Data Mining: This Stuff Really Works

2014-07
The section of the Association of Certified Fraud Examiners’ (the ACFE) recently released 2014 Report to the Nations on Occupational Fraud and Abuse I find most interesting focuses on the anti-fraud controls in place at the organizations victimized by fraud. While there is no “silver bullet” to ensure fraud will not occur, anti-fraud controls help mitigate and manage fraud risks. Making its debut on the list of anti-fraud controls in the 2014 Report is “proactive data monitoring/analysis.” First, as a data mining specialist, I am very excited to see this control on the list. While I believe it was always included as a component of many other controls, i.e. management review, internal audit, etc., categorizing proactive data monitoring/analysis individually is important. It not only illustrates the growing importance of data analytics for organizations around the world, but also highlights just how effective data analytics is for fraud prevention and detection.


The Report identifies proactive data monitoring/analysis as the most effective anti-fraud control in helping reduce fraud losses and fraud scheme duration. While the statistics below provide the numbers behind why it is most effective, it is important to note you can use proactive data monitoring/analysis to find multiple schemes in each of the main three categories of occupational fraud noted in the Report. That, in my opinion, is another reason it at the top of the list of effective anti-fraud controls. Now, let us get back to the Report and the statistical evidence. In reviewing the various anti-fraud controls charts in the report, we learn three key things:

iceberg-300x278

  1. Approximately 35% of victim organizations used proactive data monitoring/analysis as a control (Report page 31, Figure 26).
  2. Organizations using proactive data monitoring/analysis saw a near 60% reduction in median loss compared to those that did not (Report page 38, Figure 37).
  3. Organizations using proactive data monitoring/analysis experienced a 50% reduction in median duration of fraud scheme compared to those that did not (Report page 38, Figure 38).

One of the best ways to learn how to use data mining for fraud detection is to talk to others already doing so and reading articles, blog posts and case studies about how they got their program up and running. If you would like to learn more about the use of data mining for fraud detection and management of fraud risks, a few such reading materials, and a couple webinars, are below.

(Source: ACL Blog)

Tuesday, July 1, 2014 In: Hot Topics Comments (None)

Contact

Pricing

Demo