Big data for commercial banks (2024)

Who would have realized that philosopher Immanuel Kant’s notion of global governance conceptualized in 1795 would bear an uncanny resemblance to the evolving modern day financial system? Kant had envisaged a peaceful federation of independent states, bound by consensus to a set of rules designed to prevent antagonism and conflict. In a similar vein, after the financial crisis, banks and financial institutions across the world have understood there is no short cut to glory. Instead a uniform transformative journey, from an account-centric to a customer-centric framework is a better path. Such a journey is only possible through analysis of what is commonly called Big data.

Big data is commonly a combination of structured and unstructured data. In addition to structured data available to the banks about customers (for example, account number, type, balance etc.) a large quantity of unstructured data originate from a variety of relevant sources. These sources include emails, call centre data, social media patterns, websites, customer feedback, agents and so on. This structured and unstructured data together combine into the large volume of data that become useful in important decision-making activity.

In today’s competitive world, financial institutions have realized that one must understand the customer’s behaviour across different segments through analytics, to serve the right product at the right time to the right customer. Until recently, banks were sitting idle on tonnes of such unstructured data and decisions were made in an ad-hoc fashion. However, today, banks are putting an enormous amount of investment to profit by combining the increasing volume and variety of data. In a recent study by Capgemini Consulting, it was found that worldwide more than 70% of banking executives find customer centricity important, but only 37% customers believe that banks understand and react to their needs and preferences in time.

Banking analytics is moving more towards analysing unstructured data and mapping it with structured data to get a holistic view of customers and build a real-time recommender system to predict their next moves. In the age of digital consumerism, financial institutions are taking a deep dive into incredibly rich big data. This can be utilized in many ways, for example: personalized offers made by banks. Personalization in the era of digital banking is the single most important thing to maximize their profitability. There has been a growing trend of commercial banks developing and using real-time recommender systems. For example, receiving a promotional SMS offering discount for buying movie tickets using the bank’s credit card; receiving a message on your mobile saying your coffee time is coming up, or that you could use accumulated points on your credit card. Would you not be pleasantly surprised if on a visit abroad, you receive an SMS from your bank informing you about the nearest ATMs? Even more surprising, if within a few hours of arriving you are contacted by the local branch of your bank asking for any help they can provide. Such is the power of analytics. To take advantage of and understand the next likely moves of customers, it is important that recommendations be sent to the right person at the right time.

However, perfecting these real-time recommendations is not easy and requires combined use of advanced statistical methods and machine learning algorithms. While distributed computing through Hadoop is becoming more mainstream for banks, it is the expertise in using some of the most advanced models, variational bayes methods, alternating direction method of multipliers, parallel matrix factorization, that will give banks an edge in effectively retaining existing customers and increasing revenue.

What seems to be the need of the hour is merging several strands of information across systems such as data from customer relationship management, portfolio, loan, debit, credit card etc., and mapping them on a seamless 360-degree view of customers. Customer analytics is the most powerful device for banks. Research by McKinsey shows that banks with advanced capability of using customer analytics have a four to six percentage point lead in market share over banks who do not. The immediate areas where banks can leverage the value of big data analytics and maximize value are customer retention, market share growth, discovering potential affluent customers, selling the next best product pricing of products and increasing lead generation potential among others.

The role of big data analytics in facilitating central bank policy decisions is another ongoing area of research. Interestingly, the Reserve Bank of India has been assiduously espousing that all monetary policy decisions are data driven (read: structured). May be one day, such policy decisions could also be driven by the unstructured part (read: gathering inflationary expectations from unstructured sources). Interestingly, much of development economics is based on randomized experiments and in a similar vein, inflationary expectations can also be generated on a daily basis.

To sum up, the promise of big data analytics in banks lies in data-keeping with a 360-degree view and making smart use of it.

Pulak Ghosh is professor of quantitative methods and information systems, IIM Bangalore. Soumya Kanti Ghosh is chief economic adviser, State Bank of India.

Comments are welcome at theirview@livemint.com

Follow Mint Opinion on Twitter at https://twitter.com/Mint_Opinion-

Unlock a world of Benefits! From insightful newsletters to real-time stock tracking, breaking news and a personalized newsfeed – it's all here, just a click away! Login Now!

Published: 04 May 2015, 05:31 PM IST