Finding fraud at 1-800-Got-Junk was arduous in the days of paper trails, Paul Speed says. ‘You’d have to do a lot of running around. It would take hours to pull out the data.’Rafal Gerszak
As much as every company hopes it's hiring trustworthy employees, there will always be a few crooks who get on the payroll. At Vancouver-based 1-800-Got-Junk, stopping these ne'er-do-wells before they do damage falls on Paul Speed's shoulders.
For this junk removal business, it's financial fraud that poses the greatest danger. Money is frequently moving between customers, truck drivers and corporate staff, and cash can go missing if people aren't paying attention.
Years ago this used to be a big challenge, but thanks to advances in technology and data analytics, uncovering nefarious activity can be done in an instant.
Before 2008, when 1-800-Got-Junk improved its systems, tracking fraud was a painstaking process. Truck drivers would write down on paper what they collected, how many garbage dumps were made, the number of clients served in a day and more. That information was then delivered and entered manually into a computer program. To make matters worse, every 1-800-Got-Junk franchise had its own individual system.
If someone had a feeling that fraud was taking place, he'd have to search through the data and somehow figure out that money really was missing. "You'd have to do a lot of running around," says Mr. Speed, manager of franchise operations services. "It would take hours to pull out the data."
Five years ago, the business beefed up its Junk Net booking system to handle and process large amounts of data. While information still has to be manually input, the program can sort those numbers instantly and in ways the previous system couldn't.
Junk Net wasn't specifically designed to catch fraud – the system tracks 50 key metrics including sales calls, the number of repeat customers and sales conversions – but when Mr. Speed looks at certain data points together, he can tell whether something fishy is going on.
For instance, Junk Net can quickly calculate the average job size, average job transaction cost and average dumping costs over a 30-day period. If the job size and transaction costs fall but the dumping costs remain the same, then there's a chance that money is going missing.
"If the same amount is getting disposed of, but less money is coming in, then that's a good indication of theft," Mr. Speed says.
Unlike years ago, Mr. Speed says he can get this data "with a click of a button" and quickly stop the fraud before it gets worse.
As computer processing has become faster and programs have been able to hold and calculate more data, many companies have been able to significantly reduce fraud-related losses, says David Meadows, an associate partner on Ernst & Young's fraud investigation and dispute services team.
He says that traditionally most companies would look at a sampling of data to determine whether fraud was occurring. Now every single record produced by a company can be tracked. "Technology now allows a company to look at 100 per cent of the data rather than a small sample," he says.
Multinationals can not only look at company-wide financials to see whether there are any discrepancies, but they can use programs that allow them to examine e-mails for unusual keywords, too. Some words that could indicate expense fraud include friend fee, one-time payment, processing fee and donations.
If an e-mail includes one of these words, the software will then look at other data to see if it makes sense that the staffer is using these words or if a red flag needs to be raised. Ernst & Young's software tracks more than 3,000 words and new ones are added quarterly, Mr. Meadows says.
These programs can also determine whether sensitive files are being sent outside the company, if people are spending enough time on a job site and much more.
Using data analytics software to catch fraud helps a company in two main ways, Mr. Meadows adds. The first is that it mitigates a significant amount of risk – the less chance of losing sensitive information, the better – and it also can save a company big bucks.
Reports have found that large companies lose about $140,000 a year to fraud. "And those are only the ones being detected," Mr. Meadows says.
While Mr. Speed can't say exactly how many incidences of fraud occur now compared with before the software started crunching numbers, he has no doubt that his technology has helped. He also points out that if people know that he's watching, they'll be less likely to steal.
No matter the size of the company, though, less fraud means a better bottom line. "We can make sure more of the money is being reported and getting where it's supposed to be getting," he says. "That keeps the franchises and corporate headquarters healthy."