The next big data thing for banks
- 01 March, 2013 10:42
Last year saw the biggest theft of credit card data in Australia’s history. While individual consumers were shielded in that case, the $30 million incident certainly caught the attention of Australian bankers. Fortunately, their newest and most powerful tool in the war against cybercrime is big -- very big.
Enterprise data is growing at an unprecedented pace – about 80% year-on-year - a big data explosion generated in no small part by the avalanche of transactions and sentiments expressed across traditional, mobile and social channels. There is another rich source. Vast unstructured and historical data accumulated with the enterprise over the years. This data has traditionally been ignored. So with the stakes high and the quantum of data even higher, it’s not surprising that IT analysts are receiving more big data enquiries from the banking sector than from any other.
As an example of how big data has become a critical weapon in the cyber wars, let me walk you through the situation of account takeover fraud (when a criminal impersonates a genuine card holder, gains control of the account and then makes unauthorised transactions).
Today, banks can predict and prevent account takeovers in real-time, based on the analysis of customer transaction patterns. Data analysts use an event-level risk scoring mechanism to identify potential fraudulent action or deviant behaviour by tracking input from multiple sources: customer call logs, existing customer documentation, transactions and more. Data from transactions such as change in PIN number and residential address, or requests for issue of a new card are treated as events and assigned risk-rating scores too.
These aggregated scores, along with metrics such as timing between transactions and existing red flags for customers then indicate the potential risk of customer identity takeover. When these risk ratings cross a predetermined threshold, they trigger an alert that automatically initiates appropriate fraud prevention measures.
Despite the rapid escalation of data streams in recent years - more than six million Australians bank through mobile channels and online sales now exceed $12 billion per year - bankers’ drive to extract meaning from vast data warehouses is not new. For decades, banks have invested in analysts and systems to glean insights from various data sources. Over a period of time, they have created multiple discrete “puddles” of data. The problem is, connecting these “puddles” takes too long, and drawing valuable insights from them comes too late. Traditionally, they have been restricted to structured data only. Hence, vastly available unstructured and historical data continues to be a dark secret.
Banks need big data innovation that straddles the entire data and analytics value chain to unlock valuable insights and allow rapid, actionable decisions. They need access to internal and external data sources, both structured and unstructured, enterprise systems and document management systems, customer call logs and social networks – all at once, with ease, in real time.
Ever experienced the frustration of talking to your bank about your credit card, and realising that they have no idea you also hold a home mortgage with them - much less that you’ve recently had a 20-tweet exchange with its customer care team on Twitter?
A 360 degree view of customers is the next big data thing for banks, covering all historical and real-time transactions and conversations, existing customer information and external data including discussions on social networks.
One challenge in achieving that 360 degree view is that only around 20 per cent of data in organisations is structured with predictable fields of information such as customer name and account number. A massive 80 per cent – in documents and contracts, as one example - is unstructured, making it more difficult to analyse and leaving it mostly ignored.
As bankers use the latest big data tools to shine a light into this “dark data,” the potential is staggering. I expect bankers be able to better anticipate our needs, and provide more useful, relevant services. And cybercriminals take note: the focus of the big data torch will illuminate the darkness.
Uber ride-sharing service hit by Brussels court ban
Last-minute lazybones dump Windows XP
Last-minute lazybones dump Windows XP
"Ivy League Editors offers high-quality online editing services, online proofreading services, which ..."Judge blocks sales of Typo keyboard on BlackBerry request
Judge blocks sales of Typo keyboard on BlackBerry request