A merchant on Amazon's US Seller Forums asked the company in late August 2026 for a dedicated investigation and appeal function covering what it calls malicious returns: bursts of orders, refunds and negative reviews that, the merchant argues, competitors use to push rivals down the rankings. By October 4, 2026, the thread had 641 views, 15 likes and six replies, none of the visible ones from Amazon, and Ephraim Rosenberg, founder of the Amazon Sellers Group, had carried it to LinkedIn.

In Short

A seller asked Amazon to create a way to report shoppers who buy items and send them all back on purpose, saying competitors use this trick to hurt rival shops, and other sellers added their own stories about the same buyers returning goods again and again. If you sell on Amazon, a burst of returns and angry reviews can lower your ratings and push your products down in search, which means fewer sales and more expensive advertising. Nothing has changed yet: Amazon has not answered in the thread, and sellers still have no special channel for sending evidence about suspected return attacks.

Four orders, one minute, one buyer

The post carries a long title - "[Seller's Voice] We earnestly request the platform to add a special investigation function for malicious returns and rectify the malicious return suppression behavior of competitors" - and appears under the account Seller_x3bBcgVNqAdTx in the Create and Manage Listings section of the Seller Central forums, tagged Product feedback. Its author describes the business as a long-term compliant seller whose store "has recently suffered from targeted malicious returns and negative feedback suppression from competing sellers."

The central example is precise. According to the post, a single buyer purchased four units of the same item within a single day and returned all of them, which the author calls "merely the tip of the iceberg." A redacted order table attached to the post shows what that looked like from inside the seller's account. Four orders sit on Amazon.com, each for one unit, two timestamped 20:44 and two 20:45 Pacific Daylight Time on August 7, 2026. Each carries the same buyer first name, an expedited order type, a payment-complete status and a logged refund request. The fulfilment channel reads Amazon, meaning the units shipped through Fulfillment by Amazon, and the table's column headings are in Chinese.

After checking orders, logistics data and buyer account details "one by one", the seller lists four abnormal characteristics: concentrated ordering accounts, heavily overlapping delivery addresses, return reasons with no actual product quality problem, and uniform negative reviews arriving shortly after delivery. From those, it concludes that the activity amounts to "confirmed malicious violations deliberately operated by peer competitors to seize traffic and suppress rivals". That conclusion is the seller's own. The post names no competitor, and the only evidence it publishes is the redacted table.

The damage, as the seller describes it

The post describes two channels of harm. The first runs through performance indicators: groundless returns, it says, depress the store's return rate and order defect rate, lowering what it calls store weight, cutting traffic and dragging down the ranking of carefully built listings. The second runs through shoppers, as negative feedback misleads real customers and erodes sales. Underneath both sits a procedural complaint. The seller says it holds abnormal order data and account characteristics but has "no official channel to submit verification and appeal."

One technical detail matters here. PPC Land's April 2026 account of Amazon's Anti-Manipulation Policy set out the three inputs to Order Defect Rate - negative feedback, A-to-z Guarantee claims and credit card chargebacks - and its 1 percent threshold, above which sellers risk losing selling privileges. A return on its own is not among those inputs. The negative feedback that the seller says arrived alongside the returns is. Return rate carries a separate cost in Europe, where Amazon has charged FBA returns processing fees on products exceeding category-specific return-rate thresholds since February 2025, with products shipping fewer than 25 units a month exempt.

What the seller wants built

The proposal has five parts. Sellers would file special appeals against abnormally concentrated returns and suspicious batches of after-sales orders. They would upload supporting material: order records, account characteristics, logistics information and evidence of competitor activity. Amazon staff would then carry out manual verification and a systematic investigation to separate genuine after-sales problems from competitive abuse. Where abuse was confirmed, the related performance deductions and negative feedback records would be removed automatically. Finally, the accounts responsible would be restricted and penalised.

The post also asks Amazon to "put an end to malicious involution", borrowing a term widely used in Chinese business commentary for self-defeating competition, and closes by inviting other sellers to comment and "speak up together". Several did.

A toy dealer's ledger

The most detailed reply came from Seller_hLH0bAcCwYcos, who says the business has sold on Amazon for 17 years, moves 5,000 toys a year and runs a return rate of 2%. According to that seller, 70% of returns come from the same four or five buyers.

Three cases are set out in detail. In the first, a buyer returned $600 of new, sealed Transformers figures in five returns, each SKU involved otherwise drawing almost no returns; the boxes that came back, according to the seller, contained a jar of peanut butter, a cup of jello, or goods worth about a tenth of the original value. In the second, a buyer said her son had bought $600 of Transformers without authorisation and sent them all back opened, damaged and stripped of about half their parts and accessories. The seller puts the cost of each case at $600.

