Key Takeaways: FraudWatch: a Glenbrook & Mastercard Quarterly Webinar
Executive Summary
The webinar addressed the growing challenge of return and refund fraud, now ranked as the top fraud concern for merchants, surpassing traditional third-party fraud. Speakers outlined how schemes such as wardrobing, returning stolen or counterfeit goods, swapped items, false “did not receive” claims, and policy abuse are driving significant losses across merchants, issuers, and the broader payments ecosystem. The problem is fueled by liberal return policies, e-commerce growth, social media “hacks,” and increasingly by AI-enabled fraud tactics, with both first-party and third-party actors involved. MasterCard is piloting “Return Risk Intelligence,” leveraging network-wide transaction, chargeback, and fraud data combined with AI to detect high-risk return behaviors, cross-merchant abuse, and double-dipping, and to share actionable intelligence with ecosystem partners. The initiative aims to break down data silos, provide standardized categorization and reason codes, and enable proactive, collaborative fraud prevention to reduce losses and protect legitimate customers.
Speakers
- Chris Uriarte, Partner, Glenbrook
- Samantha Gordon, Senior Engagement Manager, Glenbrook Partners
- Karamjit Singh, Vice President, AI Products, Mastercard
Key Takeaways
1. First-Party Fraud: Return and refund fraud has become the top fraud concern for merchants, surpassing traditional third-party fraud, with first-party fraud now accounting for 20–30% of disputes.
2. Return Scheme Losses: Common schemes include wardrobing, returning stolen or counterfeit goods, swapped item fraud, and exploiting lenient return policies, often resulting in significant inventory, shipping, and operational losses.
3. AI-Driven Fraud: AI tools and social media are accelerating the growth and sophistication of both first-party and third-party return fraud, making detection and prevention increasingly challenging.
4. Network Data Intelligence: MasterCard’s Return Risk Intelligence pilot leverages network-wide transaction data, AI, and collaboration with merchants to identify high-risk return behaviors such as cross-merchant abuse, chargeback returners, and double dipping.
5. Layered Risk Management: Collaborative, layered risk management combining network intelligence with merchant-level data is essential to proactively detect patterns, reduce disputes, and protect the retail ecosystem.
Key Quote
The goal is simple and, you know, make security smarter, more connected and ready for the future.
Related Content
Explore Related Content.
Webinar
Watch Full Webinar here.
FAQs: FraudWatch: a Glenbrook & Mastercard Quarterly Webinar
Understanding Return and Refund Fraud
What is refund fraud?
Refund fraud occurs when a customer seeks to obtain a refund without legitimately returning the purchased product. This can include false claims such as stating an item was not received despite proof of delivery.
What is return fraud?
Return fraud happens when a customer returns something other than what was purchased or exploits a retailer’s return policy. Examples include swapping an expensive item for a cheaper counterfeit or repeatedly buying and returning items with no intent to keep them.
How is first party fraud related to return and refund fraud?
First party fraud involves legitimate customers misrepresenting themselves to gain financially, such as abusing return policies. In return and refund fraud, the fraudster and the customer are the same person, making detection challenging.
What are common types of return and refund fraud?
Common types include wardrobing (buying items to use once and return), returning stolen merchandise, swapped item fraud, 'brick in a box' scams, 'did not receive' claims, and policy abuse involving lenient return rules.
Impact on the Ecosystem
How does return and refund fraud affect merchants?
Merchants face losses from chargebacks, lost inventory, shipping costs, and customer service burdens. They may also need to invest in third-party detection tools, and policy tightening can negatively impact legitimate customers.
How does return and refund fraud affect issuers?
Issuers incur operational costs from handling disputes and may absorb losses, especially on low-value transactions. Increased fraud can lead to stricter policies that may inconvenience genuine cardholders.
Why is this type of fraud difficult to detect?
It often involves legitimate customers with established histories, making it hard to distinguish between normal and abusive behavior. Data is fragmented across merchants, and there is no standardized tagging for violations.
Scale and Trends
How widespread is return and refund fraud?
Surveys show it is now the top fraud concern among merchants, surpassing traditional third-party fraud. The NRF reported over $700 billion in merchandise returns in 2023, with a return rate of about 15%.
What role does AI play in the growth of this fraud?
AI tools can generate convincing fake documents and receipts, making fraudulent claims easier. They also enable more sophisticated schemes, increasing both the scale and complexity of attacks.
How does organized crime contribute to return and refund fraud?
Organized groups may exploit return policies using stolen payment credentials or compromised merchant systems, blending third-party fraud tactics with return fraud schemes.
Mastercard’s Approach
What is Mastercard’s Return Risk Intelligence (RRI)?
RRI is a pilot program using Mastercard’s network-wide transaction data and AI to identify risky return behaviors, such as cross-merchant abuse, chargeback returners, and returns linked to compromised cards.
How does RRI work?
It categorizes customer behavior into risk levels and provides reason codes explaining the patterns detected. This intelligence can be combined with merchant data to create a more complete risk profile.
