Every trust product eventually faces the same question: “can you guarantee there will be no scams?” The honest answer is no. Nobody can. Anyone who tells you otherwise is selling something they can’t deliver.
What you can do — and what we think is the only worthwhile goal — is raise the cost of running a scam until it stops being worth running. That’s the actual design problem. Not zero scams, but expensive ones.
How most scams economics work
A typical online scam looks like this:
- Stand up a profile in 20 minutes. Free, no documents required.
- Stage some assets — bought followers, generated avatars, screenshots from somewhere else.
- Approach a few hundred candidates. Most ignore you. A few engage. One of them pays.
- Disappear. The profile gets reported eventually; you’ve already spun up the next one.
Step one — the cost of creating a credible-looking profile — is what makes the whole loop work. Drive that cost up by an order of magnitude, and the math breaks. Most bad actors don’t care about your specific platform; they care about any platform where steps one and two stay cheap.
What raises the cost
Three things, in roughly this order of effectiveness:
- Verified identity. A real selfie matched against a real ID, with a liveness check, is much harder than a stolen avatar. It’s not impossible — it’s just expensive enough that the average scammer moves on.
- Real history. Completed transactions, tied to real buyers, reviewed by real people. You can fake one review. You can’t fake fifty mutually-confirmed interactions across a year. The cost of synthesizing that history is higher than the upside.
- Portable consequences. When the profile carries the trust, removing the profile destroys the asset. A scammer who builds a real history over months and then burns it on one scam paid much more than they got.
The honest math
The goal isn’t to eliminate the long tail of bad actors; it’s to make the long tail uneconomic. Even with all three above in place, some scams will get through — usually from accounts that built real history first specifically to burn it. Those scams have a higher per-scam cost (months of building) and a higher per-scam ceiling (a single big payday). Different shape than the spray-and-pray model, fewer victims overall, much easier to investigate.
And here’s the under-discussed bit: when most platforms raise the accountability cost, the scammers don’t go away. They move to the platforms that haven’t. Trust products compete with each other on this dimension specifically. The first ones to raise costs benefit; the laggards inherit the scammers.
Why we say this out loud
Promising zero scams to a buyer is a small lie that eventually becomes a large one. The first time someone gets scammed on your platform after you promised, you’ve lost them — and they tell ten people. Saying we make scams expensive, not impossible — here’s how is harder marketing copy but better customer service.
We’d rather under-promise and earn trust the slow way. That’s the whole product, end-to-end.