Marketplaces are where everything else in this lane finally touches grass — accounts, cards, drops, and goods all meet a platform whose entire business is trusting strangers to trade. That trust is engineered: seller ratings, buyer guarantees, escrow timing, listing moderation, and payout holds exist because the platform priced its own fraud risk into the product. Which means the method here is not fighting those systems — it is understanding which ones release money early, which ones release goods early, and which pairs let you walk away before anyone's protection window closes.
ROUTE MAP (PLAIN ENGLISH)
Marketplace fraud splits by direction. Buyer-side runs against the platform's buyer protections: purchasing with instruments that will be disputed, claiming items never arrived when the tracking says otherwise, exploiting refund-without-return thresholds, and returning wrong units into the intake stream. Seller-side runs the other way: collecting payment for goods that never ship, shipping substitutes into cheap-fulfillment flows, or converting stolen inventory into platform balance. The platform's defenses are asymmetric by design — buyer protections are generous because buyers are revenue, while seller payouts sit behind holds, reserves, and delivery-confirmed windows precisely because sellers are where the platform's money sits at risk.
The economics live in the timing gaps: escrow release windows, payout hold periods, dispute filing deadlines, and tracking update latencies each create a moment where one side has been paid, the other has not, and nobody has filed anything yet. Whoever moves inside that gap decides which protection narrative applies to them — and the platform's resolution policy, not the truth of what shipped, decides the outcome.
STEP 1 - PLATFORM CLASS AND RISK PROFILE
STEP 2 - BUYER-SIDE SEQUENCES
STEP 3 - SELLER-SIDE SEQUENCES
STEP 4 - DISPUTE GRAMMAR AND EVIDENCE
WORKED EXAMPLE - ESCROW MARKETPLACE, BOTH DIRECTIONS
Seller side, aged account (two years, 340 feedback, reserve released long ago): listing a $390 electronics lot that converts through a legitimate supplier — honest photos, accurate description, ships next-day with tracked carrier and signature confirmation. Payout releases to the name-matched account on delivery confirmation inside the platform's standard window; five such sales across the quarter establish the account's cadence as an ordinary small merchant, no holds triggered, no claim filed against any of them — feedback and history are now real collateral. Buyer side, separate identity, same platform: a $210 component purchase where the tracking shows delivered to the building's package room but nothing reached the unit's door — INR claim filed day three (inside the window, before the carrier's own investigation timeline), one polite message to the seller referencing the scan discrepancy, evidence package consisting of the tracking record and a photo of the empty package room. Platform resolves buyer-favorable on thin scan evidence — no signature existed, the delivery photo showed a lobby table, not the door — refund issued, seller's protection declined because their carrier evidence was unsigned ground service. Total across both directions: ~$1,610 in released payouts plus a $210 refund outcome, zero cross-linked accounts (distinct devices, networks, funding sources), zero multi-channel escalations, every claim filed once and never re-filed. The two roles never touch in any platform-visible way, and each account's history reads exactly like what it looks like from outside.
WORKING SEQUENCE
QUICK ANSWERS
RELATED
Marketplace fraud splits by direction. Buyer-side runs against the platform's buyer protections: purchasing with instruments that will be disputed, claiming items never arrived when the tracking says otherwise, exploiting refund-without-return thresholds, and returning wrong units into the intake stream. Seller-side runs the other way: collecting payment for goods that never ship, shipping substitutes into cheap-fulfillment flows, or converting stolen inventory into platform balance. The platform's defenses are asymmetric by design — buyer protections are generous because buyers are revenue, while seller payouts sit behind holds, reserves, and delivery-confirmed windows precisely because sellers are where the platform's money sits at risk.
The economics live in the timing gaps: escrow release windows, payout hold periods, dispute filing deadlines, and tracking update latencies each create a moment where one side has been paid, the other has not, and nobody has filed anything yet. Whoever moves inside that gap decides which protection narrative applies to them — and the platform's resolution policy, not the truth of what shipped, decides the outcome.
STEP 1 - PLATFORM CLASS AND RISK PROFILE
- Escrow-marketplace (P2P goods, vehicles, domains, digital goods): funds sit in platform escrow until the buyer releases — buyer-favorable by default, seller carries the shipping risk. Disputes resolve against tracking evidence and communication records. The soft spots: delivery-confirmation timing versus tracking carrier scans, and buyer-claim windows that close faster than slow carriers deliver.
