In age of digitisation, banks must bet on Big Data, Analytics:EY

A: We have typically seen four key challenges in banks in operationalising analytics

1) Banks typically look at analytics as an IT Function: Many banks have a technology-driven approach to analytics and run their analytics programmes as standalone IT function; this usually ends up limiting the analytics output to “informative” instead of “transformational”; and make the function an enhanced business intelligence unit instead of a business transformation enabling unit.2) Banks typically think they need perfect data to get started on analytics journey: Banks believe that not having a single repository of data, such as EDW (enterprise data warehouse) is indicative of data not being adequately organized for analytics and decision making. Thus, they miss out on the benefits of analytics that can be realized even in the absence of enterprise-wide data consolidation.3) Traditional mind-set of gut-feel based decision making: Most bank leaders prioritize experience-based judgments over data-driven insights, or if they use analytics are comfortable as long as data-driven insights help them validate the decision they have already taken. Analytics return on investment  (ROI) in both the above cases is nil. With the advancement in technology, we today have the option of optimising decision-making at customer level (i.e., n=1) instead of taking decisions at aggregate portfolio level or at 8-10 customer segment level. This needs a fundamental change in how banking leaders operate.4) Not enough emphasis on analytics implementation and associated business process changes: Banks have been laying primary emphasis on production aspects of analytics (technology, tools, data and advanced analytic skills) over the last couple of years. However, it is ultimately people in the business who need to act upon those insights in order to generate positive business outcomes. Thus the banks need to find the right balance in analytics solution development and analytics implementation across business processes.Q: How will analytics help banks to increase their revenues?

A: Analytics is no more a good-to-have, but a must-have [tool] for banks i.e., if a bank does not leverage analytics for decision-making, it is likely to be under increasing pressure on cost, margin and growth.

Let us understand this better through an example of how banks with low analytics sophistication can fall into an adverse selection trap. Compare two hypothetical banks - bank A has only basic business intelligence (BI)/reporting in place, while bank B has established advanced analytics centre of excellence and is using analytics-based decision-making techniques.

Given their limited view of a customer’s unique position and preferences, bank A will offer flat-rate pricing for all of its customers irrespective of their credit worthiness. Bank B, on the other hand, will leverage risk-based pricing models to offer better rates to their ‘good’ or ‘credit worthy’ customers, and less attractive rates to the rest of the product applicants. As a result, bank B would attract more credit-worthy customers, while bank A will only appeal to “riskier” customers who did not get preferential rates from bank B. This biased selection will result in increased NPAs, resulting in pressure on bank A’s profitability and long term business model viability.

Banks can increase revenues via deploying analytics frameworks to - improve customer cross-sell, enhance acquisition effectiveness, increase wealth management, penetration in high net worth customers etc to name a few analytics focus areas.Q: How will it enable banks to serve the customers better?

A: As customer interactions become more electronic and distant, banks need new insights into customer behaviour. By deploying advanced analytics techniques banks can predict with reasonable accuracy customer needs for a banking product/service at any point in time.This will help banks build customer preference related insight into self-service processes and restore the sense of personal relationship that human tellers once provided.Q: Will analytics lead to automation in the banking process? If so, how will banking jobs be affected by it?A: Advancements in robotics process automation (RPA) today enable banks to deploy robots i.e., functionally programmed virtual workers to execute rule-based information processes, improving accuracy and efficiency.

This will help significantly reduce costs for repetitive manual tasks and also help improve quality. By using cognitive analytics i.e., embedding machine learning algorithms on RPA platforms, we can further extend the scope of automation to help Robots learn and replicate decision-making by studying historical patterns of how humans took decisions in similar business cases, one example would be loan underwriting.

We don’t see this resulting in job losses, though new job creation for roles involving low-complexity repetitive tasks to be slightly impacted. We expect banks to use automation to improve customer service by reducing TAT and for effectively managing the volatility in business operations.Q: With the wave of digitisation in the country, what role do you see Big Data and modern data analytic tools play in helping India's SME space and the huge informal sector?

A: With the government’s push for ‘Make in India’, increased digital penetration with increased usage of smartphones and improved internet connectivity, thrust on financial inclusion with increasing penetration of financial services – we foresee a strong momentum building up to drive growth in Indian small and medium-sized enterprise (SME) space.

In such an environment, effective usage of Big Data and analytics tools can enable banks to enhance their credit underwriting processes by tapping into newer data sources including e-commerce transaction data, travel aggregator data, social media data etc, to help underwrite risks for SME business owners with low credit score or non-existent bureau scores. This will help banks cherry-pick credit worthy customers from a segment of customers that had historically never been given credit because of lack of sufficient information for banks to assess their risk.Q: What compliance issues arise for banks partnering with a third party vendor to use data analysis tools?

A: As long as customer personally identifiable information (name, address, contact details, unique identification ID, example- Aadhar) is not shared with the third party vendors and appropriate data security controls are in place while transferring data, we don’t foresee any challenges with banks leveraging support from third party to perform analytics on their data and provide the bank with actionable insights.Q: Given the lack of good primary research based data in India, won’t flawed analysis have a negative impact on decision making?

A: While banks might have started on the analytics journey recently, they’ve been focusing on data collection for the last two decades. Banks are sitting on a gold-mine of data in terms of customer demographics, customer transaction history, customer bureau data providing a perspective on external lending etc, which can help them enhance effectiveness of customer management, underwriting, collections and various processes.Q: How can analytics help banks and the economy in dealing with cash crunch due to demonetisation?

A: While analytics can’t help print money faster, use of optimisation algorithm can help optimise distribution of the scarce resource (new currency) optimally across cash dispensing locations (ATMs, Tellers in branches) across all banks such that it reaches the maximum population of India.

Here’s the detailed approach: Currently, the mandate for ATM refilling (which ATMs, what amount) is with each of the banks. This has resulted in a scenario where certain zip-codes have cash out across all ATMs (especially in Tier-2,3 cities), while certain others have multiple bank ATMs within walking distance of each other having cash.

As a short term measure (till the cash crunch is resolved), the RBI can take an “across-bank” view by using geo-spatial visualisation to identify regions where “across banks” ATMs have a cash-out situation.

Population density data can be overlaid on this to further help RBI provide directions for cash routing to certain cash starved locations to help optimise cash distribution such that the pain for a larger segment of the population is alleviated. This can be further augmented to not just optimise cash routing to ATMs, but to any cash dispensing location i.e., bank branches as well to be covered under similar geo-analytics.

Follow @shukla_05sid