BankFinancial Corp. in Chicago began using predictive analytics to develop models so that the bank can, for example, more accurately target promotions to customers and new prospects. The software it uses employs neural networks and regression routines to create these models. Regression routines are perhaps the most widely used statistics technique to estimate relationships between independent variables and dependent variables to help understand and explain relationships and predict outcomes.

William Connerty, assistant vice president of market research for BankFinancial, anticipates that the software will reduce the time it takes the bank to develop models by 50 to 75 percent. The biggest obstacle is getting transaction data and dealing with disparate data sources, Connerty says. Neural nets require large amounts of data for maximum accuracy. Unfortunately, BankFinancial works with data that originates in several separate bank systems in a variety of formats. Much systems integration and interface work needs to be done before the bank will see the full fruits of its predictive analytics tools. Connerty hopes the effort will be worthwhile. We need to increase our efficiency, our ability to deliver actionable information to decision makers, he says. Im under a lot of pressure to deliver.

Some software tools have been developed to assist businesses in gathering and integrating data for predictive analytics. KXEN Inc., a software company in San Francisco, claims that its analytic Framewotk product can greatly reduce the time it takes to define, develop, and run a model. Seymour Douglas, director of CRM and database marketing at Cox Communications in Atlanta, claims that it cuts data preparation time in half. Having clean, uniformly formatted data, however, is only the first obstacle to realizing success with predictive analytics.

Robert Berry, president and CEO of Central Michigan University Research Corp. in Mount Pleasant, thinks that lack of organization is the fundamental stumbling block that keep businesses from capitalizing on predictive analytics. Berry says predictive modeling should involve collaboration among people who have IT, analytical, and business expertise. You have to build a business-intelligence team, he says. But companies are struggling with common issues like who owns it, who manages it, and so on. How do you pull the business skills, the IT skills, and the analytical skills across corporate silos and create this team? Its not easy.

Predictive analytics is a valuable technique that no doubt will earn increasing attention of businesses of all sizes. But the current analytics systems require a sizeable, consistently formatted, integrated data warehouse plus a staff of professional who understand its potential.

1. How can predictive analytics increase BankFinancial Corps competitive advantage?

2. What is the biggest obstacle that BankFinancial Corp. is facing in operating their predictive analytics?