Ice Cream

Case studies

When a retail company suspected fraudulent activity

Deloitte Analytics uncovered it in places they'd never thought to look

The client, a retail company, was experiencing fraudulent activity resulting in some of their 100 branches reporting poor returns. The company had knowledge of one type of fraud, but recognized there may be other types of fraud that they didn’t know about. They wanted to see how analytics could help them see where fraud was potentially occurring, ways it could have been carried out, and which areas they needed to monitor more closely.

The Deloitte Analytics team worked alongside the Forensic team, using the data provided by the client, to identify types of people who were more likely to carry out fraud based on a number of characteristics.

The findings astonished the client - at least four different types of fraud were found, and the team were able to tell the client which types of staff members, and which branches, were most likely to be engaging in fraudulent activity on a regular basis.


The client had an internal audit function but could not cover all branches at once. They assumed that finding perpetrators of any fraud would be like ‘finding a needle in a haystack’, but the client assured that the analytics team had a number of tools and techniques that would reveal any anomalies in data. The client wanted to get smart about using analytics to undertake a risk assessment in order to figure out where to place their focus.

The company found that their employees were able to manipulate stock-taking processes and procedures in their systems. This, combined with the ability to provide a discount or refund to the customer and earn commissions on sales, meant that employees had the opportunity to commit a myriad of different types of fraud. These potential frauds were not detectable because of control weaknesses in their internal systems.

The challenge was to try and distinguish the different types of potential fraud and figure out how to analyze the data in order to show where any fraud was taking place, how frequently, by whom, and the costs associated. The techniques and results needed to be robust so that innocent people were not labelled as fraudsters.

The challenge

The team used a combination of the standard DTect™ facility and statistical techniques to correlate certain types of behaviors displayed by what was understood to be a fraudulent activity. For example, certain types of products which are easily sold, returned, and refunded were correlated. The types of products, as well as stock-take variances and stock-take adjustments taking place in branches were profiled.

These tell-tale signs are some of the signatures of fraud, and were used by the team to develop a lens that helped the client see the potential fraud clearly. The next step was to look at who did what by integrating this information with audit logs to compare the levels of refunds with the levels of stock write-offs, and the stature of the people processing all these transactions.

The Deloitte Forensic team provided the genetic code of a fraudster and a description of what types of things to look for. This, coupled with analytics, made it easy to see anomalies in the data and pinpoint the area in which there were issues.

How analytics helped

All the information was collated and the team were able to pinpoint to the client a number of branches and staff connected to anomalous transactions. The client’s normal internal audit process was unable to pick up on these variances. The accuracy of the analysis and prediction was high. Previously unnoticed fraudulent activity was discovered – exactly where the analytics had predicted.

The fraudulent activity meant the client was vulnerable in the region of hundreds of thousands of dollars – a substantial amount as this client was in the business of selling an inexpensive product.

Since carrying out the exercise, the client requested an ongoing agreement to continue with the analysis. The client’s internal audit team will be able to feed the Deloitte Analytics team with fresh data on a regular basis so that they can then provide them with an updated visual output that will illustrate, based on a range of indicators, a particular branch, or a particular area of the business where they will need to focus.

The solution

Did you find this useful?