Article written by:
Maria Nobile
Senior, Forensic & Dispute Advisory Services

As a forensic accountant, a common problem clients bring to us sounds something like this: “I think there may be fraud, but I don’t even know where to begin looking.” Unfortunately, many businesses or individuals become overwhelmed by the amount of information that would need to be analyzed, the time and effort it would take, and of course, the cost.

Enter data analytics. It’s not that data analytics hasn’t been historically available to solve this problem, it’s that as a result of new technology the use of data analytics has become more accessible. In the past, an investigation was a manual, people-intensive and time-consuming process that required looking through each individual transaction. All of this can lead to a very expensive project that is comparable to looking for a needle in a haystack.

Now, data analytics can help automate these processes and help users identify problem areas and focus efforts on those areas. In turn, this saves time, manpower, and costs. With data analytics, software is able to sort through what could be an overwhelming amount of transactions and narrow the focus to transactions that are suspicious or potentially fraudulent.

Some examples of the tests we may run are:

  • Benford’s Law – Benford’s Law was used to develop tests to detect fraud, errors, estimates, and biases in financial data. This is the mathematical theory that in data sets, digits are distributed a specific way. For example, in a data set, the leading digit “1” appears about 30% of the time, whereas “9” only appears 5% of the time. Any transaction amounts that fall outside of the statistical norm will be labeled as suspicious, providing a starting point for which transactions are to be investigated further.
  • Relative Size Factor Analysis – This analysis is used to identify potentially fraudulent activities in invoice payment data and utilizes the largest and second-largest payment amounts to calculate a ratio based on purchases that are grouped by vendor. For example, one would expect that large payments to the same vendor would be similar in amount. If the two largest payments to the same vendor are vastly different, this could be worth investigating as fraudulent or an error.
  • Duplicate Payments Analysis – Fraudsters often use duplicate payment amounts to the same vendor. This analysis automatically identifies duplicate payments by vendor, allowing for easier identification of suspicious transactions.

Based on the results of such tests, we now have a more manageable and relevant data set with which to work.

We hope that you never have to worry about fraud within your organization. Unfortunately, this is not realistic. Regardless of whether or not fraud is detected, with data analytics, a fraud risk assessment can provide an insight into blind spots within a company’s procedures or financial reporting. To learn more about GBQ’s forensic accounting and fraud investigation services, including fraud hotline services, contact us today.

« Back