In today’s increasingly data-driven world, companies utilize statistics collected on human behavior to direct their product development and marketing strategies. Entities are able to take the seemingly random actions of their customers and look at them on a collective basis to determine trends and patterns. Similarly, a general journal or disbursement ledger line item, when looked at in isolation, does not provide much information to an analyst during a fraud investigation. But when many line items are accumulated and analyzed as a whole, the data reveals patterns that are essential to an investigation.
The first step in performing analysis on a large set of data is to determine what information is necessary. For example, within a fraud investigation, when looking for improper expenses in relation to an entity’s disbursement policy, the data to review will often include the date of the transaction, the amount of the transaction, and the payee. In addition, characteristics such as the address of the payee and account charged can provide additional insight into the nature of the expenses.
Using the pivot table function of excel, one can then sort and summarize the data based on any of the factors to find hidden trends within the population. For example, using the fields given in the above example, the pivot table function would allow the user to find the total amount of expenses for any given vendor. The data can then be sorted from high to low by the amount of expense, to see where the most money has been spent during the time period. It would also be possible to sort by address and find out if any vendors have the same address. This could indicate that funds are being disbursed to inappropriate parties. By comparing results to preliminary expectations and guidelines using this tool, it becomes easier to identify abnormalities in a data set.
To take the analysis to a deeper level, one can also incorporate date data. This is useful in cases where an expense policy allows employee spending for business functions. By summarizing by both the day of the week and also the payee, events such as an employee making charges on a weekend can be easily identified and evaluated for appropriateness.
Overall, data analysis is useful in fraud engagements to identify patterns and trends in otherwise complex data sets. Analysts are able to leverage technology to provide a more robust analysis.
Schneider Downs can assist you with assessing your current fraud mitigation program or help you to design one. To learn more about how Schneider Downs can help you, please contact us or visit the Our Thoughts On blog.