Advertisment

Banking and Analytics: Dices, Slices or Juliennes?

Which way are banks chopping the Big Data onion? That too, without getting lachrymose about the flip side of algorithms, automation and social analytics

author-image
Pratima Harigunani
New Update
ID

Pratima H

Advertisment

INDIA: Analytics has moved from information sinks to the stove of insights in just a few years. The way this ingredient is being stirred in pots of CRM, risk management, precision pricing, decision platforms at banks and financial institutions is particularly fascinating.

S Ganesh, MD & CEO - Dun & Bradstreet technologies and data services allows us a look at the modern bank’s chopping board and gives a taste of where analytics has been and is going when it comes handling the temperature of new format of databases, privacy issues, algorithms, market readiness etc. Let's watch his recipe of caramelizing predictive insights from raw data.

Tell us about the general scenario of adoption of analytics, specially the predictive flavor of it, in the industry.

Advertisment

Analytics is a hot trend, specially in banking. We use various inputs like demographics, roles, behavior patterns at one level and macro or micro economic data or Credit Bureau related behavior to churn out useful information. Predictive analytics is being used in a large way in banking to source customers or to anticipate their preponderance to buy. It is helping in areas like cross-selling, pricing strategies, better reporting, flexible pricing models, and better repayment formats etc.

For example, we specialise in providing Predictive Analytics, Decision Management and Information Management platforms and services to Dun & Bradstreet and its clients globally. Through them, it supports Decision Platforms for various leading banks and financial institutions worldwide.

Where?

We do Risk Management, which encompasses statistical modelling of credit risks in consumer, SME and corporate space. Globally MNC banks have been using predictive analytics for a long time now and while it has not been so widespread for small or mid-size players, the trend is growing across the strata now. So far, our solutions have been implemented across the region in various banks to enable them to do high precision risk based pricing especially of commercial loans to SME's and large corporations.

Advertisment

How much of this is about algorithms?

Predictive analytics is largely about algorithms but some users have specific needs around variables or concerns, like those about false positives. Algorithms and models are fundamental pieces but they alone are not adequate. The layer of every criterion’s individual weight and correlation aspects have to be looked into too.

When you use social analytics to strengthen insights about customers, why does it not impact privacy concerns?

Advertisment

Most companies can accomplish good profiling and offer solutions without threatening privacy of a prospect. It is all about the trade-off that can be achieved and the expectations set out. Today online footprints are not the same as they used to be some years back. Application world is like a walled garden in the Internet landscape but today few things are private. Regulatory bodies have to ensure that all this omnipresent data is not used for malafide purposes, but we should also accept the fact that the world has changed, whether we like it or not. So individual discretion matters as well.

Addressing banking as a specific vertical may call for some degree of customization, or not?

It is a combination of vanilla solutions and customization. Everyone needs a base model but over and above that every financial institution’s risk-taking ability and target audience differs and those differences kick in. Every bank has to calibrate its models as per specific factors. For instance, our Rating & Scoring models are customized specifically for markets in the South Asia, Middle East, and Africa and incorporate a number of local variables that are typically not present in the developed markets.

Advertisment

What role does Data per se play in making Analytics deliver what it promises?

Data is very important. There should be enough data and of the quality which makes predictive action possible. Wrong answers can happen easily so it is not just the volume of data but its richness and its contribution to amplifying accuracy that makes a lot of impact.

What about databases and their evolution from transactional formats to OLAP ones to clusters to streaming ones?

Advertisment

This direction is in fact, very relevant to what we do. Over the last 10 to 15 years, the time scale of data has been granular and now real-time data is taking over. That calls for products that match correspondingly to this new nature, example- for EMIs on the fly.

Technology can sometimes run faster than customer behavior but it has to align with customer patterns and market demands too. That is happening now. True usage of analytics is going to influence every facet of work as we move forward. It is going to affect areas like SCM, employment, preventive healthcare etc and move beyond narrow use cases.

Without disrupting or displacing the human element?

As we get better at predictive strengths, the human judgement comes up as stronger aspect on building up the model. All technologies are mere tools, whether it was the bullock cart or present day’s aero planes. It is a human who takes decisions around it. So we just have to evolve on how to use these tools better.

analytics cio-insights smac