Big Data in Banking

Also, it would allow banking institutions to increase the number of customers and provide better customized services. The rise of websites that stand between customers and banks has the potential to eliminate the latter unless banks reinvent themselves in the Big Data context. Several challenges as scalability, security, and privacy need to be resolved before taking full advantage of Big Data.

Big Data is a term currently used to denote the huge amount of digital data (terabytes, petabytes, or even exabytes) available to companies that makes impossible its handling by traditional data management technologies. As early as 2001, the Gartner analyst Douglas Laney associated Big Data with its defining three properties, known as the three Vs: Volume, i.e., growing amount of data, Velocity, i.e., high speed of incoming data, and Variety, i.e., many formats of the available data. According to IBM, humans produce 2.5 exabytes everyday and 90% of the data has been created in the last two years alone. This huge amount of data needs new storing, transmission, and processing capacities in order to be useful for various economy sectors as for example the banking industry.

Last year, Citigroup hired Watson, the famous IBM computer that has previously beaten the champion human contestants on “Jeopardy”, an American quiz show. The reasoning capabilities of Watson can help Citigroup perform risk analysis, discover fraud, or provide financial offerings. In another example, PayPal, the company that dominates online payments, barely survived in 2011, when it came under attack by fraudsters. In order to deal with these threats, PayPal developed Igor, a computer system named after a Russian thief and hacker, to analyse unusual patterns in the data. Several other companies like Palantir and Xoom followed by building systems that aggregate data from various places and try to find connections.

The largest users of fraud-fighting computers are credit cards associations as Visa and MasterCard. While individual payments might look legitimate, by analyzing all the available transactions, computers are able to spot unusual patterns or connections. None of these systems is cheap, but they are a lot cheaper than possible fraud losses, e.g., for credit card companies the average fraud losses are 0.1% of all card transactions. Small banks that could not afford fraud detection systems had to close or sell their fraud detection systems to large credit card companies. These large companies can look at far more transactions than smaller companies and thus are better in spotting fraud. Nevertheless, small companies realizing that they can be deprived from a rich source of data, i.e., customer’s spending patterns, are currently trying to adapt general data processing software and storage capabilities in order to keep the credit card business to themselves.

The availability of large amounts of data has allowed new business opportunities to emerge. For example by analyzing customer shopping behavior, Citigroup is able to send text messages with discounts in stores and restaurants that can give the bank a second transaction and a cut of the extra spending. In another example Lloyds Banking Group is thinking to show to the customers next to a balance, also how much money is left after paying their bills. ZestCash is able to lend cash to customers with poor credit histories, by looking at thousands indicators, and Tesco offers now credit cards and loans, assessing the credit worthiness of customers based on their shopping habits. Financial advisors as Mint have an overview of customer’s financial situation in several places and provide advices on how to cut their debts.

All the previous advantages of employing Big Data in the banking industry come with some risks. With enough information about one another, investors and savers might no longer need the banks. Andrew Haldane from Bank of England notices that after music and publishing, finance could follow in cutting the middlemen. Privacy and security pose also complications as customers are afraid to lose their power over their data in the data aggregation process. Target, an American retailer, was criticized in 2012 for sending baby-related coupons to a teen-ager after discovering from shopping patterns that she was pregnant, before she announced her family. Big Data poses also scalability challenges not only because of the huge amount of data that we are dealing with but also because this data is continuously growing. These Big Data challenges need to be properly researched by both academics and industry before one can reap the full benefits of Big Data.