How to safeguard your marketplace against seller fraud

The global e-commerce market is expected to total $5.55 trillion in 2022 — and to exceed that over the next few years. Borderless commerce presents a highly lucrative prospect for online retailers to tap into new markets and increase revenue.

According to Digital Commerce 360’s analysis, marketplace sales account for 67% of global ecommerce. The world’s top online marketplaces — eBay, Amazon, Alibaba, etc — sold $3.23 trillion worth of goods in 2021.

When you look at those kinds of figures through the eyes of a criminal, it’s not hard to see why they devote so much time to endless, varied and sophisticated defrauding methods.

Seller fraud is a growing concern for our marketplace partners. This can include anything from account takeovers (ATOs), counterfeit sales and delivery fraud to lighter policy violations — but the potential end results are the same: financial losses, bad buyer experiences and compliance issues.

To fight seller fraud and reduce business risk, you need to know how to take measures to detect and prevent the most common types of criminal activity.

In this whitepaper, we’ll delve into some of these methods and explore how you can combat them to protect your marketplace from the consequences.

Types of seller fraud

Seller fraud encompasses many varied and sophisticated techniques of deception, affecting both the marketplace and its customers. It can happen at any point in the sales cycle, from when a marketplace store is set up all the way through to the shipping and returns process.

Fraudsters will work to manipulate their store’s ranking within the marketplace, influence competitors and launder money through CNP payment fraud.

Some common types of marketplace seller fraud include:

Product listings

Luxury goods might be the first thing that springs to mind when it comes to counterfeits, but product listing fraud manifests itself in many different forms: misleading descriptions, inferior quality items, non-existent products, false photographs, and misclassifications all come under this umbrella. And counterfeiters are known to prey on the vulnerable — during the pandemic, lateral flow tests, safety equipment, medicines, masks, and PPE were counterfeited at scale to take advantage of the unfolding crisis and the public’s increased dependency on e-commerce.

Shipping fraud

From failing to ship orders, to creating bogus shipping companies with dummy tracking details, shipping fraud is widespread and increasingly sophisticated. At a basic level, a seller might re-use an existing tracking number. The buyer believes they are seeing their parcel’s journey progress, but it is in fact a completely different parcel making its way across the globe. At a more advanced level, fraudsters may replicate a well-known shipping company’s website, even down to the name and URL, to throw shoppers off the scent of a non-existent shipping process.

KYC

This category involves sellers using fraudulent identities or other practices that violate a marketplace’s terms and conditions. For example, using stolen identities or altered documents to set up the store, creating duplicate stores, or setting up a store strictly to collect positive feedback and reviews so that store can be sold to the highest bidder later (known as ‘born to sell’).

Collusion and antitrust

This category involves practices designed to change the dynamics of a marketplace. For example, unscrupulous sellers damaging competitors’ reputations by leaving bad reviews, or acting as buyers of their store’s own products — known as ‘brushing’ — falsely driving up prices (‘shill bidding’), inflating sales volumes and leaving positive reviews to mislead real customers. Brushing is designed to game marketplaces’ competitive search ranking algorithms, which favour high sales volumes and glowing reviews.

Product and listing fraud

• Misleading listings
• Counterfeit listings
• Illegal and banned merchandise
• Pricing and listing deceptions
• Copied listings
• Listing as marketplace brand
• Poor-quality listings
• Non-existent listings
• Secondhand as new
• Misclassification of products
• Excessive duplication of products
• Incorrect variations
• Text on images
• Unlicensed used of marketplace brand
• Hidden links

Shipping fraud and abuse

• Not shipping
• Selective shipping
• Dishonest shipping practices
• Shipping SNAD items
• Issuing fake/used tracking numbers
• Combined shipping issues
• Falsifying inventory location
• Violating shipping and refund policy
• Creation of bogus shipping company
• Use of unsupported shipping carrier
• Falsifying where lots ship from
• Encourage offsite communication
• Hidden links

Account and KYC issues

• Duplicate stores or accounts
• Failed KYC
• Born to be sold
• Excessive violations
• Unilateral refunds
• Born to defame
• Persistent risk issues
• Settlement outside of a marketplace
• Profitability issues
• Suspension at will/for violations
• Excessive INR/return/broken rate
• Excessive poor quality
• Excessive charge-backs/recalls
• Excessive handling time

Collusion and antitrust

• Fake reviews
• Automated/bot reviews
• Purchased reviews
• Buyer-seller collusions
• Damage to competitors
• Suspicious reviews close to holidays
• Gaming the platform
• Shill bidding
• Affiliation to oneself
• Brushings
• Buyer overpayment

There are many more fraudulent techniques and types that fall under and outside of the four main categories.

