The Credit Card Accountability, Responsability and Disclosure (CARD) Act has given banks less room to maneuver as they search for profits in their card portfolios. Not only are they being pinched by new regulations, their card profits are facing pressure from changing consumer behavior. Cardholders in good standing are paying down balances while delinquency rates among riskier customers rise.

But there are immense opportunities for lenders to build new advantages by taking a bold approach to risk management and analytics. The task at hand — re-engineering card profitability — demands a whole new look at the possibilities opened up by advances in analytic technology.

Essential analytical capabilities that should be in every lender's toolbox include:

  • Macroeconomic analysis. Gross domestic product, housing prices, unemployment and other factors have strong correlations to individual creditworthiness. Banks should apply analytics to explore "what if" scenarios about future economic conditions and understand the risk/revenue sensitivity of different segments under different conditions. Current decisions would then be adjusted to account for future potential.
  • Credit-capacity insight. New CARD Act regulations require an explicit evaluation of affordability before any credit is assigned or increased. Income estimators help lenders comply with these requirements but fail to answer the most important question: Can a consumer handle more debt responsibly? This question requires a sophisticated evaluation of spending behavior, credit usage and other variables. Understanding a consumer's capacity refines consumer differentiation within risk score bands. 

    By using credit scores with a credit-capacity index, lenders can offer more credit to high-capacity, low-risk consumers, and reduce lending when capacity is low.

  • Transaction-based risk analysis. A recent TowerGroup report said that "Transactions will become more important as demand for reduced latency in business intelligence and monitoring require more granular knowledge and timely tracking at the transaction level." Transaction analytics use this data to uncover early warning signs of risk in spending patterns. This knowledge enables lenders to expedite actions, such as adjusting collection strategies before other creditors do. While most card issuers rely on customer-behavior scores calculated on 30-day cycles, transaction analytics can trigger intracycle decisions, yielding a more granular segmentation of accounts within a given behavior score range.
  • Decision modeling and optimization. The leading edge for risk management analytics is optimizing individual actions and opportunities within the context of portfolio goals and constraints. Risk managers and marketers traditionally leverage ad-hoc analytics to guide decision logic selecting segmentation criteria and thresholds, and determining actions for customer segments based on intuition, experience and observed study.

Decision modeling and optimization create the ideal analytic framework to understand and evaluate alternatives based on your objective and constraints. Active and rigorous experimentation will be a hallmark of the successful issuer. The complicated dance between compliance and competition gets a lot easier when you have algorithms plotting on your side. Such sophistication will soon become de rigueur, as it enables lenders to simulate the impact of policy changes and plan the most profitable path forward.
The fight is on for the best customers. The new generation of analytics is the enabler of rapid and accurate modeling and simulation. It speeds up learning and helps banks identify optimal strategies for acquiring and managing profitable customers.

Card issuers that embrace this approach will look back on 2010 as a turning point in the long-term competitiveness and health of their businesses.

Subscribe Now

Access to authoritative analysis and perspective and our data-driven report series.

14-Day Free Trial

No credit card required. Complete access to articles, breaking news and industry data.