-By Joseph Tarnowski
Supermarket operators continue to rank employee theft their most
severe shrink problem. In
Supermarket Security and Loss
Prevention 2007, FMI estimates that nearly 40 percent of total
shrink stems from employees who steal money or merchandise.
But this is only an estimate based on internal data -- not on
catching thieves in the act. This difficulty in accurately
measuring the losses is especially acute with sweethearting --
cashiers pretending to scan merchandise but deliberately bypassing
the scanner, and thus not charging the customer for the
merchandise. The "customer" is often a friend, family member, or
fellow employee in collusion with the cashier.
"The challenge of the sweethearting problem is that in the past it
has been nearly impossible to quantify the loss, due to the
difficulty in identifying the sweethearting events," says Mark
Gaudette, director of loss prevention at Springfield, Mass.-based
Big Y Foods, Inc., and a member of FMI's Loss Prevention Committee.
"What we do know is that industry statistics from FMI and other
sources indicate that employee theft accounts for 40 percent of all
shrinkage, and that our currently identifiable internal theft is a
very small percentage of that total."
And what retailers can't see is hurting them: According to NRF's
2006 National Retail Security Survey, store employees steal $20
billion worth of merchandise a year, an estimated two-thirds of
that, or $13 billion, through sweethearting.
Supermarkets are particularly vulnerable to sweethearting, and the
practice accounts for almost 35 percent of profit loss
industrywide.
Pinpointing fraud
Gaudette is piloting a new technology at Big Y to help boost
the amount of this "identifiable" theft and reduce shrink --
particularly sweethearting -- at the checkout.
The grocer installed StopLift Checkout Vision Systems' video
recognition software last month in a pilot in several stores that
together form a representative cross-section of its 58 units.
The StopLift software monitors existing security cameras that
capture activity at the checkout registers. Its patent-pending
computer vision technology visually determines what occurs during
each transaction to pinpoint instances of fraud at the checkout.
"The system is capable of identifying the full set of fraudulent
behaviors, including when a cashier covers up a bar code by hand,
stacks items, or carries an item above or around the scanner," says
Malay Kundu, c.e.o. of the Bedford, Mass.-based vendor. "The
computer vision software is specifically designed to mathematically
analyze the body motions of cashiers and their handling of
merchandise at the checkout. Rather than needing to be explicitly
trained, the system adaptively learns to distinguish between
legitimate and suspicious behavior."
This addresses the chief challenge of using security cameras alone:
catching the perpetrator in the act at the time of the incident.
Most security cameras are at best sporadically monitored, notes
Kundu. With StopLift, as soon as a sweethearting incident occurs,
the software flags the transaction as suspicious and immediately
reports it, identifying the cashier and the date and time of the
theft.
"We expect to have control over far more of our shrink and loss
through the use of this emerging technology," says Gaudette.
Exclusive Web Content
ShopLift's nuts and bolts
By mathematically analyzing the pixels of digitized video, the
software scrutinizes exactly how a cashier handles each item, to
determine whether she or her has properly scanned it. The system is
capable of understanding the full set of fraudulent behaviors,
including when a cashier covers up a bar code by hand, or purposely
misaligns the scanner and item so that the item isn't
scanned.
TECHNOLOGY: Goodnight, sweethearting
May 1, 2008
-By Joseph Tarnowski
Supermarket operators continue to rank employee theft their most severe shrink problem. In Supermarket Security and Loss Prevention 2007, FMI estimates that nearly 40 percent of total shrink stems from employees who steal money or merchandise.
But this is only an estimate based on internal data -- not on catching thieves in the act. This difficulty in accurately measuring the losses is especially acute with sweethearting -- cashiers pretending to scan merchandise but deliberately bypassing the scanner, and thus not charging the customer for the merchandise. The "customer" is often a friend, family member, or fellow employee in collusion with the cashier.
"The challenge of the sweethearting problem is that in the past it has been nearly impossible to quantify the loss, due to the difficulty in identifying the sweethearting events," says Mark Gaudette, director of loss prevention at Springfield, Mass.-based Big Y Foods, Inc., and a member of FMI's Loss Prevention Committee. "What we do know is that industry statistics from FMI and other sources indicate that employee theft accounts for 40 percent of all shrinkage, and that our currently identifiable internal theft is a very small percentage of that total."
And what retailers can't see is hurting them: According to NRF's 2006 National Retail Security Survey, store employees steal $20 billion worth of merchandise a year, an estimated two-thirds of that, or $13 billion, through sweethearting.
Supermarkets are particularly vulnerable to sweethearting, and the practice accounts for almost 35 percent of profit loss industrywide.
Pinpointing fraud
Gaudette is piloting a new technology at Big Y to help boost the amount of this "identifiable" theft and reduce shrink -- particularly sweethearting -- at the checkout.
The grocer installed StopLift Checkout Vision Systems' video recognition software last month in a pilot in several stores that together form a representative cross-section of its 58 units.
The StopLift software monitors existing security cameras that capture activity at the checkout registers. Its patent-pending computer vision technology visually determines what occurs during each transaction to pinpoint instances of fraud at the checkout.
"The system is capable of identifying the full set of fraudulent behaviors, including when a cashier covers up a bar code by hand, stacks items, or carries an item above or around the scanner," says Malay Kundu, c.e.o. of the Bedford, Mass.-based vendor. "The computer vision software is specifically designed to mathematically analyze the body motions of cashiers and their handling of merchandise at the checkout. Rather than needing to be explicitly trained, the system adaptively learns to distinguish between legitimate and suspicious behavior."
This addresses the chief challenge of using security cameras alone: catching the perpetrator in the act at the time of the incident. Most security cameras are at best sporadically monitored, notes Kundu. With StopLift, as soon as a sweethearting incident occurs, the software flags the transaction as suspicious and immediately reports it, identifying the cashier and the date and time of the theft.
"We expect to have control over far more of our shrink and loss through the use of this emerging technology," says Gaudette.
Exclusive Web Content
ShopLift's nuts and bolts
By mathematically analyzing the pixels of digitized video, the software scrutinizes exactly how a cashier handles each item, to determine whether she or her has properly scanned it. The system is capable of understanding the full set of fraudulent behaviors, including when a cashier covers up a bar code by hand, or purposely misaligns the scanner and item so that the item isn't scanned.