In 2004, HDFC Bank in India started thinking about investing in analytics with an aim to revolutionize the banking sector in India. It took them a few years to build the framework and once they achieved that, the overall picture of a typical customer became clear to them.
They could then identify who was an ‘active’ customer, who was a ‘long-term’ customer, and, for that matter, who was just having an account for the sake of having it. The offers were sent out accordingly, so were the add-on benefits, and all these, in turn, helped them strengthen their framework and, more importantly, the trust the customers had in them. In June 2014, the NPA for HDFC bank stood at 1.1%, one of the lowest achieved in Indian banking. The investment in analytics and, subsequently, machine learning in finance , had reaped them rich benefits. The results were there to be seen by everyone.
Other banks have followed suit, and none could complain that investing and trusting in analytics have not worked out well for them. In fact, the whole BFSI domain has benefitted from the emergence of analytics applicable data science in finance and deep learning in finance.
Application of Data science in banking
Let’s look at some of the key areas in which analytics have intruded and made massive favourable changes:
- Customer Lifetime Value : Customer Lifetime Value (CLV) helps the banks assess the customers they have. It gives predicted values of all the business the bank will derive throughout the lifetime of a customer. Generalised linear models and regression trees are often used to build such models. It involves a rigorous process of data cleaning and manipulation. Then there’s often profiling and segmentation before the final model could be fit to generate the desired outcomes.
- Customer segmentation: Segmentation of customers based on varied criteria not only can help build models on the segmented data but can also in itself provide insights and trends regarding customer data, which often is used to target the right customers with the right plans and benefits. Methods like K-means clustering, decision trees and random forest can be used to get the much sought after insights.
- Customer Support: You must have got those offers via phone banking? And, perhaps, you have even opted for that life-term insurance plan? Well, nothing to worry about. The bank has built its robust customer support system and, with the help of data science in finance and data science in banking, has been able to track and explore all your data and your investment patterns and cycles. Hence, it knows what plans you have and what loans you don’t have. It can hence come up with the offers that suit you. Automation by means of macros and arrays have helped it to identify different plans and offers for each of its customers.
- Fraud Detection: The banks and such financial institutions are wary about fraudulent activities that can affect them adversely by making them incur huge loses. The usual data science in finance,data science in banking and machine learning in finance tools are used to classify data, segment data and predict or forecast for the future. Fraud Analytics has grown out to become a separate branch in itself and often involves rigorous regression model building and forecasting techniques.
- Risk Modelling: Banks and other financial institutions have deemed risk modelling a mandatory and integral part of assessing their overall performance. Stress testing has now been done in most major banks around the world and tools like SAS, R and Python are often used to carry out the check on the financial health of the bank. It also attempts to measure the chances of unfavourable and unforeseen events affecting the bank and the likely remedies if such a situation indeed arrives.
There are other functions that analytics successfully does for the bank. But, overall, the players in the BFSI domain must successfully integrate data science in finance and deep learning in finance in their decision-making process. The important insights generated must lead to the building of implementable strategies. Every passing day enhances and, perhaps, emphasizes the importance of data science in finance and data science in the banking sector.