What is the goal of RRI?
The goal is to proactively identify and mitigate fraudulent return patterns, reduce disputes, and foster a healthier, more trusted ecosystem through collaboration between merchants, issuers, and Mastercard.
Emerging Considerations
How could agentic commerce affect return and refund fraud?
Agentic commerce, where AI agents make purchases, could increase returns by automating price comparisons and initiating returns when better deals are found, potentially amplifying policy abuse.
What new fraud vector has been identified involving virtual cards?
Fraudsters may use single-use virtual card numbers to make purchases and then return items, avoiding detection by merchants who cannot link the tokenized card to the primary account number.
Blog: Strategies to Detect and Prevent First-Party Return Fraud
Return and refund fraud has become a significant challenge in the retail and payments ecosystem, fueled by the growth of e-commerce and the widespread adoption of flexible return policies. These policies, intended to build customer trust and loyalty, have opened the door to abuse by individuals who exploit them for financial gain. Common schemes include obtaining refunds without returning products, replacing purchased goods with cheaper substitutes, and repeatedly returning items after use.
The financial impact is substantial, with merchandise returns exceeding $700 billion in 2023—around 15% of total retail sales—translating to an average loss of $13–$14 for every $100 in merchandise sold. This issue spans from first-party fraud, such as wardrobing and returning counterfeit goods, to organized third-party fraud involving stolen payment credentials.
Distinguishing between genuine returns and fraudulent activity is increasingly difficult, as some consumers perceive behaviors like bulk purchasing with high return rates as acceptable. This complexity underscores the need for effective detection and prevention strategies that protect profitability while maintaining customer satisfaction.
Preventing First-Party Return Fraud
Return and refund fraud often overlaps with first-party fraud, where a legitimate customer misrepresents their actions. Unlike third-party fraud involving stolen credentials or compromised accounts, first-party fraud is harder to detect because the perpetrator’s identity is genuine. Indicators typically include repeated high-value returns or patterns of policy exploitation. Common examples are “wardrobing,” where items are purchased for temporary use before being returned, and “brick in a box” schemes, where worthless items are sent back instead of the original product. Organized criminal groups also exploit vulnerabilities, such as cloning point-of-sale devices to process fraudulent refunds to prepaid cards. These tactics demand detection methods that can distinguish genuine customer behavior from fraudulent intent.
The operational and financial impact on merchants is significant. Losses extend beyond goods and revenue to higher shipping costs, increased customer service workloads, and expenses tied to third-party fraud detection tools. Issuers face a surge in disputes and chargebacks, often absorbing losses on low-value transactions to avoid resolution costs. Both merchants and issuers bear the burden, while legitimate customers are affected by stricter return policies aimed at curbing abuse. This environment calls for coordinated prevention strategies, with data sharing and aligned policy enforcement to stop fraud before escalation.
Effective mitigation requires policy adjustments, advanced analytics, and industry-wide cooperation. AI and machine learning can analyze transaction histories, return patterns, and behavioral anomalies to flag suspicious activity in real time. Structured evidence programs—such as proof of delivery or product condition—help issuers make informed decisions and reduce wrongful chargebacks. Merchants must balance prevention with customer experience to avoid penalizing legitimate buyers. Education campaigns can deter casual abuse promoted through social media and reinforce the consequences of fraudulent behavior.
First-party fraud is challenging to address because it often exists in a gray area between acceptable behavior and exploitation. Retailers aim to maintain customer-friendly policies to drive sales, yet these can be manipulated. Detection systems must identify when return behavior becomes fraudulent without creating false positives that harm customer relationships. Models should differentiate between occasional high-volume returns and patterns that signal abuse.
Advances in AI have intensified the problem by enabling fraudsters to produce convincing fake receipts and documents quickly, accelerating third-party return fraud. AI-driven schemes are more sophisticated and scalable, allowing criminals to exploit weaknesses at unprecedented speed. This increases the likelihood of faster-than-expected growth in fraud rates, making rapid adaptation essential. Combating AI-enabled fraud depends on integrating diverse data sources and leveraging network-wide intelligence.
Breaking down data silos offers a strong defense. Combining merchant-level SKU data with network-level transaction and chargeback information can reveal cross-merchant abuse, double-dipping, and returns involving compromised cards. Network operators can classify risky behaviors, assign transparent reason codes, and feed this intelligence into merchants’ and issuers’ risk models. This layered strategy enables proactive fraud prevention, reduces disputes, and strengthens trust across the retail ecosystem.
Return and refund fraud is a growing threat that requires the retail and payments industries to adapt quickly. As e-commerce expands and customer expectations for convenience stay high, businesses must balance protection and experience. Effective defense depends on using advanced technology, refining return policies, and strengthening collaboration among merchants, issuers, and solution providers. With AI driving more sophisticated fraud tactics and new commerce models increasing return volumes, proactive action is essential. By combining network intelligence, merchant insights, and analytics, the industry can detect abuse, safeguard revenue, and sustain customer trust—ensuring a secure, reliable environment for legitimate transactions.