The third case is stranger. A buyer purchased a rare Transformer, out of print for five years and now worth four times its original price, then returned it unopened and sealed 17 consecutive times. The product otherwise drew fewer than two returns per 200 orders, and the seller says that one buyer generated more returns than the previous 1,500 customers combined. The returns totalled $1,200, and return fees added $200, which the seller says wiped out the profit on about 20 orders. "That's a broken system," the seller wrote, noting that the buyer had blocked all messages from sellers while remaining free to keep ordering.

Why does a single collector hurt so much? As an authorised toy dealer, the business pays 80 cents on the dollar, then adds 40% to the manufacturer's suggested retail price to cover free shipping and Amazon's 15% commission, simply to break even. For every figure sold at a premium, according to the reply, at least 20 SKUs sit near break-even for up to two years before turning a profit.

Those three buyers, the seller says, were reported repeatedly over two years. Their returns covered almost 30 orders worth close to $3,000. Amazon sent warnings but did not block them, and the returns resumed around three months later. "We simply need to be able to block these bad customers," the seller wrote, adding that Amazon is the only site among those the business uses that "doesn't allow us to block fraudulent or malicious buyers." The reply also cites earlier buyer-blocking threads with 400 to 500 likes, one nearing a thousand; those figures cannot be checked against the documents. With 17 likes, it is the thread's most endorsed response.

Figures that do not fully reconcile

The numbers in the reply do not add up neatly, and the thread does not explain the gap. A 2% return rate on 5,000 annual sales implies roughly 100 returns a year, yet the seller describes the three buyers' 30 or so returns over two years as exceeding a full year of returns. The three itemised cases come to $2,400 in returns ($600, $600 and $1,200), below the total of almost $3,000 given for the same three buyers. The $1,200 in the third case also needs context: the collectible came back sealed each time, so that sum reflects reversed sales rather than lost stock, while the $200 in fees is a direct cost.

Dissent, and two smaller sellers

Not every reply agreed. The first response, posted in Chinese by Seller_6CL4jjlEJs4vR, argued that the returns were not necessarily malicious, since some buyers simply order this way, dislike the product and ask for a refund. The same seller predicted that the platform would change nothing because it wants to please buyers, and said it had already reported what it called a refund bug. That reply holds no likes and three dislikes, the only response in negative territory.

Two smaller sellers described patterns closer to the original complaint. Seller_Gp6MaNUcSuNdS, posting about four weeks before the capture, said one of the business's best-selling products fell from a 4.6 to a 4.0 rating, including negative reviews aimed at the price rather than the quality. Seller_ltSjL0MchcGcL, posting about three weeks before the capture, described three orders over four days, all returned, one marked defective, and a review rating down 0.2 points. That seller said a collectible trading card packed into the product as a bonus was not sent back, that a report to Amazon had gone unanswered, and that a competitor looked a more likely culprit than a thief.

The same reply blames Amazon's A9 algorithm for the outsized effect of three returns on a small seller. PPC Land has described the A9 and COSMO systems that govern Amazon search, though neither the reply nor the thread sets out how returns enter those systems. "Is there no recourse?" the seller asked, before closing: "Please do something Amazon!"

Two further replies, nested beneath the toy dealer's post, were collapsed in the capture and are not reflected here.

Rosenberg's amplification

The thread reached a wider audience through Ephraim Rosenberg, whose LinkedIn profile lists him as founder of the Amazon Sellers Group (ASGTG.com), based in Brooklyn, New York, and working on seller compliance and suspension problems. He shared a screenshot of the post in the week of September 7, 2026, judging by the three-week timestamp at capture. "Amazon sellers can be punished quickly when metrics move in the wrong direction," he wrote, before framing the issue as one of procedure. "If abuse is coming from another seller, a buyer, or a coordinated group, there should be a real path to submit evidence and get it reviewed." The post held 15 reactions and four comments on October 4.

It is not Rosenberg's first intervention on Amazon's enforcement design. In early May 2026 he questioned whether Amazon's approval gates and letters of authorization amounted to indirect enforcement of private distribution agreements, drawing responses from lawyers, compliance managers and sellers.

A discrepancy in attribution

The screenshot Rosenberg shared attributes the post to an account displayed as "cultivator", timestamped two weeks earlier. The forum page, viewed logged out on October 4, lists the author as Seller_x3bBcgVNqAdTx and dates the post to "a month ago". The text is identical, and the timestamps agree with each other: two weeks before a LinkedIn post made three weeks before the capture places the original in the second half of August, as does the "2 weeks ago" label beside the August 7 orders in the seller's own table. The difference in names is not explained by the documents. It may reflect how the forum displays authors to signed-in users, but that could not be confirmed.

What Amazon offers now

No Amazon response appears among the visible replies, although the page carries a filter for them. Three similar threads listed beside it each display an "Amazon replied" marker: a request to investigate a suspected professional negative reviewer, about a month old, with 38 views; a "Request for Investigation of Repeated Returns", about three months old, with 166 views; and a request to remove a malicious review, about seven months old, with 267 views. Their content is not in the capture.

