How do strangers earn the right to trade?

In February 1996, Pierre Omidyar addressed the people using his young online auction service. His letter, preserved by eBay, described a marketplace that had grown beyond his expectations. He had personally handled complaints while developing the service. The next step was a public Feedback Forum where participants could describe their experiences. The decision was modest in technical appearance: let users comment on one another. Its significance was institutional. Information about a trading partner could begin to travel beyond the two people involved in a transaction.

The letter is primary evidence of both the intervention and the founder's stated conviction. Omidyar expressed confidence in people's general honesty while acknowledging that deception existed. He asked users to participate, praise good conduct, and make complaints visible. That account should not be read as proof that most transactions were safe or that public feedback eliminated fraud. It shows how the founder proposed to address a problem. The marketplace would enlist its participants in producing information that its central operator could not observe completely.

A listing can describe an object, but it cannot fully establish whether the object will arrive, whether its condition is accurately represented, or whether the buyer will pay. Reputation offered a way for previous interactions to influence future choices. The stakeholders' interests were related but not identical. Buyers wanted credible warnings. Sellers wanted confidence and a fair chance to establish themselves. The marketplace wanted participation and completed trades. A dissatisfied participant might want redress, while another might prefer to avoid a public dispute altogether.

Research by Paul Resnick and Richard Zeckhauser examined eBay's reputation system rather than assuming its presence made it effective. Their study found that reputation profiles contained information about later performance, while also showing limitations in the simple summary measures. The important implication is that a visible score and a useful signal are not the same thing. The design must be evaluated against what people actually report, what they omit, and how later participants interpret the resulting record. The score is a measurement system with incentives attached.

Subsequent research on feedback mechanisms examined reciprocity and the possibility that fear of retaliation distorts reviews. The historical record also documents changes from the earliest open commenting arrangement toward transaction-linked feedback. These developments complicate the easy story that eBay solved trust by adding ratings. Ratings introduced a mechanism that could be studied and revised. A reputation system can reduce one uncertainty while creating another, including whether a positive review reflects a good experience, courtesy, strategic exchange, or reluctance to provoke a negative response.

The teaching interpretation is that trust must be supported by decisions about information and consequences. Who can leave feedback? When is it revealed? Can it be challenged? How are newcomers treated? A platform choosing these rules is shaping behavior, not merely displaying user opinion. For a new marketplace, the first experiment might be small enough to observe directly: follow a set of completed transactions and compare the public record with private accounts of what happened. That would test the signal rather than celebrate the existence of a feedback feature.

eBay's 1996 move remains useful because it made a hidden part of the transaction visible without claiming perfect knowledge. It transferred some information production to users, while leaving the operator responsible for the rules. The broader lesson is not that every community needs a five-star scale. It is that repeated exchange among strangers requires a credible way to learn from prior behavior. The case concerns the development of that mechanism, not a blanket endorsement of every later eBay policy or a claim that reputation can replace dispute resolution and enforcement.

For a discussion tailored to your team, explore a workshop with Julien.