Empty Box: Who's Really Swapping Products on Marketplaces

Blame the Nearest One

The scenario of product substitution is familiar to almost every active marketplace shopper. A person opens the box at a pickup point, and it's not what they ordered. Instead of original headphones—a Chinese knockoff. Instead of a smartphone—a block of plasticine matching the weight. Instead of a new jacket—a worn one. The buyer then writes an angry review, processes a return, and gets their money back. From the buyer's perspective, justice has been served, but the key question remains: who actually substituted the product?

By default, the seller will be held responsible. The cost of the item, a penalty, and logistics both ways will be deducted from their balance. Their listing will receive a negative review and drop in search rankings. And all of this happens based on platform rules, since proper investigations are typically not conducted.

Meanwhile, from the seller's warehouse to the buyer's hands, the product goes through receiving, a sorting center, reloading, long-haul transportation, a pickup point, and a fitting room. At each of these stages, different people handle the box, and the seller isn't there (unless they handle delivery themselves). The seller last saw their product when they sealed the packaging.

The scale of the problem is driven by the very mechanics of online commerce. According to data from Shift agency, buyers return between 15% and 45% of clothing and footwear ordered online, with the maximum occurring precisely on large marketplaces. In fashion, there are sellers with a 10% purchase rate—out of ten items ordered, only one is actually bought. Each return costs the seller an average of 300–1,500 rubles in reverse logistics, which is sometimes more expensive than the product itself. 

Returns have effectively become a separate expense item that sellers are forced to account for. A return with substitution also means loss of the product, which is paid for by someone who most likely had nothing to do with it.

Argument Media spoke with Veronika Panushkina, founder of the Russian children's goods brand NEONLOVE, a company selling its products on marketplaces.

Hair Ties Instead of a Hammock

Panushkina makes hammocks for strollers—devices for transporting a second child. This is a product with load-bearing attachments and a stated capacity of up to 35 kg, meaning it's something a child's safety depends on. Which makes what the brand gets back all the more bizarre.

The first case is almost comical, if not for the consequences. A customer ordered a hammock, arrived at the pickup point, and was handed a bag of hair ties. Naturally, she refused the "product." And left a review on the listing with a low rating and the words "I feel cheated." The seller took two hits at once—a lost order and a damaged rating. The customer never saw the hammock at all.

The second case was a return with the explanation "counterfeit product." Inside the box, instead of a hammock, was a piece of ordinary black fabric. And it's important to note here that the sale was conducted under the FBS scheme—a model where the seller stores goods at their own warehouse, and when an order comes in from the platform, they assemble it themselves, label it, pack it, and hand it over to a courier or drop-off point. According to Panushkina, she personally packed the item and has all the video recordings of the assembly.

Third. Between May and June 2026—several identical refusals upon delivery, all with the same reason: "wrong item in the order." What exactly was in those boxes, Panushkina doesn't know and won't find out. The goods go back to the marketplace warehouse, and the contents aren't shown to the seller.

A separate issue—returns without attachments. The hammock comes with four high-strength carabiners; they're what hold those 35 kilograms. These aren't consumables, but part of the structure. And they go missing.

"Returns periodically come back without the attachments, and there's a suspicion that some orders are placed solely to extract the carabiners and return the hammock itself," Veronika explained.

There are also examples of pure consumer fraud—"rentals." A hammock is ordered for a trip, then a dirty, used item is returned. The same story as with dresses before New Year's, except this is children's textile that can no longer be sold after such treatment.

Each incident can cost the brand the full price of the product, double logistics, and penalty write-offs.

The Platform Loses Nothing, the Seller Gets an Empty Box

Argument Media sent inquiries to the largest marketplaces regarding unjustified returns and product substitution. Ozon and Yandex Market provided comments. Wildberries did not respond to the request.

Ozon insists the problem is not widespread. The press service reported that the share of goods in completed returns versus delivered items is 1.41%. When signs of bad faith are detected, anti-fraud measures kick in—a multi-level analysis with machine learning. To return a product, the buyer must submit a request in their personal account and attach photos of the item, packaging, and tags. And regarding logistics, the wording is as follows:

"Our rules allow us to reduce cases of substitution or mis-sorting of goods in Ozon logistics to zero"

Yandex Market's press service describes the process in more detail. Warehouses have video cameras, and in disputed situations, the seller can file a claim—the platform will pull up the footage. Mis-sorting is handled by a separate "anomalies department," which determines whether it's the supplier's error or the receiving team's, and if necessary, manually recounts goods by storage cells. Pickup point employees inspect returned items for signs of use and compare them with photos from the request when accepting returns. Sellers are given 48 to 72 hours to file a claim.

