Do you Know Your Seller (KYS?) - Seller Check for Marketplaces
Seller fraud is bad for business in marketplaces. The traditional flywheel of a marketplace relies on attracting more buyers and sellers to create scale. But users lose trust if they regularly experience fraud or issues.
The figures are shocking:
- 40% of users have been scammed by a fraudulent listing
- 17% of users report being scammed using Facebook marketplace alone
- 80% of users will avoid a listing for fear of being scammed
Marketplaces need to make it as frictionless as possible for people to become sellers, expanding the universe of companies from 1,000s to millions to potentially billions.
In a world where everyone is a seller, how well do you know your seller (KYS)?
Why is seller fraud spiking in marketplaces?
Criminals always find the weakest link. As buyer-side fraud detection tools have improved, marketplaces haven’t focussed as much on adding fraud screening for sellers or merchants.
The second reason is the aperture of possible sellers widens with each possible year. As it gets easier to become a seller, it also becomes easier to list products on a marketplace.
Anatomy of seller fraud
Bad actors
- Sign up as fake sellers because the checks are relatively lightweight (often it simply requires creating an account)
- List an item that doesn’t exist (because again, often there’s no way to check)
- Collect funds from the marketplace in advance
- Leaving the marketplace as a buyer before the marketplace gets paid
There are 100s of nuanced attacks that can be done with seller fraud (see our Tackling marketplace fraud blog for more here). Simple non delivery, new account fraud, fake listings and fake reviews are beginning to overwhelm marketplaces. They’re now dealing with unhappy buyers, and a reputational risk.
The need for lightweight business verification
Banks and Fintech companies solve a lot of risk with “KYB” or “Know Your Business” checks. This involves getting detailed identity information about business owners, checking incorporation documents, and even looking into the company's background and operations to verify if they’re a real company.
I know what you’re thinking.
Marketplaces don’t want to be banks.
But they do need to fix their seller problem.
They don’t want to do KYC.
The answer is lightweight “business verification” of the seller.
A bank might add more comprehensive checks like EIN verification or verifying documents from the Secretary of State where a business is registered and call this KYB or “Know Your Business.”
Seller identity verification, without adding friction
In principle, we want to only bring businesses through high-friction KYB-style processes if we have some evidence they may be high-risk. For a new seller this can be hard to track but it is entirely possible with passive detection and building a fast lane vs slow lane.
Passive detection means looking for data signals about a potential user without requiring them to input any new information (other than would be required in a standard registration).
We always try to build a fast lane for good potential customers (sellers) and a slow lane for bad potential customers (again, sellers in this case). If the passive detection signals tell us this is a high-risk user, we can drop them into the slow lane and ask for additional information borrowing from the KYB experience.
How seller checks work in practice at onboarding
As a working example, let's assume a new seller is registering and could be high-risk; this is how it might play out.
Passive detection techniques help us figure out which lane this user belongs to.
- Device fingerprinting and network link analysis help detect if the same seller is behind multiple seller and buyer personas. For example if the user has shared devices, IPs, or emails we can see this as high risk.
- Email and phone history help in understanding if the email and phone number are new and if they’re registered to the entity they claim to be signing up for.
If we see any red flags ⛳ we can drop them in the slow lane and do further checks like.
- The “TrueIndustry” of a seller by pulling their state fillings
- The card or bank account history of a seller by asking them to enter their receiving account information
Once a user is in a slow lane, we can also take that concept further, like placing hold periods on them getting paid. Often bad actors are trying to make money quickly; the more speed bumps we can put in their way, the less likely they are to get away with it.
How seller checks work throughout the lifecycle
The key to all fraud prevention is to watch the user throughout their lifecycle on your marketplace. Every product listing, cash-out, and review is an opportunity to screen for high-risk sellers.
Passive detection techniques (on the device, for example) can be very helpful for things like fake reviews and collusion. By identifying a cluster of devices and creating reviews for sellers we already believe to be high-risk. Imagine if the same device that had just signed up as a seller also created 10 buyer accounts and started posting reviews exclusively on listings by the new seller.
The use cases are endless.
There is, however, the small challenge of making that data useful.
Zero to one in Know Your Seller (KYS).
Making the data useful is harder than it should be.
Marketplaces often rely on their payment provider for payment fraud detection and a hodgepodge of other tools to create a more complete picture of their customer. The data is split across the PSP, CRM, infosec, and countless other systems.
Ideally, you need one place for those passive detection data providers and your internal data to be combined into a set of logic (rules) and ways for your fraud team to analyze.
At Sardine we found what we built for banks (the KYB, rules engine, dashboard and device intelligence capabilities) are increasingly popular with our marketplace customers.
So if you’re struggling with more seller fraud, higher volumes of fake goods, and collusion, get in touch.
