(Guest Post by Malay Kundu)
Self-checkouts are supposed to save retailers in labor costs, because they do not need a cashier and, theoretically, the customer can do the checking out him or herself. But some retailers are finding that they may be costing more than they save, as incidents of customer theft grow.
Big Y, with 61 Massachusetts and Connecticut stores, and Albertsons, with 217 stores in the South and West, have done away with self-checkout in order to foster more human contact and better customer service rather than having customers struggle with bar codes, coupons and payment. On the other hand, CVS Caremark Corp. recently implemented self-checkout in some urban markets to make shopping faster and more convenient while saving on labor.
But are retailers really saving money?
At StopLift, we are receiving many more requests for our self-checkout video analytics for finding theft and fraud at the self-checkout, in particular detecting merchandise leaving the store with customers unscanned. Chains are waking up to the fact that the siren call of labor savings is often coupled with the danger of increased shrink. And now they are trying to do something about it.
We initially developed scan avoidance detection technology for retail chains to be implemented with manned checkouts. Our ScanItAll, patented computer vision technology, interprets the behavior of the cashier and customer by analyzing and understanding the operator’s body motions and interactions with items of merchandise at the checkout. If you visit StopLift.com, clicking on the bottom right of the video window on the home page will show you real videos of customers stealing at the self-checkout.
The dirty little secret is that it’s very easy to steal at self-checkout, and that is why the incidents among many retailers are growing. Self-checkouts are almost an immediate ‘benefit of the doubt’ afforded to every customer, honest and dishonest alike. And there is no better repeat customer than one who gets free merchandise.
Here are just a few examples of how people steal in the self-checkout:
- Leaving more expensive items in the shopping cart, while only scanning some of the less expensive items.
- Leaving items in a recyclable or reusable bag on the floor, while only scanning a couple of other items. Many self-checkouts read the presence of a recyclable/reusable bag in the output area as an unscanned object, prompting the customer to put it on the floor anyway. How easy is it to take just a few items out and scan them, before returning them to that bag of other goodies on the floor?
- The “banana trick”. Even though the customer has a $10/pound sirloin steak on the scale, with the bar code facing upward, he or she can indicate it as produce and click on “Bananas” at 49 cents a pound.
- Overloading the weight scale so that additional unscanned merchandise is not sensed by the self-checkout.
- Scanning an item with one hand while dropping another into a reusable bag on the floor.
These are just the tip of the iceberg. It never fails to amaze me just how creative and ingenious people can be in figuring out how to steal in the self-checkout. The beauty of it from the customer’s standpoint is this: Every item in their order is a chance to purposely try – or inadvertently discover – how to beat the system. And if the customer is noticed incorrectly ringing up the transaction or skipping an item, he/she can always plead ignorance about how to properly use the system.
That brings me to the self-checkout attendants. While their primary purpose is to be available to help customers at self-checkout, they are also supposed to watch and keep customers from trying to game the system. Even with complete attention and presence of mind, keeping an eye out for customer theft is a very tough job. Compounding the problem, however, is the fact that these attendants are often not around. They may be helping another customer, or they may be sent off on another task by store managers who see them as “extra bodies” just standing around. Or, as I’ve witnessed personally at my local grocery store, the attendant might be at his or her station texting. Without attendants available and attentive, not only does customer service falter, so does security against customer fraud.
The growth of self-checkouts throughout the retail and grocery industries has increased demand for self-checkout video analytics. That, in turn, is driving deployment of cameras above the self-checkout registers just as it has with manned checkouts. Although StopLift’s systems work with both kinds of cameras, while retailers previously defaulted for analog CCTV cameras, we now see them opting more and more to use our system with digital IP cameras over self-checkouts in order to be “future proof”. Since retailers also want to react directly from the shop floor with mobile smart phone and tablets, we see increased interest in our mobile smart phone and tablet applications as well.
As self-checkouts continue to grow in popularity, one thing is for sure: we will continue to see a growing market for video analytics and related technology to help retailers catch the wrong kind of repeat customer.
StopLift Checkout Vision Systems has developed software-based checkout vision systems which automatically analyze regular CCTV video from existing cameras to detect various forms of theft, training error, and operational analytics at the checkout. A pioneer in the field of checkout vision systems, StopLift has developed the first ever system capable of successfully detecting "sweethearting" between cashiers and customers (U.S. Patents 7,516,888 & 7,631,808).
Rather than take a one-size-fits-all approach, StopLift has developed targeted applications to address the specific needs of retailers from different sectors, including general merchandise, grocery, and specialty retail.
StopLift grew out of a Harvard Business School field study on Retail Loss Prevention entitled "Project StopLift". One of the study's major findings was that, while CCTV is the most widespread of all loss prevention technologies, it is often the most underutilized--it is just too expensive and time-consuming for humans to monitor or review video comprehensively.
With engineering talent and computer vision research insights from MIT, the Project StopLift team realized that video recognition could be used to automate and, thus, make possible the comprehensive examination of surveillance video.
How to Submit a Guest Post - Scroll down and click on the link on the Footer menu