Prying the lid off of warranty data

General Motors knows all about its warranty issues. And Toyota knows all about its own.

But GM doesn't have access to Toyota's. And suppliers that sell the same part to both automakers often don't have information on how those parts are performing in the field.

This is the nature of proprietary data, repeated thousands of times all over the auto industry.

But an information service company — We Predict, of Ann Arbor, Mich. — is working to pull back the curtain on parts performance data and provide transparency across corporate lines.

"The idea is to help the auto industry," said Renee Stephens, the company's vice president for North America. "Certain parts are potential cross-industry problems. But no one shares their data. So if a problem has shown up in your data but not in mine, how would I know what's coming at me?

"You don't want the next airbag crisis to hit you in the face."

We Predict gathers warranty claim data from a variety of sources, including dealerships and independent service shops. Stephens says the company has built a database of more than 500 vehicle models, with stats on at least one-fifth the population of every model. The database currently goes back 10 model years, but We Predict is extending it to 15 years.

The company is selling a predictive analysis that projects when a given part will fail, based on how it performed in the past. The data it is now offering to auto companies is essentially what insurance companies rely on to operate — actuarial data. We Predict was founded by an insurance executive, James Davies, formerly with Lloyd's of London; he is now We Predict's CEO.

That product offering sounds simple. But the potential implications are wide-ranging.

An automaker could ask for the performance data of a competitor's individual part, or of every part used by the competitor, or of every part in the industry.

Easier access to warranty data would permit one automaker to benchmark how a competitor is faring on the performance of a given part. Historically, an automaker knew only how its own parts were performing. And while a manufacturer might improve its warranty performance year over year, it still wouldn't know how that performance stacked up against rivals.

That insight will play into the hands of automotive purchasing managers, Stephens said. Buyers would be able to see how a component performed over time inside of a competitor's vehicle to determine whether to buy it.

But conversely, the information could bolster a supplier's sales pitch by providing third-party proof of a part's performance, said Stephens, who previously worked as director of quality data management at GM International Operations, and at J.D. Power.

Stephens said she encountered the firewall of proprietary parts data 10 years ago at GM when the company experienced problems with a supplier component. GM was aware that Toyota was using the same component and wondered if the quality problem was occurring there, too. But there was no way to acquire the information. Without it, GM couldn't be sure whether the problem was coming out of the supplier's plant or out of GM's.

Stephens got around the mystery by simply picking up the phone and calling a friend who worked at Toyota and asking about the part in question.

That back-channel call resolved GM's immediate question, but Stephens recognized that if automakers would only share their data — gingerly stepping around antitrust concerns — supply chain issues could be resolved faster. But legal department concerns about data sharing proved too challenging, she said.

In a more recent situation, We Predict spotted that a part being used by an automotive client was suddenly showing a higher-than-expected failure rate. And managers at We Predict happened to know that the same part was about to be sourced by a second automaker.

For proprietary reasons, We Predict could not notify the second automaker. Instead, the data company alerted its client to the situation, and asked them "if they would like to make their competitor aware of the part problem," Stephens said.

"And they did," she said. "Their attitude was, 'We're going to have to talk to this supplier — and it would be better if they were hearing about this from two customers instead of one.' "