Existing tools sit at other points in the chain, and none matches the proposal. When Amazon made prepaid return labels mandatory for all US seller-fulfilled orders from February 8, 2026, it pointed sellers to the SAFE-T reimbursement programme for refunds they believe were not their fault. Sellers criticised recent limits on SAFE-T, noted that the labels carry $100 of insurance, and one said its return rate had fallen by more than half when it stopped offering prepaid labels. The August 7 orders in the forum post, however, were fulfilled by Amazon, which processes those returns itself.

Seller Challenge, available since October 2025, gives Account Health Assurance participants three challenges per six-month period with a 48-hour review. It applies to enforcement decisions after a standard appeal has failed, not to a pattern of returns building up beforehand. Feedback recourse narrowed when Amazon moved to star-only seller feedback, announced on July 29, 2025 and live from August 4, disabling the appeal option for ratings submitted without text. The Anti-Manipulation Policy, meanwhile, bars sellers from offering refunds or anything else in exchange for feedback removal, even when a buyer starts the exchange.

Return windows have stretched too: holiday purchases from November 1 to December 31, 2025 could be returned until January 31, 2026, and sellers reported approvals 35 to 60 days after delivery.

On reviews, Amazon has published detection figures: the company said it blocked or removed more than 200 million suspected fake reviews in 2022, partly through AI. Regulation reaches the same conduct from the other side. The Federal Trade Commission's rule finalised on August 14, 2024 bars buying positive or negative reviews, though the agency clarified that it does not apply to merely hosting them. A competitor paying for hostile reviews would fall inside it. A competitor buying goods and sending them back is a different matter, and nothing in the documents points to a comparable rule for returns.

Why marketers are watching

The thread follows a case documented on August 31, 2026, in which a UK seller's whole account was suspended three days after Amazon reinstated the disputed listing. There, coordinated buyer accounts placed orders, requested refunds, left one-star feedback and filed counterfeit complaints over two weeks in late July 2026, and Amazon's forum moderator acknowledged limited visibility into buyer-related matters. The new thread moves the same logic to returns, a channel that needs no complaint at all - only an order and a refund request.

For advertisers, the exposure sits in the listing. Sponsored Products campaigns buy clicks against individual product pages, so a rating that slips from 4.6 to 4.0, or a slide in organic rank, changes the value of every click bought there without any change to bids. Since Amazon restricted review sharing across dissimilar product variations, phased in from February 12 and completed on May 31, 2026, each rating carries more weight on its own listing. A handful of hostile reviews now moves a smaller base.

The complaints also land on a seller base that has described itself as under strain. Between May and August 2025, sellers reported sales declines of 60 to 80% year on year, with one veteran naming Amazon's failure to eliminate scams and fake reviews among the causes. When organic visibility falls, paid placements on the same results page are the main route back to the same shoppers, so suspected sabotage can surface as a larger advertising bill rather than as a line item of its own.

The proposal borrows a familiar shape. Ad platforms handle invalid activity by detecting it, reviewing it and crediting the advertiser. The forum post asks for a marketplace equivalent: evidence in, human review, reversal of the metric damage. Would Amazon want a seller-facing channel of that kind, one that could also be turned against genuine customer complaints? The thread does not say, and neither does Amazon.

Timeline

Summary

Who: Seller_x3bBcgVNqAdTx, an anonymous merchant on Amazon's US Seller Forums, along with replying sellers including a toy dealer with 17 years on the platform; Ephraim Rosenberg, founder of the Amazon Sellers Group (ASGTG.com), who shared the thread on LinkedIn; and Amazon, which has not responded in the visible thread.

What: A request for a dedicated function letting sellers appeal abnormally concentrated returns, upload evidence, obtain manual investigation, have related metric deductions and negative feedback removed, and see responsible accounts restricted. Replies added accounts of repeat returners, including one buyer said to have returned the same sealed figure 17 times, alongside a dissenting view that such returns are not necessarily malicious.

When: The post appeared in the second half of August 2026, citing orders placed on August 7, 2026. Replies followed from late August to mid-September, Rosenberg's LinkedIn post came in the week of September 7, and both were captured on October 4, 2026.

Where: The Create and Manage Listings section of Amazon's US Seller Forums, under a Product feedback tag, concerning Amazon.com orders fulfilled by Amazon; and LinkedIn.

Why: Sellers say coordinated returns and reviews depress ratings, metrics and search rank with no channel for submitting evidence, and existing tools - SAFE-T, Seller Challenge, feedback removal - sit at other points in the chain. For advertisers, the same listings carry Sponsored Products spend, so a fall in rating or rank alters the value of paid clicks.