Every reward structure drifts toward rewarding itself. Test it against what it was built to produce, not against how long you have used it.
Every organization pays people for something. A bonus formula, a performance rating, a promotion rubric, a sales target, a scorecard of numbers someone chose years ago. Ask why a particular measure decides who gets paid and who gets promoted, and you usually meet four words that should never be enough: this is how we measure it. Inheritance is not a reason. It is the absence of one, wearing the costume of tradition.
First Principles is the principle, and it applies to the reward system specifically: every structure that decides who gets paid, promoted, or praised should be tested against what it was built to produce, not kept because it was inherited. Periodically, someone has to ask of each measure and each incentive what behavior it exists to encourage, and whether it still encourages it.
Because every reward structure rots the same way. A measure is chosen because it tracks something you care about. Then people are paid on it, and the measure becomes the thing they produce. The number everyone has learned to game outlives the behavior it was built to encourage, and the gaming becomes the job. Nobody decided to reward the wrong thing. The reward simply kept running after its purpose had quietly left the building.
So take the measure that matters most to pay and promotion where you work and ask the plain question: what was this built to produce, and is it still producing that? If the honest answer is that people have learned to hit the number without delivering the thing, then what you are paying for now is skill at the measure.
First Principles is what keeps Reward from decaying. Substance Over Signal decides what you pay for, and Costly Action tells you what to trust. Both can be built into a sound system, and both slowly fail if nobody goes back to check whether the measures still measure. That is why this principle covers rewards rather than every structure in the building. Testing the whole architecture is the work of Structural Reform, which closes this series.
The Precedent
In 1902, colonial Hanoi had a rat problem, made worse by the modern sewers the French administration had just built, which turned out to be ideal rat country. Paid crews could not keep up, so the authorities opened the work to everyone with an incentive: a bounty of one cent for every rat tail handed in. The tails arrived by the thousand. So, before long, did reports of rats running through the city without them. Hunters had worked out that a tail was worth more than a dead rat. They cut it off and released the rat to breed, and health inspectors found people farming rats on the edge of the city. The bounty was scrapped. It had been designed to reduce the number of rats. Measured faithfully, it was producing them.
In Other Words
“When a measure becomes a target, it ceases to be a good measure.” Marilyn Strathern, summarizing Goodhart’s law
“The difficulty lies not so much in developing new ideas as in escaping from old ones.” John Maynard Keynes, The General Theory
The Test
Which measure decides who gets paid and promoted here, what was it built to produce, and is it still producing that?
Next: the first of the two Bindings, The Integrated Leader. The person the whole architecture requires.
This is one principle of Trust Architecture, a framework for building organizations where the truth survives contact with power. New here? Start with The Framework; the full definitions live in The Principles.
Free to read and share under CC BY-NC-ND 4.0. Framework and essays © 2026 Ankush Chowdhary. See Ethics and Transparency for how this work is produced.

