Something to readReviews & performance
Noise
A book about a flaw in human judgment the authors call noise, and why two managers can rate the same work so differently
Why it’s worth your time
When two managers would rate the same work differently, part of your review is luck, and this book explains where that luck comes from. Daniel Kahneman, Olivier Sibony and Cass R. Sunstein call it noise, the unwanted scatter in judgments that should match, and trace it through courts, medicine and hiring to performance reviews, where knowing how it works can take some of the sting out of a rating.
Why performance ratings are inconsistent
The book · About 10 hours of reading · Buy or borrow
How to get it
We don’t earn anything from these links.
- Start with the part on performance ratings.
You don't need the whole book for work. The discussion of ratings explains why different raters use the same scale differently; read it first, and go back to the opening chapters on what noise is if you want more.
- Ask how your rating gets decided.
In your next 1:1, ask who besides your manager has a say in your rating, what they look at, and when it happens.
You“When ratings get decided, who else weighs in, and what do they look at? I'd like to make sure the right evidence is in front of them.”
Good moments for this: before a reviewafter a rating that surprised youbefore calibration season
Keep going with Truest
Small actions add up. Truest helps you find the next thing worth doing, make the space to do it, and keep a record of your progress. It’s a private place to work on your career, and it’s yours, not your employer’s.
- It actually happensWe’ll put it on your calendar and remind you before it starts.
- A living page in your VaultKeep it as one page you add to, next to your reviews and your files.
- It’s yours to keepEach thing you do is saved to your private record, there when your next review or job search comes around.
Anything else we can help you with this week?
Making it work for you
Every workplace is different. Here’s what to think about before you start, and what might get in the way.
Things to think about
Much of the scatter the authors describe is random rather than personal. That doesn't make it acceptable, but it can make a surprising rating easier to discuss calmly.
What might make this harder
Read the ratings discussion and a review or two. The core idea, that judgments which should match often don't, carries most of the value for work.
Questions people ask
What is noise in decision making?
Unwanted variability in judgments that should be the same, such as two managers rating identical work differently, or one manager rating it differently on different days. The authors contrast it with bias, which pushes judgments consistently in one direction.
What does Noise say about performance reviews?
That ratings are often noisy because raters use the same scale in different ways, so a rating can say a lot about the rater. The authors discuss remedies such as relative comparisons, shared reference cases and combining several raters.
Is Noise worth reading?
If your interest is work, the parts on ratings and hiring carry most of the value, and you may not need the rest.
How can I protect myself from noise in my review?
You can't control the raters, but you can give them solid ground: a written record of what you delivered, ready before the process starts, and a clear idea of who else will weigh in.
What it says, and how it holds up
Unfair judgments are easy to think of as bias, a consistent lean in one direction. Kahneman, a psychologist and Nobel laureate in economics, Sibony, a strategy professor at HEC Paris, and Sunstein, a Harvard professor, wrote about its quieter cousin, which they argue people and organizations tend to neglect. Noise is the scatter you'd see if the same case went to different judges, or to the same judge on a different day. Their examples range from sentencing and medical diagnoses to insurance quotes, job interviews and performance reviews.
At work, the ratings discussion is the part that lands. A review in The Psychologist points to its finding that appraisals often track the person rated only weakly, partly because raters don't share a sense of what good or great means. The remedies the authors group under decision hygiene include structuring judgments, comparing rather than scoring in isolation, and combining several independent views. From the employee's chair, that explains why calibration meetings exist, and why the evidence in front of the raters matters so much.
Reducing noise has costs of its own. Writing for the LSE's US politics blog, Kaibalyapati Mishra points to them, from expensive audits to rules that override judgment that matters, and draws on Gerd Gigerenzer's view that mental shortcuts don't always lead to biased or noisy decisions. Kahneman and colleagues also wrote a 2016 Harvard Business Review article on noise.
When the scatter has a direction, it isn't noise. If you see a pattern, such as people like you consistently rated lower, that's bias, and it calls for notes and advice rather than statistics. If you give ratings yourself, the book may be more directly useful still, since it's written as much for the people running these processes as for those on the receiving end.
You might like
More in Reviews & performance
See allRead this alongside your role page
Tell Truest what you’re dealing with
We read every one, and we write guides for the things that come up most. Leave your email if you’d like to hear when we do.
Picked by Truest and described in our own words. The original belongs to its creator. Last updated October 9, 2026. We sell a career membership; where that’s relevant above, we say so.