The Perfect Store

The single most important thing Omidyar built was not the auction software. It was the Feedback Forum. After every transaction, buyer and seller could rate each other — positive, negative, or neutral — and leave a line of comment. Those ratings accumulated into a public score attached to every username, visible before anyone bid a cent. A seller with nine hundred positives and no negatives was a known quantity. A brand-new account with a suspiciously cheap Rolex was not. Reputation, on eBay, became currency.

This was a quietly radical solution to the oldest problem in trade among strangers: how do I know you won't cheat me? For most of history the answer was proximity — you dealt with the merchant in your town because he'd be there tomorrow and your neighbors would hear if he robbed you. eBay had no town. So it manufactured the equivalent. Cheating someone now cost you your rating, and your rating was the thing that let you keep selling. Honesty stopped being a virtue and became, for anyone who wanted to stay in business, simply the smart play.

Trust, once built, was easy to spend and hard to replace, and eBay's hardest years were the ones where the technology strained. The most dangerous stretch came in June 1999, when the site crashed and stayed down for twenty-two hours. For a marketplace, going dark meant every running auction froze, every seller lost sales, and every buyer wondered whether the whole thing was as solid as they'd assumed. The stock fell hard. The engineering had simply not kept pace with the growth, and for a company whose entire value was reliability, an unreliable site was an existential threat.

Whitman's response revealed what kind of company eBay had become. Rather than issue a terse corporate statement, executives got on the phones with users, credited fees, and treated the outage as a broken promise to a community rather than a service interruption to be managed. They over-invested in the technical backbone afterward, hiring aggressively and rebuilding the infrastructure so the site could carry loads no one had planned for. Underneath the folksy tone was a serious operating discipline.

Step back from the auctions and the Pez and what eBay actually solved was the problem the early internet kept tripping over. The web made it trivially cheap to connect a buyer in Ohio with a seller in Portugal. What it could not do, on its own, was make either one believe the other would hold up their end. Everyone in the 1990s could see the connectivity. Very few figured out the trust. eBay's real invention was not a way to hold auctions online — it was a way to make strangers accountable to one another at internet scale.

The mechanism was almost boring in its elegance. Give every participant a persistent identity, attach to it a public record of how they've behaved, and let that record be the thing they can't afford to burn. It is, in a sense, how reputation has always worked in small communities — but eBay engineered it into software and thereby scaled it to millions of people who shared no geography, no language, no prior contact. Reputation became portable, legible, and permanent, and on that foundation a genuine economy could stand.

eBay never did become the thing it feared — a warehouse, an inventory, a store with its own shelves. It stayed, at its core, the odd little experiment Omidyar coded over a holiday weekend: a place that owned nothing and sold nothing, and made its money renting out the trust it had learned to grow. The broken laser pointer that a stranger wanted anyway turned out to contain the whole idea. People will trade almost anything, with almost anyone, if you give them a good enough reason to believe.