Covering Disruptive Technology Powering Business in The Digital Age

Albert Einstein told us that “time is relative”. This is especially true when it comes to data analytics.

Imagine that I gave you a list of 45,000 credit card transactions, 44,999 of which are fine, one of which is fraudulent, and you need to find the fraudulent one. How long will it take to analyse that data and find the fraudulent record? Maybe you could get a team of people and work your way through the list, trying to get typical spending behaviour data from some of the card owners’ transactions. How long would this take, perhaps 2 days? 20 days?

Did I mention: the 45,000 transactions happen every second, and you need to identify the one bad one in real-time, blocking the transaction before it completes?  Adding the factor of time to the equation completely changes the scope.

This is a real life problem. In 2018, it was estimated that 45,000 Visa transactions happened every second globally. Add in Mastercard, Amex, debit cards and now the plethora of e-wallets, the question of how banks and businesses detect fraudulent transactions becomes immensely challenging.

The vastness of this lake of electronic transaction data, combined with the rate at which new transactions pour in, provides possible anonymity for those looking to cheat the system.

The only way to combat this is with a modern-day data platform. Big Data has evolved rapidly; the days of analysing enormous sets of static, unstructured data with Hadoop have been replaced with a range of complementary technologies that enable faster real-time analysis of massive streaming data sets. It is this modernised strategy that allows the banking sector to stay ahead of cyber fraud.

Cloudera is the leader in utilising open-source Apache project technology to create real-world data analytic solutions that raise the bar on how large amounts of data can be ingested and analysed in real-time.

For Cloudera, it is not just about real-time streaming analytics. It is also about enabling a holistic approach to data by providing an enterprise-wide data platform for clients. This allows them to bring more to the party when it comes to fraud detection. In addition to the speed and scale, they can use this cloud-based data platform to break down data silos across organisations. In the case of Bank Danamon, this meant customer behaviour information could be combined with transaction details to identify possible fraud, based on historical and expected customer spending patterns.

It doesn’t end there. Data is the fuel of learning. In the case of Cloudera, this means machine learning. With Bank Danamon, machine learning was built into the total solution which enabled the bank to start predicting potential fraud risks before they impact customers. The bank was able to develop preventative triggers and proactively alert customers to potential fraud.

We may not quite be at a stage where fraudulent transactions are found every time, the moment they occur, but we are moving towards that capability. Bank Danamon is using Cloudera Data Platform to ingest and analyse 1TB of new transaction data each day. This has allowed them to identify potential fraud significantly faster than in the past, where it was not uncommon for fraud to be discovered only when customers received and reviewed their statements.

In the data-driven world, a single second can be the difference between success and failure, lost business or new customers, catching fraudsters or letting them slip through your fingers. Technology exists to capture every “data second” and Cloudera is helping many companies to build platforms to capture and use that data as it gets created. Find out more here.

(0)(0)