Applying data to reduce digital financial services fraud
Better data analysis is the key to stopping fraud before it happens
Fraud has become a significant problem for organisations across the continent, with a Myriad Connect study in Kenya last year finding that 61% of financial services CIOs cite fraud as one of the top 5 challenges facing their institutions. Over 1 in 5 Kenyans have been the victims of financial fraud including mobile money, fraud, online shopping fraud and SIM swap fraud, or know someone who has. Virtually all surveyed would consider moving to a financial service provider who offered more protection against fraud, as well as paying a fee for more secure financial services. Despite the relatively high number of respondents who have fallen victim to fraud, 66.8% said they felt their bank, mobile operator or mobile money provider did enough to protect them from these types of fraud.
For enterprises such as financial services, this changing environment characterised by massive volumes of data and growing fraud risks presents both challenges and opportunities. The challenges lie in securing and protecting massive volumes of in-transit and stored data relating to customers’ personal information and transactions. As well as challenges in identifying attempts at fraud, preventing fraud and theft, and prosecuting criminals in the event fraud takes place.
The opportunities, however, are to be found within the same proliferation of digital data, which presents significant opportunities. These include new approaches to big data analytics, behavioural analytics, machine learning applications and emerging technologies, to find the patterns underlying fraud and more effectively track down and prosecute those responsible for it.
Data analysis techniques have long been used for fraud detection, but the more organisations and consumers use digital services, the richer and more prevalent the data that becomes available for analysis. Now GPS delivers geographical information, interlinked applications can help identify culprits, and social network analysis can be harnessed to understand connections within syndicates and behaviour patterns before or after fraud is committed. Certain digital financial services, like online banking or mobile banking apps, simultaneously offer a wealth of data that can be collected and analysed to gather critical insights into where, how and by whom fraud is being carried out; and a proliferation of technology that can be introduced for fraud prevention.
In South Africa alone, the South African Fraud Prevention Service indicates that new fraud listings increased by 56% in 2017, but losses by companies that took proactive measures such as audits, a fraud hotline, fraud risk assessments and proactive data monitoring were 12% to 56% lower. Discovery Health has famously saved hundreds of millions of rands each year by applying analytics to its claims data in recent years. Last year alone, the company uncovered fraud and recovered misspent funds totalling over R568 million.
The Association for Certified Fraud Examiners (ACFE) notes that data analytics is especially useful when fraud is hidden in large data volumes, and can be applied across financial and accounting, customer, vendor, HR and external benchmarking data, as well as to internal communications and documents. Data on employee attendance, spending and access to physical sites and networks can also support fraud identification and raise red flags about potential internal fraud, says ACFE.
Given the importance of data analytics, 82% of CIOs in Myriad Connect’s survey use real-time insights and tracking tools and platforms, to monitor and manage fraud. Data analysis should form an essential part of security professionals’ toolkit that can help prevent and detect fraud – helping to mitigate risk, catch fraud faster and reduce the impact it can have on organisations.