Strategic Economics

Reputation Systems and the Economics of Trust

Treat a rating system as an incentive mechanism with a computable budget rather than as a feature — and learn to read the silence in your feedback data, which carries more information than the stars do.

  • Advanced
  • 13 min total
  • 13 chapters

What decision this helps you make: Whether your reputation system is currently strong enough to make honesty the rational choice for the participants you are worried about — and which of its four levers to move if it is not.

What this topic is

A reputation system is the machinery a marketplace uses to make strangers safe to transact with. It works by turning a one-off interaction into a repeated one: your behaviour today is visible to everyone you might deal with tomorrow, so cheating stops being profitable. That is the entire mechanism, and it means a rating is not a description of quality but a price on future misbehaviour.

Why it matters

Trust is the product a marketplace actually sells. The fee it can charge, the transactions it can support, and the range of participants it can admit are all bounded by how safe it makes dealing with someone you have never met. And the mechanism has precise failure conditions — it stops working for anyone who is about to leave, for anything the buyer cannot observe, and for any market where honest reporting is not itself in the reporter's interest.

Who should learn it

Anyone designing or fixing ratings, reviews, verification or dispute processes on a marketplace; operators trying to understand why their ratings are all five stars and their complaints are rising; and sellers deciding how much a reputation is worth defending.

What you will understand

  • The honesty condition: the exact comparison a participant makes between cheating now and earning later
  • The four levers a platform has to widen that gap, and which are cheapest to move
  • Why silence in your feedback data is a stronger signal than the stars, and how to use it
  • The endgame problem: why reputation cannot discipline anyone who is leaving, and what to use instead

Prerequisites

Common misconception

"Our average rating is 4.8, so quality is high." The average rating on almost every mature platform is high, and it is high for reasons that have very little to do with quality: unhappy participants disproportionately say nothing rather than complain, reciprocal rating invites retaliation, and everyone involved understands that a middling score functions as a penalty. What carries information is not the stars but the pattern around them — who declined to rate, who never returned, and which transactions produced no feedback at all. A platform reading its own averages is reading the one part of the data that was designed to be flattering.