Dealer Services, a company that lends money to dealerships acquiring used cars, is trying to use predictive analytics to make money when its rivals can’t.
The company now studies more complex data than it did in the past in order to better predict which loans will pay off. The “Aha” moment came in 2008, when, as the economy tanked, the private lender began to look beyond simple data points such as loan volume and number of customers. “There was underlying data from dealers that indicated something was brewing,” says CIO Chris Brady, including where a new customer had previously gotten loans. If a dealership was coming to Dealer Services because it no longer had a line of credit at a bank, that was a warning sign of what was happening in the used-car market, she says.
To read more on this topic, see: Forrester: SAP, Others Will Make Analytics Acquisitions and To Hell with Business Intelligence: 40 Percent of Execs Trust Gut.
Traditional business intelligence (BI) might point you in a direction, but predictive analytics aims to uncover a treasure map, says David White, a senior research analyst at Aberdeen Group. That’s because BI identifies relationships between a few data points, while predictive analytics evaluates how many factors work together. BI vendors are now offering predictive analytics tools that used to be available only from niche vendors such as SAS and SPSS.
White knows of a department store chain using predictive analytics to formulate more profitable coupon campaigns by targeting the right customers. If a store sends a coupon to a customer who was going to make a purchase anyway, the store is no further ahead. But send the same coupon to a shopper who wouldn’t have otherwise come in, and you’ve made money, White says.
Dealer Services launched in 2005, and it grew so fast that within six months, Brady says, it had met its three-year goals for revenue, number of loans and customers. The company doesn’t reveal financial figures, but it now has 70 offices across the United States and serves 11,000 dealerships. It also finances dealers who lease cars and who sell power sports vehicles such as snowmobiles. This growth showed that the market for used-car loans was ripe, Brady says, but the company needed better BI to understand all that was happening.
Dealer Services originally analyzed data in the same manner as in the rest of the industry, working off basic reports it wrote internally and some Microsoft Excel spreadsheets. Brady says this approach led to some individuals and departments using different numbers for the same reports, which slowed down decision making and hampered forecasting.
Brady brought in Information Builders tools to do real-time analysis of how loans are performing. Managers now study data they hadn’t paid much attention to before, she says, such as the age of the loans on used SUVs and trucks. Some dealers were buying those vehicles at the same pace that they had in flush times, but the vehicles sold more slowly, raising the specter of more loan defaults. That made Dealer Services change how it monitored those dealers, she says. “The number of loans can be big for good reasons or bad,” says Brady. “If you don’t know the difference, you’re in trouble.”
Contact Senior Editor Kim S. Nash at email@example.com.
Do you Tweet? Follow me on Twitter @knash99. Follow everything from CIO Magazine @CIOMagazine.