New SAS procurement fraud solution detects and prevents suspicious behaviour through data analysis

The solution red-flags irregularities within data, providing a return on investment in just three weeks

Procurement fraud is the second most prevalent form of economic crime in South Africa, with 59% of companies surveyed by PricewaterhouseCoopers (PwC) being affected over the past 24 months. So said Louis Strydom, a chartered accountant at PwC, who was unpacking the findings of the Global Economic Crime Survey at yesterday's launch event of SAS' new procurement fraud solution.

Even more alarming was the finding that 77% of internal fraud in South Africa is committed by senior management, and that 59% of local companies have been affected by procurement fraud in the past two years - the most prevalent forms being vendor selection (76%), payment processes (60%), bid processes (57%) and vendor contracting (52%). Fraud risk management is the most effective method of detecting fraud - although this only uncovered 17% of all cases - while 80% of fraud detections are made by methods within management’s control.

Data analysis can improve these numbers, said Yolande Byrd, Director at FACTS Consulting, which worked alongside SAS to develop the solution. For Byrd, there is a general lack of understanding within companies when it comes to fraud detection, as they are unsure of what to look for, and how. Further, there is a lack of internal skills to manage the fraud detection process, which is often left up to the IT manager.

The SAS procurement fraud solution performs constant background analysis of a company’s systems – aggregating data from all sources - and flags suspicious behaviour or patterns, such as relationship links, duplication of information, chronological order of transactions, irregular transactions and ghost vendors. It gives organisations the ability to proactively monitor and detect potential fraudulent activities and also serves as a preventative measure. For Byrd, organisations can no longer afford to be too trusting of employees and suppliers – making them aware that their activities are being monitored will go a long way to reducing instances of fraud.

The analytics underpinning the solution comprise the following elements:

  • Business rules: A coded understanding of the way that an organisation does business and what fraud looks like so that the solution can detect rule breakers;
  • Anomaly detection: Detecting transactions or interactions that are outside of the organisational norm;
  • Predictive analytics: Using the results of previous analytic processes to forecast fraud trends;
  • Text analytics: Analysing unstructured text-based data, such as keywords within emails;
  • Database searches: The incorporation of databases of known fraudsters or other specifics to be red-flagged; and
  • Social networking: Highlighting the relationships between individuals and entities based on common links such as bank account numbers, telephone numbers, addresses or company directors.

The solution has a very high return on investment, says Dion Harvey, Sales Director at SAS, with most companies able to recoup their expenses in fewer than 12 months after implementation. A pilot project at a client that deals with many suppliers produced results in just three weeks, enabling the company to repay part of the cost of the solution with the money recovered from fraudulent activities.

Once the solution is implemented – which can take up to six weeks – it looks into the client's systems and continually monitors transactions for fraud, explains Byrd. A number of tests are run against the data that look for any irregularities in transactions. The solution is also able to detect fruitless and wasteful expenditure, such as duplicated payments. It then brings all data together for each supplier, department and employee and flags any loopholes according to a risk score. This allows companies to immediately see where the biggest fraud areas are. If something needs to be investigated further, it can be sent on to a case management system, or to an accounting or investigation department.

While the solution was built within a fraud framework, businesses can add different modules and apply it to other areas of the business, for example, subscriber fraud within a telecommunications company, or payroll or financial statement fraud.

While costs are difficult to determine, as they vary across organisations, Byrd explains that, if a company uncovers just 10% of fraud, that recovered money will likely more than pay for the solution. The solution is Web-based and interfaces with existing systems, such as case management systems. Plug-ins ensure it is compatible with existing systems, such as Oracle.

There is no such thing as small fraud, concludes Byrd. Small fraud is just big fraud waiting to happen, as people become greedy. Perhaps the biggest motivator to uncovering fraud is that a 3% reduction in fraud equates to a 60% increase in profits.

The solution is available immediately through SAS.


About SAS

SAS is the leader in analytics. Through innovative analytics, business intelligence and data management software and services, SAS helps customers at more than 80,000 sites make better decisions faster. Since 1976, SAS has been giving customers around the world THE POWER TO KNOW®.


Editorial contacts:

Back to Top