On Feb. 9, Redner’s Markets Inc. loss prevention employee Ryan Christman saw a woman who appeared via surveillance video to be underscanning high-value items.
Christman was familiar with the woman from footage he reviewed of previous thefts, all or most of which involved gourmet cheese and cured meat products, at the Redner’s Warehouse Markets grocery store.
He called the local police department, saying he had a shoplifting suspect in his office. According to the criminal complaint, he told the police officer who responded that the woman had been stealing from the same grocery store for the past couple of months.
She would substitute the lookup code for bananas — pound for pound one of the cheapest products in the store — for more expensive items.
Christman later provided the officer with exact dates of the prior thefts. The officer checked the footage for herself.
From Aug. 14 through Feb. 7, according to police, Redner’s reported a total of 45 of these types of thefts totaling nearly $5,000 committed by the woman.
“She admitted to me and the police that she ran a business, on the side, and a lot of the products that she was stealing funded that business,” Christman said in an interview with the Reading Eagle.
The thefts would likely have gone unnoticed if not for the use of artificial intelligence.
With the help of predictive analytics software, a pattern was identified, including when she was likely to shop again.
Sure enough, as predicted, the woman returned the next day.
Her case was one of those highlighted by Redner’s executives during the interview in the corporate office in Ontelaunee Township on their use of AI to catch shoplifters.
For the past year, Redner’s has been employing the technology to identify anomalies in shopping patterns – red flags for further investigation of a customer’s transactions.
In a business in which keeping prices low is essential, it’s not efficient to have loss prevention agents at every store, much less watching every transaction.
AI saves Redner’s loss prevention officials countless hours by gleaning from various data points to flag transactions or sets of transactions for a closer look, said Eric B. White, Redner’s director of marketing and communication.
Each transaction creates an electronic log, data which is fed to various departments in different ways, from inventory control to operational control to loss prevention.
“So the software helps glean through it and points to areas where you may want to look at these or you may want to take a look at this and that’s when you can really take a deeper dive,” White said.
Unlike the traditional shoplifter who leaves the store with a cartful of groceries without going through the checkout or hides items under their coat or in their handbag, White said, this kind of retail theft amounts to white-collar crime.
“It’s calculated,” he said. “Someone’s calculating how they can manipulate the system to get what they need on multiple occasions.”
Redner’s officials are astounded by some of the backgrounds of people who have been caught, with the help of the AI software, repeatedly shoplifting. Among them are a health care professional, a scholastic athletics coach and a former school board member — people who, one would expect, have the means to pay for their own groceries.
The health care professional had been stealing from two Redner’s Markets stores for several months last summer and fall, employing a checkout-scanning ruse that is difficult to detect in real time.
Her sleight of hand involved scanning inexpensive items such as Kool-Aid powdered drink packets and Tic Tac mint candy while making it appear as though she had scanned higher-value items such as diamond steak and Triscuit crackers. The register would ring up Kool-Aid packs that cost less than a quarter apiece instead of the steak.
Loss prevention officials figured out that the woman worked at two different offices.
“We could basically predict which store she was going to go to on which day by her schedule and by our software,” Christman said.
In mid-November, after being fed the data, Christman’s team learned there was a high probability that she would be at the store at 3205 North Fifth Street Highway the next day.
Like clockwork, she returned to that store as predicted and was detained and later charged by Muhlenberg Township police with retail theft.
Redner’s staff provided police with evidence that she had been stealing from the store between July and November at both the Muhlenberg store and the Redner’s Warehouse Markets store in the Sinking Spring Shopping Center along Route 422 in South Heidelberg Township.
South Heidelberg police separately filed retail theft charges against the woman.
While predictive analytics software is extremely helpful in flagging suspicious behavior, it still takes human intelligence and tediously poring through surveillance footage to prove thefts have occurred and predict when the offender will next be back, officials said.
Jay Schaeffer, Redner’s senior loss prevention supervisor, is adept at analyzing video to identify when a customer’s scanning behavior doesn’t line up with their transaction log.
Self-checkout is preferred by at least half of the customers, so Redner’s would prefer to continue to offer it as a convenience to customers, White said.
Plus, it’s a small percentage of people who are dishonest, so it’s neither efficient nor good from a customer-relations standpoint to watch every transaction as if a theft is about to take place.
“Our mission is to create great rewarding experiences by being a great place to shop and work,” White said.
With 44 stores throughout three states, however, even a small percentage of transactions that involve theft adds up to a significant loss, costing Redner’s an estimated $1 million annually, said John Flickinger, Redner’s director of risk management and loss prevention.
The data show shoplifting is prevalent even in more affluent areas.
“We’re finding these thefts that are occurring in locations that we don’t necessarily have many issues with,” Flickinger said.
With grocery prices in the news a lot these days because people have tangible evidence of inflation at the checkout, supermarkets are under pressure to curtail theft to keep prices down.
“It’s very important for us to keep our prices as low as we can for our guests,” Flickinger said.
Retail stores have to factor theft into their pricing, especially when shoplifting is impacting their ability to make their margins, he explained.
“It’s very frustrating and it affects everyone,” Flickinger said.
