Management decisions have superlinear returns. So do management mistakes.

An operational reading of Paul Graham's essay: performance isn't linear because systems aren't linear. A wrong structural decision doesn't cost 10%. It costs the entire company.

Paul Graham argues that returns in many areas — technology, science, careers — are non-linear. Small differences in quality at the top produce enormous differences in outcome. The essay speaks mainly about individual performance and startups. But the management version of this principle is harder and less discussed: organizational structure decisions have negative superlinear returns. A wrong decision about who has authority, how incentives work, or which problem the company prioritizes doesn't cost proportionally to the size of the error. It multiplies through every subsequent decision made using the wrong structure as its foundation.

The reason is that organizations are amplification systems. A rule that rewards the wrong behavior doesn't produce one wrong employee — it produces a set of behaviors that, over time, redefines what is considered correct inside that company. An approval process that adds latency in the wrong place doesn't delay one decision — it delays every decision that depends on it. A culture that punishes those who bring bad news doesn't silence one person — it silences the entire early warning system. The damage isn't linear because the structure isn't a sum of independent parts. It's a system where each part affects the others.

The practical implication is that the most important decisions in an organization are rarely the most visible ones. Not the growth targets or the product you're about to launch. They're the decisions about how the company will decide — who has voice, who has veto, what counts as success, what's tolerated and what isn't. These decisions become invisible because they normalize quickly. After six months, they look like culture. After two years, they look inevitable. The superlinearity of the damage begins there: at the moment a bad structural decision stops being perceived as a decision and becomes a premise.

Leo Bentier