Both platforms advise sellers to do the same thing—record on video the assembly, packaging, and handoff to the courier, and communicate with support about every incident.

Five points where substitution can occur

The most detailed risk map of where products can be substituted was presented to Argument Media by Svetlana Karpova, head of the Marketing Department at the Higher School of Management faculty of the Financial University under the Government of the Russian Federation. Her rule is simple: "The probability of substitution is inversely proportional to the density of control." Substitution happens where the digital trail breaks.

The hottest point is the pickup location at the moment of return. The customer goes into the fitting room with the item and returns a cheap copy, worn item, or trash.

"The blind spot begins at the pickup counter. There are no in-line scales there. The camera only captures the general view, not the barcode in the customer's hands. Pickup point employees physically don't have the time or ability to thoroughly verify serial numbers, seals, or original tags on hundreds of items per day," Svetlana Karpova emphasized.

The overwhelming majority of substitutions, in her opinion, happen "in the last ten meters"—at the pickup point or already at home.

The second zone is warehouses and sorting centers. The more times a box gets transferred, the greater the chances it will be opened somewhere without camera coverage.

"Unscrupulous employees can carefully open factory packaging, remove the valuable item and replace it with a copy of similar weight, then repackage the box."

Returns processing is especially vulnerable. An item from a previous buyer arrives at the warehouse, no one really inspects it properly, and it moves on to the next customer.

The third zone, which gets talked about least of all, is parcel lockers. There are no witnesses at all here. A courier loads a compartment, a customer picks it up. "The courier might put the wrong item in the compartment. The customer might pick up an empty box and write to support that 'there was a brick inside,'" Karpova describes. Without a recording of the opening at the moment of transfer, it's impossible to prove anything.

The fourth zone is what happens at warehouses after returns. Olga Borisova, associate professor in the Department of Corporate Finance and Corporate Governance at the Financial University under the Government of the Russian Federation, told Argument Media about possible risks:

"Employees in 'blind spots' not covered by cameras can open a box and replace its contents. Perfume, cosmetics, and small electronics are considered most vulnerable. At some warehouses, the situation is made worse by the absence of a quality video surveillance system."

And there's a fifth point that the industry doesn't like to discuss—employees at the pickup points themselves. Yulia Dzhanibekova, author of the Telegram channels "Russian e-commerce Market" and "Retail," shared industry information:

"Pickup point owners describe schemes where employees register purchases under their own names or those of acquaintances, receive the original product, and then return a counterfeit or different item. As a result, the original stays with the dishonest employee or their accomplices, while a substitute goes back to the warehouse."

The list of high-risk categories is nearly identical across all experts. Smartphones, headphones, graphics cards, watches, jewelry, perfumes, branded clothing. The common thread: expensive small items that are easy to swap for similar ones.

Nobody knows exactly how much

Determining the precise scale of the problem for sellers is extremely difficult, since tracking such statistics is quite challenging. Alexey Moskalenko, president of the Association of E-commerce Representatives (APET), believes the problem affects virtually all sellers, and cites an example from auto parts: a customer orders a part, installs it on their car, then puts the old one in the box and sends it back as "didn't fit." Artem Sokolov, president of the Association of Internet Trade Companies (AKIT), sees the opposite trend: "We're noting a decline in such complaints from sellers, as platforms have introduced their own preventive measures to stop such abuses."

Who's right? Alexey Molodykh, general director of the International Association for Marketplace and Entrepreneur Development (MARP), explained to Argument Media why the debate fundamentally cannot be resolved.

"No one is systematically tracking comprehensive market-wide statistics on this specific problem—namely, product substitution or damage during returns—and it would be incorrect to cite a precise share of sellers or sum of losses."

Dzhanibekova confirms the same thing from the numbers side — there's no public data on what share of substitutions is whose fault.

It turns out that the problem around which sellers, associations, marketplaces, the Ministry of Economic Development, and the government are arguing has still not been measured. Everything else grows out of this absence of numbers.

