The browser is becoming the entry point to retailers’ shrink data and allowing faster reactions to theft.
One of the biggest issues that confronts chain retailers is employee honesty. Most employers realize that no matter how closely they screen prospects, they will end up hiring a few dishonest employees; the odds favor that outcome in a pool of several thousand people. And so the challenge becomes how to spot those who are dishonest, alert the opportunistic thieves who are basically honest to the fact that there’s a chance of getting caught, and not implement procedures that will disrupt the company’s primary job of selling merchandise. On top of all that, retailers are cautious about not overloading their IT infrastructure to catch a few thieves.
To address those concerns, some retailers are adopting web-based tools that can move a lot of data inexpensively to an analysis program and make results of the analysis available to loss prevention managers, store managers and others at any time, any place, at a low cost. “The web is our access point to the database, all we need to get at all the information we’ve gathered in there is a browser interface,” says Peter Gerhardt, vice president of finance of Toronto-based Town Shoes Ltd., which operates 64 stores with 1,200 employees across Canada and uses the FraudWatch technology from Triveristy Inc.
A buried paper trail
In spite of advances over the past few years in technology that makes it easier to catch thieves, employee theft continues to be a major problem at retail companies. The latest National Retail Security Survey from University of Florida researchers reports that for the second year in a row, employee thefts hit a record level. Preliminary results show that shrink-the term for unaccountable losses of all kinds-reached $31.3 billion in 2001. The employee portion of that accounted for $15 billion-48% of all losses. The next highest category was shoplifting, which cost retailers $10 billion and accounted for 32% of losses. Following were losses due to paperwork errors, 15%, and theft by vendors, at about 5%.
With employee dishonesty accounting for the biggest portion, it makes sense to tackle that aspect first, analysts say. And with credit card and debit card fraud being the most likely area where employees can steal a lot, it makes most sense to start there. That’s where Town Shoes began after it installed a FraudWatch web-based system from Toronto-based Triversity. “We found-not unexpectedly, but disappointingly-that our employees were stealing from us,” Gerhardt says.
The most common type of theft the company uncovered was the reversing of transactions and crediting the sale amounts to an employee’s own credit card. It’s not a terribly sophisticated type of fraud-in fact, it leaves a most incriminating paper trail-but employees were getting away with it because the volume of transactions kept loss prevention managers from conducting in-depth analyses of transactions to try to spot fraud. Each of Town Shoes’ 64 stores processes 200 transactions a day for a total of nearly 13,000 each day.
Working with Triversity, Town Shoes loaded six months of transaction log files into FraudWatch. It then sorted refund transactions by card number and dollar amount. “We dug into the database to see if we had supporting sales transactions,” Gerhardt says. “If you find unmatched sales, you know you have a problem.”
A policy breach
Town Shoes found what it suspected it would find: On the first pass it identified four people who were giving themselves refunds. “They figured they could take a bona fide sale and give a refund to their own card and they’d never get caught,” Gerhardt says. “And they were probably right. There is no way we would have caught people if we had had to move and process all the data manually.”
After the success with credit card and debit card fraud, Town Shoes turned to what’s known as cash post voids, in which a clerk simply voids a transaction rather than issues a refund, then pockets the cash. Again, it spotted dishonest employees pretty quickly, including one who had taken $5,000. It has identified as many as five other employees who were stealing via cash post voids. “If you have one person in each store doing $5,000 in cash post voids, that quickly becomes a real number,” Gerhardt says.
Town Shoes identified cash post voids as a problem, but the FraudWatch system was not written in such a way as to isolate cash transactions, says Martin Seaton, Town Shoes controller. Rather, it reported all voids-cash, credit or debit. To get to the cash voids, loss prevention staffers sorted on voids, then copied all voids and exported the data to a Microsoft Excel spreadsheet, then sorted the data further to identify cash voids. From there, they were able to further sort the data by date and store, which uncovered the patterns they were seeking.
Cash post voids should be caught by managers who are following company policy and reviewing all such transactions daily, Gerhardt says. But the investigation of cash post voids uncovered another problem that was easily solved as a result of having the data: Store managers were not reviewing all voids. “This gave us something we could use in pointing out to managers that they were not complying with company policies,” Gerhardt says.
Any time, any place
While such information could be collected and disseminated without the Internet, the web makes the job easier and more timely, Gerhardt says. For one thing, the web allows Town Shoes investigators to know exactly what is happening while it’s happening. And the web gives them access to the information on a real-time basis in such a way that store personnel may not know what’s happening. “This can be done from any laptop,” Gerhardt says. “So if someone is in a store and needs access to records, they can pull it down from their own laptop.”
Town Shoes has not computed its payback yet on the Fraud Watch technology. “The payback will be significant in terms of deterrence,” Gerhardt says. Town Shoes pays $75 a month per store-$4,800 a month total. At nearly 400,000 transactions a month, the fee is equal to about 1.2 cents per transaction. “I assume that for every one we catch, we’ll deter another four or five,” Gerhardt says.