KEY POINTS:
Businesses have a sophisticated new tool for increasing their bottom line. It's called forensic data analysis and it saves companies' money by combating fraud, theft, waste and abuse.
More than half the respondents had been the victim of fraud in the last KPMG Fraud Survey, and its incidence typically rises markedly during downturns.
Forensic data analysis involves examining organisational data collected in the normal course of business to identify patterns that match known fraud profiles or other anomalies.
Data sets include the general ledger, accounts payable, payroll, inventory, purchase orders, POS data and loyalty scheme records.
The patterns may be logical, such as vendors having the same mailing address as employees.
There may also be numerical and statistical patterns in corporate data, like duplications of specific digits, digit patterns and combinations, specific numbers, and round numbers.
The methodology even uses neural-net and other data mining technologies to develop models for fraud detection, prediction, and prevention where fraud patterns are lacking or obscure.
Mark Leishman, head of KPMG Forensic says: "To increase your bottom line, you need to understand what makes up the bottom line, and identify where improvements can be made. Businesses focused on sales in the boom, but now they're looking at why they're losing money."
Leishman says every business he has analysed has had data problems that open it up to fraud or poor management, with misconduct occurring in 50 per cent of those organisations.
Even banks can have issues _ Leishman has detected a manager overriding the loan rejection system and using outdated property valuations. The tool can be used to pick up mortgage fraud and insurance scams.
In retail, Leishman says fraudsters are getting smarter by using a variety of strategies, often alternated with each other or applied simultaneously, to defraud retailers.
These include: returns fraud; use of fake barcodes or incorrect codes to mis-state stock and sales; "sweet hearting" _ where friends and family use staff discount cards or return goods fraudulently; fraudulently obtained gift cards; and "wardrobing" _ where an item is "bought" at a lower price and refunded at full sales price.
Store to store comparisons can be revealing. Leishman analysed a store chain and identified an unusually high level of returns for a product trade in, which led to the identification of false returns and poor accountability in the trade-in controls.
Another of his recent analyses identified employees using their own personal rewards cards to obtain points when customers did not present a reward card at the time of purchase.
"These schemes, operated by external parties and employees separately or in collusion, ultimately cost the retailer and the consumer millions of dollars each year," he says.
KPMG Forensic analyst Ann-Marie Keller likens fraud detection to searching for a needle in a haystack _ with the tools and methodology in which she is expert being "like a magnet".
Of course not all anomalies detected in testing involve criminal behaviour.
System errors, data entry errors, missing data, unauthorised purchases, breaches of internal policy and general mismanagement can often be detected through forensic data analysis.
Identifying and correcting these problems can result in substantial savings, decreased future costs and streamlining of the existing procedures.
Forensic data analysis makes company audits more efficient and the results it delivers are cost effective, Leishman says.
It works in businesses of any size and the process typically takes two to three weeks.
