At Toyota Financial Services, Analytics Aid Distressed CustomersAt Toyota Financial Services, Analytics Aid Distressed Customers

Agile Analytics: 11 Ways To Get There

Toyota Financial Services is the global organization that provides retail financing, leasing, and dealer financing to vehicle dealers and their customers in 33 countries around the world. TFS USA is one of those companies, and it plays a significant role in overseeing Toyota's sales finance companies in the Americas.

Like so many other companies and consumers, this arm of Toyota's business, which was first incorporated in California in 1982, experienced the dire consequences of the financial crisis of 2008. Many more consumers fell behind on their car payments during the crisis. It put the business and the customers it wanted to serve under a new kind of stress.

As the Great Recession of 2009 wore on, car purchases fell sharply, and more people became delinquent. It was time for a change. TFS has an $80 billion portfolio.

In the wake of the economic downturn, the company had a record of more than 100,000 customers at least one day delinquent on their loan and lease payments. The portion of delinquencies more than 60-days-past-due grew 25%. The traditional collections process didn't provide answers about why delinquencies were surging, and it couldn't recommend the best way to deal with each customer.

The crisis was enough to push Toyota to take a fresh look at how it approached collections and handled consumers who were behind on their payments. That re-evaluation changed everything.

"Before 2008, we used the textbook approach to collections," Jim Bander, national manager of decision science at Toyota Financial Services told InformationWeek in an interview. That meant when someone missed a payment they entered the early-stage collection strategy. During the grace period he or she might receive a virtual message via robocall, Bander explained.

As that customer got more delinquent, Toyota Financial Services would use a predictive dialer to call every number Toyota had for that consumer as quickly as possible. Later, when another payment had been missed, that account would be assigned to a collector for management.

It wasn't an ideal way to manage delinquent accounts. It used a broad brush and treated all delinquent customers in the same way.

Weathering Global Financial Downturn

To find a better way in a time of financial crisis across the entire economy, Toyota Financial Services established a multi-phase plan to use analytics to assess risk and microsegment consumers with the goal of lowering delinquencies and keeping more drivers in their cars.

The first phase of the analytics program was about optimizing collections, the second phase was predictive analytics, and the third phase is prescriptive analytics, according to David Eddy, Divisional Information Officer for Corporate Services at Toyota Financial Services, who spoke with InformationWeek in an interview. But the whole is greater than the sum of its parts.

The entire project "combines prescriptive, predictive, forecasting, and optimization all into one framework so that we can optimize allocation of collections resources to customers being contacted," Eddy told InformationWeek.

Eddy, Bander, and their teams collaborated to create a new approach that included multiple technologies to assess individual consumers for their risk. FICO had developed an algorithm that allowed TFS to estimate which customers needed attention, and the best way to approach each of them.

Other technologies included SAS for statistics and predictive analytics, Oracle software and database software, IBM Pure Data (formerly known as Netezza), Tableau Software integrated into the user interface, Informatica for data integration, VMware for virtualization, and others. The solution relies on multiple technologies from multiple vendors. It resides in Toyota Financial Services' many data centers.

The analytics and technology, together with the cultural participation of the company, has led to a pivot in how the company approaches delinquent customers.

"The Collections Treatment Optimization allows us to microsegment our customers based on risk and other attributes," Bander said. Low-risk customers, for example, are treated differently than high-risk customers. Low-risk customers may not be contacted when they first miss their payments. They may not be assigned to collections management until they are much further along in delinquency.

By treating low-risk customers with a lighter touch, Toyota Financial Services freed up greater resources to spend on higher-risk customers. These customers are contacted earlier in the delinquency with a personal phone call when they may only be 12 or 13 days past their due date, Bander said.

The analytics "have allowed us to come up with some fairly small tweaks that has allowed us to grow our portfolio by 9% without adding staff, and in the process keep people in their cars," Bander said.

Keeping People in Their Cars

That's the number he says he is most proud of. In the first year 1,600 Toyota, Lexis, and Scion owners got to stay in their cars because of how the process changed. Some 10,000 customers avoided getting a credit marker on their credit bureau report. In all, Toyota Financial Services estimates that 50,000 external customers and 500 internal employees benefited from the project.

The path to get to this point wasn't a straight or easy path, Eddy said. The Toyota Financial Services team working on this project learned more about data, and making sure they had the right data for the project. Some data has limitations, Eddy said, so you have to make the best use of the data that you do have.

Another challenge was making sure that the teams that were building the analytics and implementing the processes stayed close to how that process was implemented in the real world. Bander said that to keep the team grounded, they were colocated in a cluster service center.

"That's so we could know how things affected the customer," he said. "The coolest thing is the trust that happens when a customer service rep actually saw me walking in the hall and told me he had a type of call that he hadn't received before and didn't understand what it meant." That close contact gave the analytics team an early view into how changes were being received by customers.

Jim Bander, National Manager, Decisions Science