The following is an excerpt from my book The New Rules of Consumer Privacy.
When it comes to the future of retail and face recognition, I’m fortunate to have a front-row seat for the revolution. Years ago, the industry’s early adopters recognized that face recognition could be used to create a better experience for customers. With the technology now proven, many more are gearing up to help create more secure purchases and prevent shoplifting and in-store violent crime.
Consumers are just beginning to experience this shift. AI is going to be the fuel that will power data services in the very near future. Consider the retail marketplace and the age-old transaction platform that brings buyers and sellers together. Not much has changed since the very beginning of this exchange. Of course, we have been collecting data about the people and their merchandise for many years. However, this has largely been a human-powered endeavor with only relatively simple tools at our disposal, like spreadsheets and mathematical programs.
Now consider what AI brings to the equation: a level of data collection, analysis and real learning that we’ve never seen before, performed at a speed that seemed impossible only a decade ago. Setting aside the creepy nature of a restaurant knowing what you want before you have made a conscience decision to choose the oatmeal with the berries, imagine a world where machines anticipate your ever need and, in fact, make better decisions about your life than you can. We’ve barely seen the beginning of the power behind machine learning and neural networks.
In retail, more than almost any other sector, consumer privacy considerations are paramount, and winning retailers need to be responsible stewards of customer biometric data. I believe this will require a balance between correcting misconceptions about the technology and entering into a contract of sorts with consumers. Here’s how the balance between innovation and privacy will play out in retail.
Organized Retail Crime
The term organized retail crime (ORC) describes coordinated shoplifting and return fraud by multiple individuals. According to a National Retail Federation (NRF) survey, 94.6% of all surveyed retailers believe they have been impacted by ORC. Responding companies reported an average of loss of $726,351 per $1 billion in annual sales volume due to ORC activities. ORC remains the leading source of inventory shrink, outpacing employee theft and other forms of shrinkage, while the average theft for retail stores is a whopping $559 per incident.
One potential reason for the epidemic may be rising theft felony threshold rates (see map). In Florida, Massachusetts, Virginia and New Jersey, stealing a purse worth $300 or more is a felony. But in many other states, recent rule changes have removed the deterrence for the average theft. In Georgia, the felony threshold is $1,500, while in Wisconsin, it’s $2,500. In my home state of California, it’s now $950. While some groups have lauded the lighter sentences, most large retailers have seen ORC skyrocket.
In a FaceFirst shoplifting recidivism study of biometric data spanning a six-month period, FaceFirst found that 60% of documented shoplifters were detected entering at least two separate locations of the same retail chain, while 20% visited three or more locations. It seems that shoplifters are incredibly loyal to their favorite brands.
The study challenges data from the National Association of Shoplifting Prevention (NASP) indicating that the vast majority of shoplifting is not premeditated. Unfortunately, it isn’t just stores that lose when ORC hits. Consumers and employees are hit even harder. Sometimes the results can be tragic. The NRF report reveals that ORC gangs are exhibiting more aggressive and violent behavior than ever before.
Why is there so much retail violence? It typically starts with shoplifting. When a well-meaning employee tries to run down someone leaving with hundreds or thousands of dollars in merchandise, a violent altercation often ensues. All too often, the shoplifter, the employee and even customers get hurt. In 2016 alone, there was more than one retail death per day in the United States, and 2017 saw a 15% increase. There’s also a monetary component, with the cost of a single retail death – including lawsuits, business closure, investigations, and so forth – being almost $2.5 million per incident. These costs are almost certainly passed on to consumers.
For big box retailers, pharmacies and department stores, face recognition surveillance has been an extremely effective solution. Customers of my company, FaceFirst, have decreased external shrink by up to 34% and in-store violent interactions by 91%. Here’s how it works: using a database of known offenders, software and long-range cameras, the system identifies known ORC members the moment they enter the premises. Real-time alerts are sent to in-store personnel.
A new crime is prevented through a simple greeting by a store employee (“Good evening, Mr. Reynolds. May I help you?”). That’s right. The act of identifying an offender by name and offering customer service typically results in the offender leaving peacefully. That’s something that would be completely impossible without face recognition.
With the majority of adults supporting the use of face recognition in retailers’ fight against crime, face recognition surveillance will be standard technology for nearly all major retailers within five years.
In terms of privacy, retailers should follow the TRUST model to the letter. Most companies are already doing a great job with employee training, responsible data handling and anonymous data purging. Where retailers need to improve across the board is transparency. Although only two U.S. states currently require public notification by law, we recommend that retailers always disclose when biometric surveillance is in use, either by signage or some other public notice. And in their biometric surveillance policy, retailers should indicate that anonymous surveillance data is routinely purged.
They should also know that their privacy is being protected. In addition to the Privacy by Design precautions discussed elsewhere in this book, FaceFirst is also exploring additional ways to make it impossible for even a system administrator to access personally identifiable information (PII) from biometric data.
Mobile payment adoption among U.S. consumers appears to have stalled. According to a survey from PYMNTS.com, fewer than one in 20 consumers who have one of the major mobile digital wallet apps use the app when the opportunity arises. The reason? So far, such apps aren’t really more convenient than traditional cash or credit.
Massive fraud hasn’t helped, either, leading the U.S. Small Business Administration (SBA) to publish this 2018 article: “6 Ways to Reduce Mobile Payment Fraud for Your Small Business.” Just when retailers and consumers alike thought mobile payments would solve the massive credit card fraud problem, a new generation of scammers erased their hopes.
Enter face recognition. Alibaba’s Ant Financial affiliate launched secure payment service in Hangzhou, the location of the company’s global HQ, where it is being trialed with KFC. The payment process requires that customers first sign up for the Alipay app and enable face recognition. A 3D camera located at the point-of-sale scans the customer’s face to verify their identity, while there is a phone number verification option for additional security. In California, burger chain Caliburger uses facial recognition to allow customers to pay for food and accrue rewards with a loyalty points system.
As similar systems are rolled out, privacy advocates will no doubt be concerned about introducing a biometric identifier in the payment process. But given the high levels of fraud, and the relative failure of mobile payments to curb the problem, using biometrics for authentication is a highly desirable solution that gives consumers an additional level of protection and security.
Order Peter’s book The New Rules of Consumer Privacy here.