How Toyota Customer Data Spun OutBy Sean Gallagher | Posted 2002-06-13 Email Print
Know the Risk: Digital Transformation's Impact on Your Business-Critical Applications REGISTER >
Toyota Motor Sales has finally gotten a handle on its customer data after hitting some serious potholes.
If you own a Lexusor know someone who doesyou know how much value the luxury auto brand places on customer service.
Take a noteworthy case in 1998. Toyota Motor Sales USA (TMS) responded to a recurring service problem with its Lexus vehicles by contacting owners, then picking up their vehicles, taking them in for repairs and leaving loaner cars as replacements. When done, the owners' cars were returnedfixed, washed and tanked up.
That level of service was made possible by TMS's Corporate Customer Information System, an application used by the company's call center in Iowa to help handle warranty, roadside assistance, prepaid maintenance and other service requests.
But 1998 didn't go altogether smoothly. Relying on data in that system, TMS began to mail checks to Lexus owners to replace troublesome tires. The checks, for more than $400 each, in some cases went to people who didn't even own a Lexus. One errant check even found its way to a Toyota auditor, for a vehicle he hadn't owned for a while.
"You can imagine the repercussions of that," says John Gonzales, data quality manager at TMS. "We can't afford to be giving money away to people who shouldn't be getting it."
The glitch was symptomatic of a bigger problem. The system depended on customer data stored in "roughly 15 databases in different parts of company," says Gonzales.
Just to get to all the data about a customer, call center employees would have to navigate through four or five mainframe applications, while customers waited.
The glitch resulted in a mandate from Toyota's office of the president for a centralized, single customer database. And Gonzales was tagged to make it happen.
"The main goal was to service customers' calls quicker," he says. But also, "as the volume of calls went up, we didn't want to increase the number of people in the call center."
Finding a way to pull off the consolidation turned out to be no mean feat. Gonzales and his team looked at a procession of products purportedly offering "business intelligence," but none fit the bill.
Then, a solution was nearly dropped into his lap. While attending a conference, Gonzales saw a presentation on Informatica's data extraction, transformation and loading software, used to pull data from a wide variety of sources into a central data warehouse.
He invited Informatica reps to make a pitch. A proof of concept test was done, and Gonzales was sold. Key to the buy was Informatica's low learning curve: "We were coding within a month," says Gonzales.
By April of 1999, Toyota had Informatica's PowerMart installed. But building the centralized customer database wouldn't be "plug and play." It would take another six months for TMS to get the exchange of data working well.
The biggest nightmare was cleansing the data. "The amount of work to do was frightening," he says. "Some databases had wrong vehicles, wrong motors, wrong addresses."
Plus, there were several million records to be checked. To ensure accuracy, Toyota had to turn to outside sources in some cases, including states' departments of motor vehicles. In some difficult exceptions, customers had to be called directly to straighten records out. The resulting data warehouse filled 250 gigabytes of storage, and grows daily.
The total cost of the rollout? More than $2.5 million.
But Toyota believes it was worth it. Toyota has been able to avoid increasing the size of its call center operation, while vehicle sales have grown from 750,000 a year to 1.7 million. The centralized database is now used to help make decisions throughout the company.
Gonzales' job may never end. The data scrubbing continues to this day, with data from Mexico the latest to be rolled into the system.
Yet he couldn't be happier. After 24 years at Toyota, Gonzales is excited, because "Toyota realizes the value of this data."
Especially when it's correct.