- Managed-retail marketplaces (the big-box platforms): buyer protection is near-absolute — A-to-Z style guarantees, refund-without-return on low-value items, prepaid return labels that pay for your own shipping back. Sellers eat the loss on claims, which is why seller-side on these platforms is about payout holds and category restrictions rather than trust.
- Classified and local platforms (peer-to-peer, cash-adjacent): no escrow at all — payment happens off-platform (wire, P2P transfer, cash), so platform protection is theater and the real contest is counterparty trust. Both sides here can lie with zero platform exposure, which cuts both ways and is exactly why these platforms score and ban aggressively once patterns surface.
- Service and gig marketplaces: deliverables are subjective — "work not as described" is the claim surface, revision cycles stretch delivery windows, and platform arbitration mediates a judgment call. Lower per-ticket value, higher claim ambiguity, quiet enough for cadence.
STEP 2 - BUYER-SIDE SEQUENCES
- Item-not-received (INR): the claim that tracking must answer — carrier scan shows delivery to the address on file but the buyer never received it. Resolution hinges on scan integrity: signature-required shipments beat unsigned ones, GPS-confirmed delivery photos beat door-knocks, and platform policy often resolves INR in the buyer's favor when scan evidence is thin. The method's honest version runs where the buyer controls the delivery context; the platform's defense is scan quality, not the claim's wording.
- Not-as-described / SNAD: photograph the discrepancy honestly — missing components, wrong variant, genuine defect — because a true evidence package reads like a disappointed customer and passes on pattern, while template phrasing gets caught at the pattern layer regardless of merit. Refund outcomes: full refund with return label (goods come back — cost, not profit), refund-without-return below cost thresholds (pure margin — the platform decided the item is not worth the shipping), or partial-with-keep settlements that both sides accept.
- Wrong-unit and swap-back: return authorization granted, what goes into the box is not what came out of it — intake serial capture on electronics, restocking inspection, mismatch codes propagating to fraud lists. This is the most instrumented buyer-side move and prices accordingly: reserve it for high-ticket, track serial state, never near other claim activity on the identity.
- Payment-instrument layering: the dispute that never gets filed because the claim resolved — but where the instrument itself was never legitimately the buyer's, the platform's eventual chargeback exposure lands on the account anyway. The platform joins instruments to accounts permanently; an account whose funding source later charges back carries that history through every future claim it ever files.
STEP 3 - SELLER-SIDE SEQUENCES
- List-ship-convert timing: seller payouts release on delivery confirmation or escrow completion — the window between buyer payment and payout release is where seller-side risk concentrates. New accounts face longer holds and lower limits; aged accounts with history release faster and at higher ceilings, which is why seller infrastructure gets seasoned with genuine small sales before any ticket that matters.
- Substitute and short-ship economics: ship a cheaper unit, a replica, or an empty parcel into tracked delivery — the tracking says delivered, the buyer opens the box to something else. Detection lives in buyer evidence (unboxing photos beat counter-evidence consistently) and in return-serial state; the pattern that survives is honest-looking merchandise with genuine descriptions everywhere except the unit's actual contents, never repeated across a cluster of buyers.
- Digital-goods delivery: no carrier, no scan — delivery is a message, a key, an account handoff. Platform disputes on digital goods resolve against message logs and redemption state, so delivery evidence is entirely what the parties wrote to each other. This class has the fastest conversion and the fastest bans; it is also where platform detection concentrates hardest because the tickets are small and numerous.
- Reserve and hold mechanics: platforms hold percentages of sales for new sellers, extend holds on category risk, and release reserves on schedule — modeling the actual release math (percentage, window, clawback conditions) before choosing ticket size separates a funded flow from one whose money is frozen mid-rotation. Read the reserve policy like a contract, because that is what it is.
STEP 4 - DISPUTE GRAMMAR AND EVIDENCE
- What reviewers actually read: timelines, tracking numbers, message logs, and photos with metadata consistent with the story. The winning packages are boring and complete — order date, expected date, what happened, what was tried, what evidence exists. Templates and emotional phrasing pattern-match to coached claims instantly regardless of factual merit.
- Platform policy beats truth: resolution follows the published guarantee, not the metaphysics of what arrived — a buyer with weak evidence inside a strong guarantee window wins more often than a seller with perfect scan evidence outside their protection window. Reading the current policy (it changes quarterly) is the actual skill.
