Shoplifting is harmless, right? It’s nothing more than a victimless petty crime.
Besides, it doesn’t really hurt the retailers because they just write off their losses. Shoplifters don’t even face jail time.
If you believe these so-called “facts,” you are buying into the myths surrounding shoplifting that have very little to do with the reality of the crime.
Consider this. According to the National Retail Federation, the loss of inventory from retail stores due to shoplifting and employee theft costs the U.S. retail industry nearly $48.9 billion a year. Moreover, the average cost per shoplifting incident is $798.48.
That’s not exactly “petty.”
According to the National Association of Shoplifting Prevention, there are approximately 27 million shoplifters – one in every 11 people – in our country today. More than 10 million people have been caught shoplifting in the last five years, but shockingly, only one in 478 shoplifters are ever caught and only half of those are turned over to the police for prosecution.
Especially troublesome is the fact that 10 percent of the total dollar losses due to shoplifting are attributable to “professional” shoplifters who steal solely for resale or profit as a business. These include hardened criminals who steal as a lifestyle, international shoplifting gangs who steal for profit, and drug addicts who steal to feed their habit.
That last group is particularly worrisome, given the escalating rate of opioid addiction in the US.
It is estimated that more than two million Americans are now addicted to prescription pain killers, while nearly 600,000 have a substance-use disorder involving heroin. People addicted to these drugs often steal from large retail stores with the intent of returning the stolen merchandise (with no receipt) for a gift card which they can then resell for cash.
While there is no nationwide research showing the connection between opioid addiction and shoplifting, many police departments have direct experience with it. The Knoxville, TN police department found that between 83 and 98 drug overdoses in a three-month window in 2017 were linked directly to gift cards.
Given those staggering numbers, is there anything retailers and the police can do to stem the tide of shoplifting?
Increased on-site security is one option, but many retailers aren’t in a position to pick up the tab. According to the Department of Labor, the average cost for a two-person security team patrolling a typical anchor store in a suburban mall is $51,000 per year – approximately 12.5 percent of total operating costs.
Increasingly, though, police—along with some retailers—are turning to a technology solution that has demonstrated its effectiveness in cracking down on criminals: facial recognition software. Starting last year, a number of retailers across the country began displaying signs to inform customers that management is using facial recognition software, turning the store into a certified safe zone.
While facial recognition software has been in use for more than a decade, retailers and many smaller police departments only began to consider it as a viable tool for targeting shoplifters in the past year or two as prices have dropped.
Facial recognition software works by using image processing and machine learning algorithms to match a photo of an unidentified person (“probe” photo) against a database of photos of identified persons who previously have been convicted of shoplifting or other crimes. The face-identification algorithms in the software will produce a list of possible matches, with each match having a score that indicates the quality or likelihood of a match.
In the past, low resolution, poor lighting, motion blur, off-angle faces, facial hair, and other scenarios have challenged these algorithms to produce a good match. Advances in the technology based on algorithms such as “deep learning,” however, have produced significant gains when processing challenging probe photos.
Despite such advances, even the best facial recognition systems are unlikely to generate just a single match from something like a store security camera photo. Instead, the system will generate a list of possible matches. The police working the case will then need to use standard investigative methods to either rule out or further investigate each match, just as they would with any investigative lead.
In other words, the software isn’t doing anything that wouldn’t occur during a normal police investigation. It is simply doing what investigators would do, but faster and with a higher degree of accuracy.
It is equally important to note that the way in which facial recognition software is currently deployed offers little threat to privacy concerns and limited potential for abuse. Most systems immediately discard images of anyone who isn’t a match for a known shoplifter.
Lack of information or even misinformation, however, can cause a reaction on the part of the public. Could the retailer, for example, collect information on everyone who walks into the store, their buying habits, and so on?
Worse still, could that information be sold to others? As a result, it is important for retailers and law enforcement to fully understand the spectrum of possible uses of the technology, as well as how the public may perceive those uses.
Facial recognition software has the potential to change the rules of retail, generating leads in a great many cases that might otherwise go unsolved. And while each case may not be high profile, in aggregate they represent a staggering amount of criminal activity.
Given that, stopping even a single shoplifter could prevent tens of thousands of dollars in future theft.
Nick Coult is Senior Vice President for Law Enforcement and Public Safety at Numerica Corporation. He is one of the creators of Lumen, a platform for law enforcement search, analysis, and data sharing. Numerica is currently beta-testing a version of their Lumen software called Lumen FR, which integrates next-generation facial recognition algorithms directly into Lumen. Numerica anticipates that Lumen FR will be available by early summer. Readers’ comments are welcome. For more information, visit https://www.numerica.us/