Analytic Challenges in the Banking Industry

Information Management Special Reports, November 2004

Soumendra Mohanty

In recent years, the banking industry has evolved dramatically, driven by changes in the business and economic environment, in legislation, in competitive pressures and in enabling technologies. With a wide variety of products, such as credit cards, mortgages, home equities, lines of credit, savings and checking accounts, insurance and investment products, banks need to anticipate and satisfy the changing needs of their customers. In addition, financial institutions need to be able to estimate and review risk and compliance with regulations such as Basel II and mandatory capital requirements. More than ever, banks need better understanding of key indicators and best practices for decision making in all areas of operations, including:

• Customer acquisition and retention
• Sales and service improvement
• Pricing and ROI analysis
• Risk management and fraud prevention
• Financial flow, valuation and forecasting
• Regulatory control and compliance

In today's demanding marketplace, leading financial institutions have no way to differentiate themselves except by taking advantage of the information locked up in their enormous volumes of data from transactions, daily operations and demographics. Timely analysis of this data will help these enterprises manage all facets of customer interaction, investment, risk, regulation, and asset evaluation to enhance customer experience and increase profitability.

The Analytic Challenges

Traditionally, financial institutions have invested money and effort in predictive and descriptive models to understand key influencers and changes in the business by analyzing the data collected in daily business operations. This approach may be used to design reports and executive dashboards as well as to understand risk and fraud, determine marketing ROI and improve business operations at every level.

However to do this, expert technical analytic teams spent weeks searching for information in a sea of data, using reporting, OLAP and traditional data mining software. To be effective today, an enterprise modeling solution must be fully automated, making it easy for an executive to search for information in terabytes of data.

Marketing and CRM

Modern financial institutions have long used predictive modeling to generate lists for direct mail, e-mail and telephone campaigns. But only a fraction of business questions were modeled because traditional tools are slow, expensive and unreliable in the hands of business users. With state-of-the art analytic technology, marketing departments can identify all underserved market segments and create, test and launch highly targeted product and service offerings for small groups of customers and prospects. Shifting from product-centric to customer-centric thought processes, leading banks are now able to design and recommend products that deliver the right solution to the right customer at the right time.

Customer Profitability: Assess the overall cost of maintaining a customer and weigh it against the predicted revenues from this customer and the predicted length of the relationship.

• Calculate customer lifetime value based on predictions of activity across all channels and products;
• Segment customers into groups according to their behavior and predicted profitability, enabling the generation of highly targeted messages;
• Assess the effect of new campaigns, products or services on the lifetime value of groups or individuals.

Acquisition and Cross-Sell: Match a specific product offering to each customer's needs and probability to purchase.

• Increase the "customer share of wallet" with cross-selling highly profitable products such as credit cards and overdraft protection;
• Expand the customer base through highly optimized acquisition campaigns;
• Use personalized recommendations to make customers aware of additional products and services and anticipate their need for the next logical product.

Retention: With many alternative offerings in a competitive and increasingly global market it has become easier than ever for customers to compare products and switch banks.

• Scoring and identification of each individual in the entire customer base for their propensity to stay with the bank, to leave and to which offers they will respond and when;
• Create specific retention campaigns for individual customer segments or product groups. Determine the important factors for increasing customer satisfaction and loyalty;
• Identify high churn risk customers in real time during inbound interactions such as call center conversations.

Operations and Finance

The Basel II accord obliges banks to reconstitute historic data with regard to elements of both credit risk and operational risk for long maturity dates (five to seven years). In the future, banks will be obliged to store and be able to reconstruct additional information, for various purposes, including updating internal models, reporting to supervisory bodies and communications intended for the markets and financial investors. These challenges can create new opportunities for financial institutions.

There are three routes to transform a "defensive" regulation into a tool for creating value:

• The calculation of the cost of risk at the level of individual units, implicitly imposed by Basel II, will allow financial institutions to establish a more accurate pricing system.
• Banks that perform well in the area of financial communication will be given better ratings by the financial markets and will be able to negotiate better refinancing rates.
• This enhanced management of margins will allow financial institutions to achieve a more dynamic and profitable management of their business portfolio.

Risk Management: As risk patterns change, the detection mechanisms must evolve with them (static filters have only short term value). New methodologies in terms of built-in deviation detection that alerts users or automatically triggers the creation of a new model as incoming data changes should be in place.

• Detect the main risk parameters per country and per sector;
• Develop a reliable behavioral score model to evaluate the probability of customer default;
• Perform back testing of credit ratings by analyzing the impact of the parameters taken for the rating estimation on the probability of default;
• Evaluate the quality of portfolios, comparing the quality of business between different areas and industry sectors.

Page 1 of 2.