- One claim, one narrative: re-filing a denied claim through a different channel — payment app dispute, chargeback, email escalation — is the multi-channel double-dip that joins every account the identity touches. Denial means the ticket's platform cycle is over; the appeal, if any, is one appeal with genuinely new evidence.
- Communication as evidence: every message between buyer and seller is a potential exhibit — sellers who write admissions ("send it back and I'll refund" without qualification) lose claims they would otherwise defend; buyers whose messages contradict their claim description lose on their own transcript. Write every message as if it will be read aloud in resolution, because it will.
WORKED EXAMPLE - ESCROW MARKETPLACE, BOTH DIRECTIONS
Seller side, aged account (two years, 340 feedback, reserve released long ago): listing a $390 electronics lot that converts through a legitimate supplier — honest photos, accurate description, ships next-day with tracked carrier and signature confirmation. Payout releases to the name-matched account on delivery confirmation inside the platform's standard window; five such sales across the quarter establish the account's cadence as an ordinary small merchant, no holds triggered, no claim filed against any of them — feedback and history are now real collateral. Buyer side, separate identity, same platform: a $210 component purchase where the tracking shows delivered to the building's package room but nothing reached the unit's door — INR claim filed day three (inside the window, before the carrier's own investigation timeline), one polite message to the seller referencing the scan discrepancy, evidence package consisting of the tracking record and a photo of the empty package room. Platform resolves buyer-favorable on thin scan evidence — no signature existed, the delivery photo showed a lobby table, not the door — refund issued, seller's protection declined because their carrier evidence was unsigned ground service. Total across both directions: ~$1,610 in released payouts plus a $210 refund outcome, zero cross-linked accounts (distinct devices, networks, funding sources), zero multi-channel escalations, every claim filed once and never re-filed. The two roles never touch in any platform-visible way, and each account's history reads exactly like what it looks like from outside.
WORKING SEQUENCE
Code:
platform pick: escrow P2P for claim timing gaps, managed retail for refund-without-return thresholds, classifieds only off-platform with eyes open -> seller side: seasoned account first, honest listings, tracked+signature shipping, reserve math modeled before ticket sizing -> buyer side: claims inside window, evidence complete and boring, one channel one narrative, SNAD over INR where scans are strong, wrong-unit last and alone -> discipline: instruments never shared across roles, accounts never joined by device/network/address, every ticket filed exactly onceQUICK ANSWERS
- Which side pays better — buyer or seller? Seller side carries the larger tickets through payout releases on seasoned accounts; buyer side carries higher frequency through refund-without-return thresholds and thin-scan INR resolutions. The platforms themselves set the ratio — managed-retail buyer guarantees are generous because sellers fund them, while escrow platforms protect sellers because buyers fund them. Pick the side the platform is protecting less.
- Why do new seller accounts get funds frozen? Reserve policy — platforms hold a percentage of new-seller sales for a defined window as fraud collateral, then release on schedule. It is written into the seller agreement, not an enforcement action; modeling the exact percentage, window, and clawback conditions before sizing tickets is basic flow math, and seasoning with genuine small sales is how accounts graduate to faster release terms.
- Does delivery tracking end INR claims? It decides them — but only as strongly as the scan itself. Unsigned ground-service scans with a generic delivery photo resolve thin against buyers; signature-required with address-confirmed GPS evidence resolves hard. The claim's wording matters least; the carrier evidence quality decides the outcome, which means the shipping service level chosen at listing time writes the dispute's future.
- Can one identity play buyer and seller roles? Not on one platform — roles share risk models, and an account that both claims against sellers and sells to buyers patterns like a layered operation the moment either role flags. Separate identities, separate devices, separate networks, separate funding sources, no visible edge anywhere between them, ever.
- What kills marketplace accounts fastest? Multi-channel double-dips — a denied claim re-filed as a payment-app dispute or chargeback joins every account the identity touches into one reviewable graph. Single-channel discipline (one claim, one narrative, one appeal max) is not politeness; it is what keeps tickets from ever becoming cross-account cases.
- Are digital goods worth it? Fastest conversion in the lane and the hardest detection concentration — no carrier evidence, message logs as exhibits, small numerous tickets, and platform moderation tuned precisely for the class. They work on cadence with clean infrastructure; they ban clusters on velocity. Ticket economics are excellent; discipline has to be better than every other category precisely because the platform is.
RELATED