policing themselves. You trade soul-crushing surveillance for old-fashioned professional pride.