24th May 2006 
 
Using simple data analysis techniques can help detect fraudsters that exist in Australian businesses said Peter Morris, KPMG Forensic, at the  
Chartered Accountants Business Forum in Melbourne.  
 
"The average fraudster is generally male, in their thirties and been employed at the company between four to seven years. The fraud often  
takes place over two years to the amount of $200K. In addition, the fraudster is less likely to take holidays for long periods so no one  
sits at their desk. Knowing these generalisations do not help detect a fraudster, but looking at trends over a period of time using data  
analysis can," Morris commented.  
 
Proactive forensic data analysis process allows companies to review large amounts of data. For example listing the top overtime earners  
amongst employees will highlight any abnormalities.  
 
Common fraud risks:  

• Dormant accounts: a review of data entries that were posted to G/L accounts not used on a regular basis
 

• Nights and weekends: a review of data entries that are created during non-business hours as many fraudsters do not like to be observed
 

• High dollar / round dollar: an analysis of high / round dollar journal entires
 

• Key words: metrics and analysis on the use of key words such as restate, reclass, reverse, and reserve
 

• Unusual account pairings, a list of data entries that have unusual account pairings such as liability to liability movements.