Scam Listing Detection

What is scam listings?

A scam listing is a fake product, service, or rental ad posted to a marketplace to take payment or personal data without ever delivering. The seller usually pushes the deal off-platform, asks for upfront payment, and disappears once paid.

Also known as: fake listings, fraudulent ads, advance-fee marketplace scams.

How it works

A scammer posts an attractive offer, often a rental, a hard-to-find item, or a vehicle priced below market. The goal is to move the conversation to email, WhatsApp, or SMS fast, before the platform's moderation can flag anything. From there the buyer is asked to wire a deposit, pay with gift cards, or send money through a friends-and-family transfer. The item does not exist. The same listing text and photos are often reused across dozens of accounts, which is one of the clearest ways to catch them.

Warning signals

• Price well below market — A deal that undercuts every comparable listing is the most common hook.
• Pressure to go off-platform — The seller asks to continue on WhatsApp, email, or SMS within the first message.
• Upfront payment by untraceable methods — Requests for wire transfers, gift cards, or friends-and-family payments.
• Reused images and copy — The same photos or description appear across multiple accounts or regions.
• Brand-new or thin seller account — No history, no past sales, created days before the listing went up.
• Refusal to meet or verify — The seller invents reasons they can't do a video call, an in-person handover, or a verified inspection.

Real-world examples

• A rental apartment listed 30% under the local average, with the 'landlord' abroad asking for a deposit by bank transfer to hold it.
• A used car priced thousands below market, where the seller claims to be deployed military and routes payment through a fake escrow site.
• Concert or event tickets sold through a freshly made account that vanishes after the buyer pays by gift card.

Why it matters

Scam listings cost buyers money directly and cost platforms the trust that keeps them transacting. One viral scam thread on social media can do more damage to a marketplace's reputation than the fraud itself. Chargebacks, refunds, and support load follow every successful scam, and regulators increasingly expect platforms to show they actively work to prevent it.

How ModPilot detects scam listings

1. Custom rules as a first filter — Block or flag listings that include off-platform contact details, known scam phrases, or payment-method requests before they reach a queue.
2. AI analysis of listing content — A model reads the title, description, and images together to score how closely the listing matches known scam patterns, with a confidence value attached.
3. Duplicate and reuse checks — Matching reused photos and copy across accounts surfaces scam rings that single-listing review would miss.
4. Escalation on low confidence — Borderline cases route to a stronger model or a human reviewer instead of guessing.
5. Logged decisions — Every approve, reject, or escalate is recorded with reasoning for audits and appeals.

Most scam-listing programs fail because they review one listing at a time. The fraud lives in the pattern across accounts, not in any single ad. A description that looks fine on its own becomes obvious once you notice the same photo set running under five seller names in three cities.

That is why screening at submission, paired with reuse detection, catches more than a human queue ever will. The queue is for the genuinely uncertain cases, not the ones a rule could have stopped on the way in.

Frequently asked questions