IBM-SAS Package Helps Automakers Mine Warranty Data

Spurred by new government reporting requirements and the need to operate warranty programs as efficiently as possible, automakers are making increased use of business-intelligence software to mine their huge stores of warranty data. And IT vendors are developing products to help them.

IBM and SAS Institute Inc. this week unveiled a package of data-management and -analysis software and services for collecting and analyzing product-safety information and reporting it to the National Highway Traffic Safety Administration under the Tread (Transportation Recall, Enhancement, Accountability and Documentation) Act. The legislation, passed because of the agency's failure to quickly spot defective Firestone tires that allegedly led to SUV rollovers in 2000, requires automakers to report all safety-related information such as warranty claims, consumer complaints, and production data starting April 1.

The IBM-SAS quality-warning package combines SAS applications for analyzing structured data such as warranty-claim statistics with software and services from IBM, including data-integration tools and a rules engine application for analyzing unstructured data such as handwritten reports.

The software will help automakers and parts manufacturers comb through terabytes of data--one early user of the product is pulling data from nearly 60 databases--for accurate analysis and reporting to the traffic safety agency, says Larry Stolle, IBM's global business development executive for the automotive industry. Stolle could not disclose early adopters of the product. Sold by IBM and SAS, the package is priced according to the implementation.

The auto industry will ultimately use these tools for more than Tread reporting, Stolle predicts. "If they stop there, they lose a lot of the value they can gain. They will fail to exploit the investment made to meet a legislative requirement," he says. Stolle expects automakers to use the system to spot potential problems with products, compress problem-resolution times, and improve manufacturing processes. Cutting just 10% from the time it takes to identify and correct problems with auto parts can save millions of dollars, he says.

While some automakers already analyze warranty-claims data, that analysis is limited to structured data such as codes for specific repair jobs, says AMR Research analyst Kevin Prouty. Such efforts are unable to make use of unstructured data such as written comments from mechanics. Automakers also have so much data that analyzing it and acting on the findings can take months. Products such as the IBM and SAS offering should help remove both impediments, Prouty says.

Ford Motor Co. is already leveraging nearly a terabyte of warranty-claims data to help 10,000 Ford dealers improve their service operations. Ford has built what it calls its global warranty measurement system, which incorporates a Teradata data warehouse with 24 months of warranty data and Information Builders Inc.'s WebFocus business-intelligence software.

Ford uses the system to analyze the data to see how many repairs individual dealers are performing and the cost of those repairs. The system then compares those results to other dealers in the region (or within the country for European dealers). That makes it possible to spot instances in which dealers may be performing unnecessary repairs or are spending too much--and therefore charging too much--to service cars. "Our objective is to identify as early as possible a dealer who is trending in the wrong direction," says Jim Lollar, the automaker's global warranty operations IT manager. Ford then works with the dealer to correct the problems.

While the program helps Ford lower its warranty expenses, Lollar says the real goal is to help dealers improve service provided to customers. Ford has provided basic printed reports to dealers for two years, but last year, the automaker implemented a system that provides dealers with Internet access to monthly WebFocus reports with charts and graphs and the ability to drill down into the report's underlying statistics.

Ford is also leveraging the data warehouse for other tasks, such as forecasting future warranty costs and identifying gaps in dealer technician training. Says Lollar, "That's a huge amount of valuable data."