The seller fraud cycle

Seller fraud is likely to be part of a repeating cycle across one or many marketplaces. This is of particular concern to marketplaces as it can be difficult and laborious to detect, despite its cyclical nature. It exploits platform vulnerabilities, causing direct and indirect impact on even the most sophisticated marketplaces.

Payoneer seller data found that 67% of all seller fraud cases were connected to prior cases, and 57% could have been prevented if the bad actors had been flagged after their first offense.

According to this data, these schemes can cost marketplaces upward of 4% of revenue — as well as threaten compliance practices and damage brand reputation and consumer trust.

Fraudulent sellers often set up multiple stores in advance, so they have dormant stores to reactivate when one is shut down. They can use stolen or bogus identities to attempt to circumvent know-your-customer (KYC) checks.

The fightback against fraud

Marketplaces need robust fraud prevention strategies in place to combat these multifaceted challenges. And today’s competitive e-commerce landscape means that customer experience is key to retaining lifetime value, driving loyalty and growing sales.

A balance needs to be struck between delivering seamless buying and selling experiences — meeting expectations for speed, convenience and security — and implementing an effective fraud prevention infrastructure.

“The key to managing fraud effectively comes down to data. Using buyer, seller and transaction data wisely is the number one way to identify, combat and prevent seller fraud.”

Maxim Polyachenko,
Senior Director, Global Risk Operations, Payoneer

Some marketplaces opt for overprotective fraud prevention policies that aim to protect their platforms from repeat offenders. But these policies come at a price: they create friction with customers due to impacting on onboarding speed (30+ days), causing delayed payouts and seller deboarding. On top of this, they are often unsuccessful in preventing fraud/money laundering and are costly due to the maintenance of associated teams.

But it doesn’t need to be like this. With direct visibility into transactional data, marketplaces can detect patterns that will prove invaluable in preventing fraud. Payoneer’s unique perspective of the full payment ecosystem leverages data to design a four-pronged attack on attempted fraud:

• Flagging suspicious payment behavior in-platform:
  enabling marketplaces to investigate sellers preparing to or currently committing fraud.
• Flagging cross-marketplace fraudulent behavior:
  enabling marketplaces to prevent sellers with a fraud track record from onboarding to begin with.
• Flagging store and/or account linkages:
  enabling marketplaces to prevent repeat fraud, and gain insight on group behavior.
• Our unique oversight of sellers’ ecosystems and ability to track the flow of funds mean we can identify the ultimate beneficiaries and act quickly to halt fraudulent networks.

Leveraging varied data points to identify fraud patterns

The power of Payoneer’s data network

We have a unique, birds-eye view of sellers across multiple marketplaces, with data points that identify connected sellers and accounts. Our unique positioning allows for detection of store connections and fraudulent activities, even cross-marketplace. Parameters such as device fingerprints, ATM usage and behavioral patterns can be combined to create a visual depiction of the seller ecosystem, identifying potentially fraudulent connections.

Connecting the dots

Tracing connections between active stores

This is an example of what, on the surface, looks like 127 different independent stores. But using our data points, we can easily trace connections between stores and determine that rather than 127 sellers, we have 25 sellers with multiple stores, some with varying geographical locations. In the centre, you’ll see three different stores, from three different countries, operated by the same seller. This powerful visibility into marketplace networks and ecosystems is an invaluable tool to fight fraud.

• 127 stores turned out to be owned by 25 sellers, with multiple stores each
• Group sizes vary significantly (in this example from 2 to 15, but overall, to 000s)
• Some stores have long tenures, while others are disposable

Repeat fraud

Our datasets also surface connections between terminated stores and their replacements, even when fraudsters have gone to great lengths to cover their tracks.

Here you can see the connections between 25 terminated stores and 58 new stores, created by the same sellers using assumed identities or alternative details.

Connected stores opened after initial stores shut down

• Out of 25 closed stores, 48% had at least one connected store active after closure
• In total, 58 stores active after initial store closed
• Most exhibit patterns of store closure and opening for a prolonged period

Seller fraud typologies

Hidden links: what you see is not what you get
This type of fraud can be hard to detect or prove if you don’t have robust datasets to identify it. Sellers usually use this method to avoid detection if they’re shipping counterfeit or illegal goods: they’ll list a product as one thing but ship another. Buyers reach this product listing via a network like a messaging channel or social network, so they will be aware that what they are purchasing isn’t what is pictured in the listing — and all the information about the true nature of the product or transaction is off-platform.

Offered on a marketplace

Shipped to a buyer

This allows fraudsters to hide all the information from financial partners, marketplaces and measures to screen or detect abuse and fraud.

Red flags like unrelated product reviews or coded instructions in the order notes are giveaways, but this method of manually identifying cases of hidden links is not scalable.

