Catch Me If You Can: 3 Retail Fraud Schemes Stopped by Data Analytics
Data analytics can detect all types of retail fraud.
Every year, retail fraud gets harder and harder to catch.
Whether driven by greed or desperation, modern retail criminals have a multitude of different tactics for slipping under the radar and stealing billions in cash and merchandise.
But there is one fraud countermeasure that perpetrators have yet to find a way to beat: Data analytics. Nearly all fraudulent actions taken by criminals create records in retailers’ data, making it a treasure trove of information retailers can leverage to identify activities sapping profits and margins. The trick lies in finding the right analytics solution, one that can recognize exactly what data behaviors and patterns indicate fraud and alert the right person.
In my 12-plus years in retail, I’ve found the most success in detecting and preventing fraud with prescriptive analytics, the software methodology that analyzes data and tells you:
What is happening.
Why it happened.
How much it is costing you.
What to do about it.
Who should do it.
As examples, here are three major real-life retail fraud cases that prescriptive analytics identified and helped eliminate:
The asset protection (AP) team for a convenience store chain received an alert from its new prescriptive analytics solution regarding some strange activity at a Texas location. According to the pattern (an algorithm that identifies signs of fraud), this store appeared to be processing significantly more refunds than similarly performing stores in its region.
The solution found that the refunds were above average in terms of both frequency and dollar value, also indicating that two specific managers had authorized almost all of them. Prompted by this alert, the AP investigators launched an investigation.
Interviews with both authorizing managers revealed a fraud scheme at play. The location’s assistant manager, who had some financial knowledge, had found a way to steal cash by padding his store’s inventory and “returning” the non-existent products to himself. Worse, CCTV footage showed the assistant manager teaching the store manager to pull off this scheme too. All told, the duo had stolen nearly $30,000 via the phony refunds.
The use of prescriptive analytics allowed the AP team to identify and resolve this case quickly, before the losses racked up even further.
Cross-Country Call Center Caper
An apparel retailer adopted prescriptive analytics as part of an initiative against e-commerce fraud. The solution’s machine learning and AI capabilities identified and alerted its asset protection team to suspicious behavior at a call center.
As many retailers do, this chain allowed its customer service representatives (CSRs) to send “appeasement” shipments – free replacement merchandise, along with a discount offer or gift card in the hope of correcting a previously negative customer experience. The solution identified several CSRs who were appeasing many, many more e-commerce orders than would be expected if they were operating within benchmark levels.
It further identified the suspect CSRs were repeatedly shipping the appeased orders to just five addresses. The prescriptive action for the AP team was to investigate the suspect CSRs for potential organized retail crime (ORC) activity.
The prescriptive action led AP investigators straight to the root cause – these employees had indeed formed an ORC ring. Working together, they’d legitimately buy a product online and, after receiving it, would call the center to complain they never got the product. Thus, they ended up with a $20 gift card and two products that they could sell online for cash. Worse, the ring had even enlisted additional individuals to apply for jobs at the call center, meaning that more conspirators were being hired every week.
With this information, the retailer took disciplinary action and broke up the group, saving an average of $50,000 a month.
A little over a year later, this same capability again paid dividends. The retailer re-configured the pattern to monitor its newly opened call center in Latin America for excessive appeasements. Sure enough, many CSRs began appeasing far more orders than would be expected according to benchmark levels, shortly after the call center was opened and the pattern deployed.
An investigation uncovered evidence of ORC. After their shifts, the conspiring employees would contact customers they had serviced earlier in the day, offering to send them free products in exchange for a kickback. In just a few short weeks, the retailer had lost nearly $75,000 to this scheme.
The call center operator was able to act swiftly to remove the dishonest employees from the organization. It also issued the retailer full credit for its losses.
A hardware retailer deployed a pattern via its prescriptive analytics solution to provide better visibility into SKU movements. The pattern quickly identified a single store at which a select few SKUs showed low sales but high rates of returns and exchanges.
A check of CCTV footage revealed a cashier returning and exchanging the items in question (the return activity was what drove down the net sales and triggered the alert), with no customers present when he did so. He repeated this behavior multiple times throughout the week, clearly indicating to the AP team that something was amiss.
An investigation and interview with the cashier revealed that he was stealing popular items from his store and selling them via his own online resale business.
The extensive return/exchange activity had an interesting explanation too – occasionally an online customer would send back the stolen items as defective. If said customer wanted a replacement, the cashier simply brought the defective item back to his store and exchanged it for a new one. If the customer didn’t want a replacement, the cashier would refund the stolen item to himself and start over.
The prompt alert empowered the retailer’s AP team to intervene while losses were still relatively low and stop the scheme.