Fraud remains one of the most significant challenges across industries, especially in retail. The 2023 NRF Retail Security Survey reports that retail shrinkage accounts for 1.4% of retail revenue, equating to a staggering $112.1 billion in losses. Since fraud, waste, and abuse contribute significantly to this shrinkage, addressing these issues is crucial for retailers seeking to reduce losses and maintain profitability. I recently had the opportunity to explore trends and solutions with our partner Break In Data (BID) at SAS Inspiration Day in Warsaw. Here are our top takeaways from the event:
Most People Do Not Realize How Long It Takes to Detect Fraud.
During the presentation, we asked the audience how long they thought it took to detect most frauds. The first suggestion was one day, then "a few weeks," but the reality is quite different. According to the Association of Certified Fraud Examiners (ACFE) 2022 Report to the Nations, it takes around 16 months on average to detect fraud. This significant delay underscores the potential for substantial financial losses over time. By leveraging advanced analytics, artificial intelligence (AI), and machine learning, retailers can significantly expedite the fraud detection process, reducing losses and improving response times.
The Biggest Fraud Issue Is (Still) Theft.
Allegra Adinolfi from BID shared insights from a recent roundtable involving retailers from different industries, including the fashion industry. All participants identified theft as their biggest fraud issue, closely followed by loyalty card fraud and fake returns. While it might be tempting to assume that most fraud today involves sophisticated technology, much of it remains low-tech. However, technology, especially generative AI, can still play a crucial role in combating these traditional forms of fraud. For instance, AI is already being used to analyze in-store video footage, identifying potential shoplifting incidents and helping retailers respond proactively.
Fake Returns Are Being Encouraged and Supported Through Online Content.
Fraudulent returns are a growing problem, with fraudsters even sharing detailed guides on online forums about how to carry out fake returns. This phenomenon, often referred to as 'fraud as a service,' enables fraudsters to exploit retailers on a larger scale, making it a prime target for organized crime networks. Some of these guides include step-by-step instructions on how to approach customer service, what to say to secure a refund, and even how to keep the product. Retailers need to be aware of this trend and strengthen their return policies and fraud detection systems to counteract these schemes.
Ghost Entities Are a Widespread Issue.
There are two primary categories of ghost entities. The first involves fictitious or non-existent suppliers submitting invoices for goods or services that were never provided. The second category pertains to ghost employees, a form of HR fraud, where nonexistent personnel are added to payroll, resulting in unauthorized payments. I am fascinated that there is basically a whole world of ghosts out there that may be creating significant amounts of waste for the company. Many retailers do not believe that they have this problem, but when we start work with customers, we often find that up to 5% of procurement is impacted by fraud, waste and abuse in some way. This is confirmed by the Association of Certified Fraud Examiners, that estimates that up to 4,76% of annual spend is at risk of fraud.
Collusion Can Be Detected Through Network Analysis.
Collusion doesn’t necessarily show up when you ask retailers about the frauds and issues they encounter because it is almost invisible. You might observe the financial losses, but it’s often unclear that individual losses are linked. However, using network and link analytics, it becomes possible to expose complex networks connecting customers, employees, and suppliers.
For example, in a retail store environment, a store manager might collaborate with a supplier to manipulate inventory records. The supplier delivers fewer products than ordered, but the manager approves the full invoice as if the correct amount had been received. In exchange, the manager receives cash, gifts, or services from the supplier. At the store level, this can manifest as inventory discrepancies, shrinkage, or inflated costs attributed to errors or mismanagement. This type of collusion is often hard to detect, as both the supplier and the manager ensure all paperwork appears legitimate. However, by leveraging network analytics, unusual patterns between purchase orders, deliveries, and inventory data can expose the hidden relationship between the manager and the supplier.
New Types of Fraud Are Emerging, and AI Plays a Dual Role
Fraudsters are constantly evolving, and one of the emerging threats involves the misuse of generative AI. While AI offers tremendous potential for fraud detection and analysis, fraudsters are also harnessing its capabilities, such as creating synthetic identities. This reality highlights the need for retailers to employ a diverse set of techniques and technologies, combining AI, machine learning, and human insights, to stay ahead of fraudsters who are equally tech-savvy.
Data Quality Is One of the Biggest Challenges in Combating Fraud
One of the most significant challenges in combating retail fraud is related to data quality. Both Allegra Adinolfi and Andrea Ferrari emphasized that data is pivotal in effectively countering fraud. However, two primary issues stand out: missing data, where critical information is not captured or lost, and dirty data, where inaccuracies or inconsistencies reduce the data's reliability. Retailers must ensure the integrity and completeness of their data to derive actionable insights and strengthen their anti-fraud strategies.
One of the most powerful methods for generating actionable insights is through advanced analytics solutions. By combining BID's expertise in retail with SAS’s advanced analytics and machine learning capabilities, retailers can transform raw data into valuable insights that directly impact store performance and reduce losses.
For example, in a case of unexplained inventory shrinkage, BID and SAS solution can integrate various data sources—such as sales transactions, inventory levels, employee shift records, and customer return data—into a unified analytics platform. Using advanced analytics, these systems can detect patterns and anomalies, such as specific employees being consistently present during high-loss periods or irregularities in refund transactions, which may point to internal theft or fraudulent returns.
SAS’s powerful analytics models can then apply predictive algorithms to forecast potential areas of risk, allowing retailers to proactively address vulnerabilities in their operations. Interactive dashboards allow loss prevention teams to visualize this data in real time, helping identify loss hotspots or operational inefficiencies that may be contributing to revenue leakage.
Conclusion: The Value of the BID and SAS Partnership
The collaboration between BID and SAS equips retailers with cutting-edge tools to detect and prevent fraud while optimizing operational efficiency. By generating actionable insights, retailers can enhance profitability, improve operational processes, and fortify their defenses against emerging fraud trends. In today’s fast-paced retail environment, leveraging such advanced analytics solutions is not just beneficial—it’s essential for maintaining a competitive edge and protecting the bottom line.
PS: In case you'd like to see this key note presentation in its entirety, you can do so by clicking this link.
... View more
