Guest Column | February 23, 2009

Combatting Employee Theft At The POS

Written by: Patrick Sobalvarro, President & CEO, IntelliVid

It's not widely known outside retailers, but employee theft comprises the highest percentage of retail shrink — more than that from traditional shoplifters or outside organized retail criminals.

But while retailers don't like to openly speak about the threat from within, they have been seeking ways to address the perplexing problem estimated to cost them billions annually in the U.S. alone. A 2006 National Retail Security Survey revealed that employees are responsible for about 48% of all shrinkage, and loss prevention experts say 60% - 80% of that internal theft happens at front-end registers. In other words, despite current surveillance efforts, about one-third or more of all inventory shrinkage takes place at the registers.

Such theft may involve cash, credit cards, gift cards or something called sweethearting. This last scheme occurs when employees give free or discounted merchandise to friends or family. They accomplish this through cancellations -- "as is" items, returned items, discounted items, void/no transactions or manual entries -- at the register. Sweethearting is particularly difficult to detect and easy to accomplish at POS terminals.

Retailers typically rely on transaction-monitoring and exception reporting systems to combat POS crime, using manual identification methods. In such cases, investigators need to know in advance exactly which transactions to look at, or review volumes of video manually to find the suspected transactions. In the weeks that elapse before staff can review tapes, employees continue to steal.

Beyond the time lag, using exception reporting alone typically takes too long to catch individuals — often a year — because of the huge number of false positives that must be checked and then discarded. Often, by the time a dishonest employee rises to the top of the report, he or she has a long history of having stolen merchandise.

In addition, most loss prevention departments are over-taxed, and with so many lines of goods to be checked and number of false positives, typically only the first few lines of reports can be reviewed in the time available. Meanwhile, shortage numbers continue to rise while thieves remain employed and continue stealing.

That's a real problem, since while the average retail employee is employed for three months at a time, research shows that internal thieves are employed much longer — some 11.7 months on average, according to research. Over that time, they may steal $25,000 or more from the retailer.

The good news is that in recent years, point-of-sale video analysis has been coupled with exception reporting to uncover fraudulent transactions instantly. Using video analytics, retailers are able to filter out the false positives to identify the people and merchandise involved.

Video analytics systems use sophisticated video processing algorithms to allow a computer to "understand" what's happening on the camera. The computer tracks both people and merchandise, and it can determine, for example, if a person is standing next to a point-of-sale or if merchandise is placed on a counter. The video analytics data is stored in a database so it can be reviewed later in conjunction with POS transactions and exception reports.

By synchronizing video analytics along with video recordings with POS transactions, loss prevention professionals can provide a visual verification along with the items appearing on the terminal transaction. This detects fraud such as sweethearting through a real-time alert and visual evidence that an item has bypassed the scanner.

In one recent case, a major luxury retailer used point-of-sale video analytics to address a high level of internal theft. The retailer found that using exception reports was extremely time consuming and it often took two weeks or longer to search and build evidence for a single case. There was no way to determine which incidents were more questionable than others — for example, when transactions took place without customers or merchandise being present — so security personnel had to sort through scores of false positives.

The retailer then implemented a video analytics system to flag suspicious behavior and correlate it with atypical transactions. It used the video analytics system to flag when questionable returns or void transactions took place at the terminal without customers being present. The retailer was then able to produce instant evidence for fraudulent events. In fact, it has been able to increase the number of internal investigations concluded in 2007 vs. 2006 by some 80%. As a result, the system paid for itself within a few months.

While exception reporting has been useful to loss prevention professionals for identifying certain kinds of internal theft, the fact that so much of internal theft still occurs at registers points to holes in the system, such as sweethearting, where, because there is no paper trail, little can be detected by data mining alone. It's thus imperative that retailers consider point-of-sale video analysis as an important tool in combating employee theft.

Patrick Sobalvarro is president and CEO of IntelliVid (, a video intelligence application provider.