Want to keep learning?

This content is taken from the Kogod School of Business at American University's online course, Business Analytics: The Data Explosion. Join the course to learn more.

Skip to 0 minutes and 1 second So we’ve seen that the Data Explosion has had a huge impact on a variety of industries. And now we’re going to focus on one – fraud and fraud detection. In my experience, I’ve seen fraud investigation techniques evolve, from being focused on finding issues in physical documents, to spotting fraud in the data. I’ve also seen a significant change in just the amount of information one can consider during an investigation. Now that records are electronic, an investigator can really look at anything. Every paper file can be turned into an electronic file and put into a document review tool to allow the investigator to really search and tag the evidence. So data analytics is a big part of fraud investigations now.

Skip to 0 minutes and 43 seconds And we’re going to learn more about it.

Skip to 0 minutes and 50 seconds When analysing data, it’s important to know the difference between anomalies and fraud. Anomalies are not intentional, and they are usually all over the data. Think about clerical errors, things like that. Fraud is intentional, and found in very few transactions. So searching for fraud during an investigation in the data, it’s like trying to find a needle in a haystack. Data analysis techniques are most productive when they’re proactively trying to find fraud. There are many techniques that one can use, but we’re going to focus on one – Benford’s Law.

Skip to 1 minute and 27 seconds Benford’s Law says that in any large, randomly produced set of natural numbers, a certain pattern will occur, which is what is illustrated here. Around 30% of the numbers will begin with the digit one, 18% with two, and so on, with the smallest percentage beginning with nine. The law is applied in analyzing the validity of statistics in financial records, and is commonly used to look at a large set of invoices and spot fictitious vendors. This technique is useful for fraud detection, because human-generated numbers usually do not match Benford’s Law. Perpetrators of fraud may not understand Benford’s Law, and incorrectly assume that the probability of each first digit is 1 in 9.

Skip to 2 minutes and 13 seconds Now, armed with this knowledge, a fraud examiner can detect fraud by looking for exceptions to Benford’s Law. So let’s talk about the advantages and disadvantages of Benford’s Law.

Skip to 2 minutes and 24 seconds Some of the advantages: it’s a very inexpensive method to implement and use. Since it’s applied to all data in the company’s database, when you’re doing an investigation, if your perpetrator is still working at the company, the perpetrator is less likely to know that you’re trying to detect their fraud.

Skip to 2 minutes and 42 seconds Some disadvantages: Benford’s Law only broadly identifies the possible existence of fraud. It fails to narrow the possibilities into a manageable field of promising leads. It does not identify the nature or the perpetrator of the fraud. So there are some limitations to data analysis. But it is a really useful technique in fraud investigations.

How can data help us to detect fraud?

We’ve seen that the digital transformation, and the Data Explosion that resulted from it, have had an impact on many areas of our lives.

Let’s now move on to examine how this huge increase in the amount of data available to us has impacted one area in particular: fraud detection. In this video, we discuss Benford’s Law, a data analysis technique which can be used to identify fraud.

Share this video:

This video is from the free online course:

Business Analytics: The Data Explosion

Kogod School of Business at American University

Get a taste of this course

Find out what this course is like by previewing some of the course steps before you join: