Noise is Daniel Kahneman’s last major book, co-authored with Olivier Sibony and Cass Sunstein. Where Thinking, Fast and Slow focused on bias (systematic errors in judgment), Noise focuses on the other half of the error equation: the random scatter in judgments that should be identical. When two judges sentence the same offender to wildly different terms, when two underwriters price the same risk differently, or when two doctors give different diagnoses for the same set of symptoms, that variability is noise.
The book’s most striking contribution is demonstrating how much noise exists in professional judgment. Kahneman and his co-authors present research showing that insurance underwriters assessing the same case vary by over 50% in their premium estimates, that judges’ sentencing varies with factors like the local football team’s win-loss record, and that forensic scientists examining the same fingerprint reach different conclusions depending on what they were told beforehand.
The distinction between noise and bias is the conceptual core. Bias is a systematic lean in one direction; noise is random variability. Both produce errors, but they require different fixes. Reducing bias means changing how people think. Reducing noise means standardizing processes: using checklists, structured decision protocols, and independent assessments before aggregating opinions.
For founders, the practical implications are significant. Every company makes judgments: hiring decisions, pricing decisions, customer support responses, performance evaluations. If different people in your organization would make different calls on the same input, you have a noise problem, and the fix is not training (which addresses bias) but process design (which addresses noise). The book’s concept of a “decision audit,” where you test how consistent your organization’s judgments actually are, is something most companies could benefit from.
The writing is clear but the book is long, over 450 pages, and some readers find it repetitive. Kahneman’s co-authors bring expertise in organizational behavior and law that add breadth, but the prose lacks the personal touch of Thinking, Fast and Slow. Still, the central concept is important and underdiscussed, and the book makes a strong case that noise deserves as much attention as bias in any conversation about improving judgment.
