No Exceptions

HMV is proud of its commitment to being an innovative user of technology, including loss prevention solutions.

But when it realized that it was behind the curve in regard to monitoring and exception reporting, the London-based entertainment chain added a hosted system to improve detection of fraud occurring at point of sale.

“Over the last 15 years we were at the top 10 percent in terms of use of loss prevention technology to deter loss, including being within the top 10 percent for our electronic article surveillance usage, digital closed circuit television rollout and development of innovative benefit denial solutions,” says Colin Culleton, HMV’s head of risk and LP. The one area where it struggled was with its homegrown exception reporting solution, one that HMV felt lacked sophistication and intuitive knowledge.

This point only became clearer as the chain made the transition from paper gift certificates to electronic gift cards. While the switch was a competitive, strategic decision, it also opened a Pandora’s Box for dishonesty – among both associates and shoppers.

HMV’s homegrown exception reporting system managed to detect issues, but it was difficult to uncover these incidents in a timely manner.
Besides wanting a system with more data mining capabilities, HMV’s ideal solution needed to have both canned rules and the flexibility to add new ones “so nothing could fall outside of specific parameters,” Culleton says.

The chain began canvassing the marketplace almost three years ago, and chose a new system shortly thereafter. But after a proof of concept test, Culleton felt that the solution was “over-engineered,” so the retailer continued searching for more user-friendly software. Secure, a hosted solution from London-based Sysrepublic, fit the bill. The software is supported by sophisticated mapping technology that allows retailers to ask questions “in their own language” with the help of Explorer query language, says Nathan Smith, Sysrepublic’s co-founder and CEO-Americas.

“Users can actually write queries in sentences, versus programming code that needs to be detected,” he says. “It doesn’t require massive levels of support or training, and can easily be rolled out to chains with 10 or 20 stores or those with thousands of locations.”

The query language and integrated knowledge set is programmed to find suspicious transactions, then drill down through a data set related to established parameters. “We have created a system that does the detection work for analyst managers,” Smith says. “It also slashes the time it takes for loss prevention managers to get to the heart of the issue.”

HMV’s version of Secure is a hosted software-as-a-service (SaaS) solution that uses an open-source configuration. Sysrepublic opens the software up to its retail clients and integrates their rules into the system “to better eliminate noise and keep incidents from falling outside the norm of parameters,” Culleton says.

And being a hosted solution, HMV doesn’t have to worry about maintenance, enabling the IT department to focus on mission-critical projects.

Software pilot
HMV initially piloted the software last fall during a four-week proof of concept program that included 30 stores and one LP advisor, who pulled exception reports and daily alerts and compared them with previous reports.

“We saw a large number of exceptions, including refunds and a growing issue surrounding our electronic gift cards, and we noticed these immediately,” Culleton says.

The HMV LP team quickly set a tolerance level for transactional activities in key risk areas. The system applies these as rules, giving the chain the strict guidelines necessary to detect fraud, “rather than forcing our team to hunt down discrepancies blindly and head out on wild goose chases,” he says.

Positive results prompted the chain to roll out the solution chain-wide, and it went live just before Christmas 2008 — a peak time for potential loss due to the significant increase in transaction volume. The chain now has seven LP advisors (each responsible for approximately 45 stores) accessing the system regularly.

Besides achieving a return on investment in less than a year, every flagged instance has been investigated and solved.

While Culleton wouldn’t specifically quantify the results, he says the solution has helped HMV become more attuned to fraudulent credit card activity. “This includes instances that we typically wouldn’t have been alerted to until there was a chargeback request.”

Mapping trends
Currently, the chain is exploring how to expand the solution into its other channels, beginning with e-commerce, where www.hmv.com has seen an 18 percent increase in business over the last 12 months.

By applying Secure online, the chain can map trends by geographies and addresses, “keeping us alert to whether it is a postal issue or a problem that stems from individual customers,” Culleton says.

In addition, data mining is helping HMV wage a war against dishonest employees, and providing an opportunity to curb unintentional internal loss through employee training. The chain also has pinpointed when critical promotions are not being applied during checkout.