As e-commerce eats into physical retail’s share of the total retail market, there has been a coincidental increase in incidences of returns fraud. This development, while shocking, is not entirely unexpected. While returns have always been a part of doing business in retail, returns fraud did not cast a long shadow as it does today.
This is thanks to e-commerce, which creates the perfect environment for bad actors to mushroom in this space. Unlike product returns to physical storefronts, returns arising from e-commerce tend to make their way to a distribution center, where there is a higher chance of it being checked by an employee who cannot identify product nuances like their peers in a retail store.
The National Retail Federation (NRF) estimated that more than $100 billion in merchandise was returned fraudulently in the US in 20231, amounting to about 13.6% of the overall returned goods that year. On a microscopic level, these numbers are starker — companies lose an average of $13.70 to return fraud for every $100 of returned merchandise accepted.
Uncovering the Tactics Used in Return Fraud
While businesses continue to strengthen their defenses against fraud, return fraudsters have kept up with the times as well. They employ a litany of tactics, weaponizing evolving customer behavior and generous return policies to exploit loopholes in retailers’ systems.
Wardrobing is a common strategy in the fashion and apparel sector, where fraudsters ‘buy’ a product to return it after using it for a specific event — like a wedding, party, or photoshoot. With the ‘no questions asked’ return policies prevalent in the market today, returning the purchases saying they have changed their minds or that it is unwanted is frequent.
While box switching or swapping products for their likeliness is one of the oldest fraud strategies, bad actors are going a step above today by replacing parts with a cheaper or defective version, especially in electronic products. This can go unnoticed by retail staff and can get added back to the inventory. If resold, the business might face the ire of genuine customers who end up with these defective products, impacting brand value.
The rise of omnichannel has complicated product returns for businesses, as customers can now buy a product via one channel and return it in another. Fraudsters may buy products online, use them for a while, and then return them to a physical store, or vice versa. Fake or altered receipts from fraudulent websites or stolen payment card information can be used to return goods in-store when never purchased.
Leveraging Technology to Resist Fraud at the Point of Entry
In a scenario where bad actors are increasingly adept at committing fraud, technology can come to the rescue. To begin with, employing a returns management system (RMS)2 is crucial to plugging gaps in the returns process while providing clear visibility into the cost of returns operations and the capital lost to returns fraud. A quality RMS can mitigate fraud by managing and orchestrating each aspect of returns management, ensuring workflows and policies are repeatable and standardized.
Data is crucial here. The end-to-end product and logistics data an RMS collects and measures provide businesses with visibility into every item’s purchase history, detailed product information, and customer profiles. This data can enable AI-based solutions to spot returns fraud in real-time.
For instance, AI can flag suspicious accounts based on return rates and the type of products being sent back. High-value items are especially prone to return fraud, and AI can detect red flags when returns do not align with a typical buyer’s behavior or if there is a short purchase history. Technology can also detect return timing anomalies, alerting the business when the returned product falls outside typical return windows or within suspicious time frames.
AI can also analyze customer sentiment through chats, feedback, social media, or reviews. For instance, inconsistent or contradictory statements on different communication threads can signal that the return request might be fraudulent.
Tech-Based Product Inspections to Identify Fraudulent Returns
While AI-based detection systems help flag returns, businesses are unlikely to outright refuse these return requests as not all flagged returns tend to be fraudulent. However, these alerts help ensure the flagged items are inspected more closely than other returns.
Technology like computer vision plays a huge role in expediting the inspection process and increasing the accuracy of detecting fraud. Computer vision can be trained to assess the condition of returned items based on images or videos taken at the time of the return. Analyzing visual features like wear, scratches, dents, missing parts, or altered packaging can quickly determine whether an item has been used or tampered with.
A common fraud tactic is product swapping, in which the fraudster returns a different or inferior product instead of the one they originally purchased. Computer vision can match returned items with original products through advanced image recognition algorithms. It can also verify the authenticity of returned goods — especially high-end or luxury items — by analyzing patterns, logos, colors, stitching, and materials that are typically difficult to distinguish manually.
Computer vision can automate the sorting and routing of returns by assessing their condition and categorizing them as ‘eligible for refund’ or ‘requires further inspection.’ This reduces human workload, allowing them to be more selective about the products they inspect. All the visual data collected by these systems can be fed into a broader fraud detection system, enabling AI to learn and evolve by recognizing patterns of consumer behavior and the state of product returns.
Personalizing Return Policies for Mitigating Fraudulent Behavior
While technology can help improve how returns are handled and processed, the resulting operational data can be used to tweak return policies, matching them to individual customer behavior. AI can analyze a customer’s purchase and return history to create tailor-made return windows. For instance, if a customer with a high return-to-purchase ratio is found to consistently return items within a week of purchase, the system can shorten their return window for future purchases while offering more generous return periods to loyal, low-return customers.
Further, AI can help offer dynamic return policies based on product category and customer history. High-end electronics may have a stricter return window than clothing, or items with high rates of fraudulent returns could have tighter return rules. This allows flexibility while enabling the retailer to monitor items at a higher risk for abuse.
The idea is to restrict bad actors who abuse the system. Once a profile is flagged, returns from that account must go for a more thorough review each time, besides increasing restrictions — like needing to return items with receipts or items in original packaging. In cases of genuine confusion and mistaken flagging, the AI system can help agents provide an efficient resolution, minimizing friction for the customer.
Ultimately, as a retailer looking to reduce returns fraud, using technology in returns management operations can bring about comprehensive results, purpose-built returns management systems3 can reduce the impact of returns fraud on your bottom line, along with radically improving the process of collecting, inspecting, and adding returned items back to the inventory.
1.
2.
3.
As Chief Commercial Officer at ReverseLogix, Chuck plays a pivotal role in driving the company’s growth strategy. He is responsible for the company’s marketing, sales, and go-to-market strategy.
Chuck has spent over 20 years in supply chain technology, helping to build market-leading businesses. He leverages a long track record of scaling companies to help ReverseLogix achieve peak growth through integrated marketing, customer-centric sales, and go-to-market strategy. Before joining ReverseLogix, he served as the director of product marketing at Plex Systems, a SaaS-based ERP platform.
Prior to Plex Systems, he was vice president of marketing for 3G, a hypergrowth transportation management system (TMS) start-up, led global product strategy for HighJump (now Körber), and held marketing leadership positions at Lawson Software (now Infor) and RSM.