BankFinancial, a Chicago-area banking chain, was sure that the best prospects for its high-balance checking account were people aged 55 and over.
That was an accurate guess. But finding them was a challenge until the bank employed Clementine, a data-mining workbench from SBSS.
And then it discovered something when planning its first 2,500-piece test: That certain people under 25—those who had received settlements or inheritances—were also fine prospects for the product.
Thanks to that newfound predictive ability, BankFinancial pulled a 6.77% response rate, according to Matthew Jacobs, market research analyst for the firm.
But that was only one part of a massive effort to make the bank more competitive in its commodity business, according to Jacobs' boss, William Connerty.
"Eight years ago, Excel was the only analytical tool we had," said Connerty, speaking last week at the National Center for Database Marketing conference. "We had no statistical tools, no data quality tools. Data was frozen in legacy systems, and there were a lot of errors in address-type fields."
That wasn't good enough for Chicago's third-largest thrift. The bank, which has 16 branches and $1.6 billion in assets, jumped into predictive analytics, "an evolving process" that continues to this day.
The first thing BankFinancial needed to do was to extract transactional data. It also had to identify households (it now serves 57,000 with a total of 71,000 customers), append demographic data and gender-code and geo-code.
The goal was to predict "who's likely to buy and churn," Connerty said.
That wasn't all. The bank has tried everything from focus groups to lateral-thinking technology. Mapping has been useful not only in picking sites for branches, but for targeting direct mail, Connerty continued.
When Clementine, a system that starts at $75,000, was deployed roughly 2 ½ years ago, it was immediately put to work locating customers for the $10,000-plus checking accounts.
The metrics were impressive, and they continued to be during the rollout. The bank is now mailing 10,000 pieces a year for the product.
The persons chosen via Clementine generate 9.4% more income and a 9% higher balance per account than other customers. And 4.3% of the total deposit dollars for that account type have been generated by the campaign, according to Jacobs.
In addition, the bank was also able to mail 79% less than it normally would have, while reducing its modeling time by 80%. And it generated the largest influx of cash deposits in one week that it had ever seen—$16,000 per customer. Finally, 98% of the account holders were retained after two years.