But when multiple data points are layered together, you can start to build a clear picture of when a seller’s listing might not be all it seems:

Indicators to help detect hidden links

Check websites and ads

Domain registration, dates, reverse WHOIS lookup, similar sites

Prices

Significantly lower (<~20%) than on manufacturer/official website

Shipping

Global shipping of localized branded goods (exclusive importers), shipping companies in use

Screen shipping company

Size, location, pricelist, registration, owners, website (dates, templates, traffic), reviews

Reviews

Clues on the packaging, carrier, localization (voltage, AC plugs)

Geography

Seller’s whereabouts and inventory location

Flow of funds

Tracing the money flow

Seller’s experience

Seller’s experience and past activities

Connections

Seller’s linkages – accounts, stores, entities, persons

Triangulation fraud

This method is much more complex than the previous example, presenting an elevated level of risk to the store and the marketplace. Basic triangulation involves three parties: the buyer, the seller and the manufacturer (or distributor).

• Buyer places an order from the seller
• Seller then places the order from the manufacturer, supplying the buyer’s address but using compromised financial means: for example, a stolen card, cheque or debit
• The manufacturer ships the item directly to the buyer
• A chargeback occurs
• The manufacturer loses out on the funds and the product

The seller can build up their store’s sales and positive reviews because they have fulfilled the requested order. However, this is risky because it’s highly likely that law enforcement will open an investigation into the stolen payment methods.

Quadrangulation fraud

Advanced triangulation adds an extra layer of complexity — so it becomes quadrangulation. Fraudsters will seek out sellers with strong reputations and offer their services as a fulfilment partner. So, the seller continues to list items, but the fraudsters will step in to take care of shipping and fulfilment, splitting the revenue with the seller. This distances the fraudster from the process slightly:

• Buyer places an order from the seller
• Seller sends the order details through to a fulfilment partner (fraudster)
• Fulfilment partner (fraudster) then places the order from the manufacturer, supplying the buyer’s address but using compromised financial means: for example, a stolen card, cheque or debit
• The manufacturer ships the item directly to the buyer
• A chargeback occurs
• The manufacturer loses out on the funds and the product (again)

Triangulation 2.0 (quadrangulation)

In triangulation, the seller was also the fraudster. But in this example, the fraudster is one step removed, which makes detection that bit harder.

3D or cubic scams

With triangulation and quadrangulation, there were two/four parties in the chain. This next example works the same as quadrangulation, except the fraudster operates outside of the quadrangle by selling gift cards to the (innocent) fulfilment partner.

• Buyer places an order from the seller
• Seller sends the order details through to a fulfilment partner (fraudster)
• Fulfilment partner, separately, has bought gift cards from a fraudster
• Fulfilment partner (fraudster) then places the order from the manufacturer, supplying the buyer’s address but using compromised financial means: for example, a stolen card, cheque or debit
• A chargeback occurs
• The manufacturer ships the item directly to the buyer
• The manufacturer loses out on the funds and the product
• The fulfilment partner is investigated for using stolen/illegitimate gift cards

Going 3D – cubic scams Parallel layers of fraud

As you can see, this type of fraud poses high risk to all parties involved: the seller, fulfilment partner, manufacturer and consumer. But the fraudster is so removed from the entire cycle of placing the order, selling the product, and shipping that it’s incredibly difficult to detect.

Preventing this type of advanced fraud again comes back to the data: layering the nine key indicators above to accurately predict and prevent fraud, protecting your marketplace from financial losses, reputational damage and further risk.

Refund-as-a-service scams

The latest emerging trend is ‘refund-as-a-service’ scams, which are spreading like wildfire and look legitimate to many marketplace buyers. The ‘service’ they advertise involves facilitating partial refunds — 50-90% — for purchases, while allowing the buyer to keep the product.

Once the buyer has sent over their purchase receipt, the fraudster will either fabricate a return by sending an empty box to the marketplace, supplying an altered tracking ID, or claiming lost in transit — or use stolen financials to pay the marketplace and request a refund to the original payment method. The marketplace will refund the buyer, who pays a fee to the fraudster. The marketplace suffers a chargeback/recall after the fact.

Preventing this type of advanced fraud again comes back to the data: layering the nine key indicators above to accurately predict and prevent fraud, protecting your marketplace from financial losses, reputational damage and further risk.

Conclusion: the value of provider partnerships

Leveraging partnerships with payment providers and fraud solutions can help marketplaces gain better insight into their transactional data and into the sophisticated fraud schemes that stand behind corrupt seller practices. In learning them, marketplaces will be able to outsmart them.

Payoneer partners with the world’s leading digital brands to solve their biggest international risk and compliance challenges. Working in tight collaboration with our clients allows us to offer sophisticated detection and better enforcement, tailored to the specific needs of each client while drawing upon our extensive cross-platform network of data. With this data, we can identify and target common repeated fraud patterns, connecting the dots to detect highly sophisticated fraud ring activities that threaten clients across our ecosystem — and keeping our customers compliant with the regulatory requirements in every country in which we do business.