(Page 2 of 3)
Chan began entering refund data into a spreadsheet to see if he could identify patterns. Once he had a large quantity of data input, he found that he didn’t have the resources to support it or to distribute it to his retail customers. Eventually, he developed the web-enabled FraudWatch system, which he sold to Triversity a year ago. Building the software in web-based protocols and delivering the product over the web so retailers do not have to install special networking systems or software are key to making data-intensive security systems work, Chan says. “This product would not exist without the web,” he says. “The web is the medium that made this affordable.”
In capturing the toner-cartridge margin thief at Staples, the web allowed local police, the private investigator, the Staples loss prevention department, store management and Triversity all to coordinate data quickly and efficiently, Chan says. “There was an enormous amount of data required and having a web-based system meant they didn’t have to ship paper,” he says. Further, he adds, the local police were able to log onto to the FraudWatch system and make their own judgment about the quality of the evidence before charging the thief. “Before the web, a chain would have had to build its own private network to be able to move data like this around and that was just not affordable,” he says.
The margin thief ended up paying Staples a percentage of what he had taken as well as the cost of the investigation. In addition, Switzer says the thief identified a few other criminals who were operating the same scam.
Apart from being an affordable solution to the question of data networking, having a web-enabled system makes the process faster and allows retailers to analyze much more data. “Everything we do is 100% web-deployed. It takes us three minutes to retrieve and process the transaction log of everything that happens in a supermarket in a day,” says Larry Miller, president of Trax Software and Consulting Co.
Into the voids
Trax licenses its data retrieval and analysis software to retailers, although it can provide it on an ASP basis if the retailer prefers. Trax’s Shrink Trax product gathers data from point of sale devices, deliveries to the store from vendors, in-store camera monitoring systems and data devices within different departments, such as meat, pharmacy or photo. Shrink Trax applies data mining techniques to the data it gathers, looking for patterns and relationships, in addition to looking for particular relationships that the retailer sets.
From the data, Shrink Trax creates reports for each store, prioritizes the results, recommends actions that store managers should take and delivers it all over the web. Part of the value of having the system web-based is the speed with which retailers can act, says Shawn McKimens, manager of shrink control at Norfolk, Va.-based Farm Fresh Supermarkets, a 36-store division of Supervalu Inc. “With Shrink Trax, we can notify the store immediately if there’s a problem,” McKimens says. “Shrink Trax gives it a sense of urgency.”
Farm Fresh has used Shrink Trax primarily to identify cashiers who were not following procedures. For instance, it helped identify cashiers who were incorrectly-either intentionally or not- selling all 12-packs of Coca Cola at a sale price when the promotion was limited to two 12-packs at the sale pirce per customer. During the promotion, after the cash register rang two 12-packs at $1.98, it would automatically ring all subsequent 12-packs at $3.98. Often a customer would complain, pointing to the $1.98 promotional price. Cashiers would then void the $3.98 and key it by hand at $1.98. “We were able to stop it within a day,” McKimens says. “Before, we would have caught it if someone had noticed the behavior going on. But even if we had been able to analyze all the numbers, we still wouldn’t have known what was happening until the end of the reporting period, and that might have been too late.”
Immediately identifying a loss problem is only one major benefit of using a web-enabled loss prevention system, retailers say. Another is that new users can quickly learn how to use such a system. Holmes estimates the time to learn the FraudWatch system is less than three hours. “There’s no training involved,” he says. “It’s all point and click. You’re only five clicks away from looking at any transaction in the entire company. And our only additional cost was setting up 22 Internet accounts for our district managers.”
Identifying a problem with a particular item is only a small portion of the analysis that most web-based loss prevention systems provide. Such software usually analyzes daily transaction logs from stores and looks at several dozen data elements. Among them: number of no-sales per cashier, number of customers per no-sale, sales per customer, items per customer, number of voids, voids as a percent of sales, number of store coupons and their percent of sales, same for vendor coupons and double coupons, number of refunds, number of refunds lacking customer phone numbers, number of items not scanned, sales per hour, ring time per order, tender time and more. “We could never have dreamed of going through every transaction like that before,” Holmes says.
Training on track
With hard evidence from such a range of transactions in hand, retailers then can take action. Whether the problem is fraud or simply not knowing the rules, retail managers can use the data as an opportunity to re-train cashiers. “We’ve identified a lot of training issues, people who just didn’t know what the rules were,” Holmes says. Bata averages 8,000 employees throughout the year. Many are younger, entry-level retail employees. “We then are able to put together a plan of action to address the issues,” he says.
Not only does the coaching help the cashier whose behavior the system flags, but it also communicates to all staff that their behavior could be monitored as well. “Of your cashiers, 10% would never take a thing, 10% will always take something and 80% will go either way, depending on circumstances,” says Holmes. “The store staff knows we’re monitoring their transactions. We’re saying to that 80%, ‘Don’t bother.’”