Marketplaces rely on repeat transactions between buyers and sellers in order to scale. Verifying the identities of buyers and sellers is critical in generating platform trust and maintaining this flywheel. 

Recently, seller fraud has been increasing: for example, up to 34% of Facebook Marketplace listings could be scams. Marketplaces have responded by beefing up seller verification strategies. While well-intentioned, these manual anti-fraud strategies can increase false positive rates and delay approvals, which negatively impacts the marketplace’s revenue prospects.

Read on to learn how marketplaces can implement robust seller verification strategies while maintaining frictionless onboarding experience for legitimate sellers.

‍

Types of seller fraud

Seller fraud at onboarding can be divided into two main types: third-party fraud – known as seller impersonation fraud – and first-party fraud.

‍

Seller impersonation fraud

The increase in data breaches, sophisticated social engineering, and the prevalence of the dark web have made it easier than ever to impersonate legitimate businesses. Bad actors can create new accounts - or take over existing accounts – using stolen identities and get to work quickly. Account takeover (ATO) attempts as a portion of total fraud increased 79% between 2021 and 2022.

Platforms have implemented protocols such as identity verification to crack down on seller impersonation fraud. However, some of these checks – such as checking email address domains – are static and do not account for the dynamically changing nature of fraud, allowing many fraudsters to slip through the cracks.

‍

First-party fraud

A seller intentionally creates a fake listing and collects revenue without ever delivering the product (“non-delivery scams”), or shipping a counterfeit product.

How do these bad actors get access to platforms? They often misrepresent their identity in order to bypass conventional know your business (KYB) checks, which can be pretty light on marketplaces. For example, they might lie about their industry or previous business history to appear more legitimate.

‍

How are platforms currently addressing seller fraud?

In addition to basic risk models, platforms have responded to increases in seller fraud through manual anti-fraud strategies designed to weed out bad actors. 

Typically, risk analysts are the first line of defense against seller fraud. During seller onboarding, risk analysts will review each applicant’s information and assess its likelihood of validity using a number of tools and their own judgment. Unfortunately, this can lead to biases in the onboarding process, and a high rate of false positives. For example, analysts might perceive applications with gmail.com email addresses as fraudulent, when in reality many legitimate sellers use their personal email to conduct business. If analysts mark these applicants as fraudulent, they are turning away credible business customers and leaving money on the table.

Why haven’t risk teams at marketplaces upgraded from these manual, subjective fraud mitigation strategies? Simply put, there hasn’t been a more efficient process – until now.

‍

Dynamic seller verification at onboarding

CorShield is our proprietary fraud model that helps leading ecommerce marketplaces crack down on seller onboarding fraud, reducing downstream losses. It addresses both types of fraud discussed above.

‍

Seller impersonation fraud

CorShield screens for questionable identities by cross-checking a seller’s provided information against known information on the business. For example, if a new seller signs up for a platform using a mix of stolen business information and a fake email address, CorShield will assess the likelihood that the email address is actually controlled by the business. Other fraud models just check for suspicious email domains, but CorShield goes beyond these tools, dynamically cross-checking ownership for email, IP address, phone numbers, business addresses, and more.  

‍

First-party fraud

CorShield flags first-party fraud by automatically triangulating known information on a seller and matching that against a seller’s provided information. For example, if a seller’s onboarding application mentions they offer marketing solutions, but Google reviews mention they actually offer gambling products, CorShield will surface this mismatch.

For each seller it analyzes, CorShield automatically generates a fraud score and top reasons for the score. This cuts down on the manual analysis risk teams are used to, helping them catch bad actors faster.

‍

Want to learn more?

Reach out if you’d like to learn more about Coris’s fraud models and how we can help your marketplace reduce fraud and improve customer trust.