Tech News

Monday, 24 August 2015 | MYT 6:29 PM

SAS: Use data analytics to combat fraud

Speedy solution: Langley says businesses have to be able to react to fraud and respond right away, and this is made possible through the use of data analytics.

Speedy solution: Langley says businesses have to be able to react to fraud and respond right away, and this is made possible through the use of data analytics.

The fraud landscape today has changed and current preventive systems in place are struggling to cope with evolving technologies, according to Rohan Langley, fraud & money laundering detection specialist of SAS South Asia.

The figures he presented were staggering. Citing survey results by Certified Fraud Examiners (CFEs), Langley revealed that organisations lost about 5% of revenue, estimated at more than US$3.5tril (RM14.3tril) to fraud each year globally.

Langley was in town last week for the SAS Forum Kuala Lumpur 2015, which saw 500 business representatives and experts in data analytics gather for a one-day event that focused on current trends and developments in the data analytics industry.

The forum featured representatives from Google Malaysia, HP, Zalora as well as Tharkal One, all of whom shared how their organisations could benefit from the use of big data and analytics in their day-to-day businesses. 

In a press briefing, Langley shared that among the most commonly targeted sectors for fraud are banking and financial services, government and public administration, as well as manufacturing. 

And to him, this is precisely where data analytics solutions can come in handy.

With real-time monitoring, data analytics are able to pick up valuable information such as IP addresses, customer demographics, and typical behaviour patterns from multiple channels, all in one place.

Langley said that businesses need to be able to react to fraud and respond right away, and this is made possible through the use of data analytics solutions, as opposed to dealing with delays with traditional IT department investigations. 

Analytics can spot anomalies in behaviour at the transaction level for a large group of customers; as compared with current security measures such as two-factor authentication and anti-virus software which target one customer at a time.

Langley explained that before fraud can be detected, the solutions system would need to have data on normal or expected behaviour versus suspicious events. 

“This will vary depending on the type of institution deploying the software. Therefore, a company will need to define detection models and rules during the implementation, then continue to actively monitor behaviour after the system goes live,” he added.

As new types of fraud threats emerge, the institution may need to tune existing models or even establish new ones to handle these new risks.

Langley shared that most companies today have different systems in place and this has resulted in “siloing”, with no sharing of data between internal units.

What’s more is that fraud cases are also usually handled on an event or customer-to-customer basis, but Langley raises the question as on what will happen should all customers be hit.

He said that in most fraud cases, it’s usually too late for any measures to be taken, and worse still, the repercussions from media coverage far outweigh the cost of the fraud cases themselves. 

Langley believes that moving forward, data analytics will grow increasingly useful to help combat fraud faster and more effectively compared with traditional methods in place today. 

Three solutions already available today are the SAS Fraud Management and SAS Enterprise Financial Crimes Framework for Banking, which provide real time stop and block capabilities ideal for the financial industry, and SAS Fraud Network, which is suitable for any industry type requiring fraud detection capabilities. 

Another forum participant, SAS Asia Pacific chief technology officer Deepak Ramanathan shared that presentation of captured data has improved with the use of visualisation techniques such as graphs, charts and central dashboards which are able to show real-time results.

“We can’t discover much with just numerals, and the challenge has always been the end product – how can I get the discovery (data gathered) out to the field (in presentable form)?” said Deepak, adding that analytics don’t mean much, if nobody is able to understand them. 

He said that analytics could help businesses make scientific decisions based on proper data and move away from decisions made on gut instinct.


Powered by