The seller pays not because they're at fault, but because it's cheaper that way

If it's unknown where the substitution occurred, someone still has to pay. The legal construction is well-known: in their offers, the platform calls itself an information intermediary, the goods remain the seller's property until delivery, so the risk is on them. But the real reason, explains Svetlana Karpova from Financial University, isn't about jurisprudence — it's about finances.

"Investigating each incident requires matching video footage, weight data, and employee testimony. The cost of such an audit exceeds the price of the goods. In this scheme, the seller acts as the cheapest insurer. Payment for this 'policy' is already built into the logistics commission."

Yulia Dzhanibekova says the same thing in different words: "The seller pays not always because they're at fault, but because in the standard model they are the simplest and legally established center of responsibility." There's no "presumption of guilt" in the law, she emphasizes. There's an operational model in which money is withheld from the seller's balance immediately, and then they run around proving their case.

Proving it, it must be said, under rather peculiar conditions. Borisova describes this construction as follows: the burden of proving substitution lies with the seller, but they have no access to the marketplace's internal video archives. The platform's logic is simple — if the wrong product left the warehouse, it means the seller mixed it up.

What's being proposed: from non-returnable goods to end-to-end tracking

A regulatory resolution is near. On October 1, 2026, the law on the platform economy (289-FZ) takes effect. And in February 2026, at a strategic session with Mikhail Mishustin, the Ministry of Economic Development presented a package of return measures — a list of non-returnable goods, the seller's right to approve returns of quality goods and charge the buyer for return shipping, photo and video monitoring of returns with seller access to materials. The associations have their own position, which differs in places from the authorities' proposals.

AKIT President Artem Sokolov explains the root of the problem simply: offline there's a list of non-returnable goods, online there isn't one, you can return absolutely everything within seven days. The rule was written as a guarantee for buyers, but according to him, "it has become grounds for abuse of rights in a number of cases." This means amendments are needed to the consumer protection law — its own list of non-returnable goods for distance selling and clear criteria for preservation of merchantable condition: packaging, labels, seals, identification marks. And the solution should be the amendments proposed by the Ministry of Economic Development.

APET is taking the same path, but adds an argument about security. The association doesn't want to restrict honest buyers.

"The buyer should have the opportunity to receive goods at a pickup point, familiarize themselves with it, inspect it, check it. But if the goods have already been taken beyond the pickup point or, for example, taken into a fitting room, then after that it should not be subject to unconditional return," association president Alexey Moskalenko told Argument Media.

MARP is targeting something else — traceability. Alexey Molodykh named three priorities to Argument Media: mandatory photo and video documentation of the product's condition at each stage of transfer linked to the order, verification of completeness and seals in the buyer's presence upon return, not afterward, and a buyer rating system for return integrity, accessible to the seller before making a decision. He formulates the meaning of the first point this way:

"This shifts the burden of proof to whoever broke the chain, not to the seller by default"

The difference between these approaches is fundamental. Non-returnable goods and paid delivery reduce the flow of returns. Traceability doesn't reduce anything, but it allows you to find out where exactly the goods disappeared. A ban on returning opened cosmetics won't help a seller who received an empty box from the warehouse.

Proof costs more than the product

Today the entire marketplace operating system is configured for one task — retain the buyer, make returns quick and indisputable. This makes sense from a turnover perspective, but such a setup has a price, and it's paid by the link with the least ability to object. The seller most often pays not because a court or investigation established their guilt, but because they turned out to be the cheapest place to assign the loss. As long as figuring out where exactly the goods disappeared costs more than writing off its value from the seller, nothing will change — no participant will voluntarily spend more on investigation than the hammock itself is worth.

Hence the limit of current measures. Both Federal Law 289, the Ministry of Economic Development package, and the associations' proposals work mainly with the flow of returns and with "platform-buyer" relations: what can be returned, at whose expense, within what timeframe. This will reduce the number of disputed situations, but won't answer the question that begins every substitution story — at what point did the product disappear. Only something that doesn't yet exist as a mandatory standard can answer this: end-to-end documentation of every hand-to-hand transfer, in which the loss falls on whoever's stage broke the chain, not on whoever's easier to write off.

The platforms can do this themselves — the question isn't about technology, it's existed for a long time, but that under the current economics it's unprofitable for them. This means either traceability becomes a common rule of the game monitored by the regulator, or disputed situations and "seller-marketplace" interactions will remain in their private communications field, which won't reduce the number of disputed situations.