Early Warning Systems:

Does anyone still need to be sold on the benefits of early warning systems in warranty claims processing? In brief, the theory is that warranty claims can provide an early warning of a production problem. The sooner defects can be detected and corrected, the fewer the number of products that will contain them. The proper analysis of current warranty claims can prevent future warranty claims from ever leaving the factory.

It makes perfect sense. It hardly requires much explanation. Automakers and their suppliers have been among the first to endorse the concept. Now they've already moved onto the next step: standardizing their warranty processing systems so that the warnings can be communicated sooner and more easily, especially when numerous companies and multiple types of computer systems are involved.

Computer and major appliance manufacturers are next. As with the automakers, they really don't need to be sold on the benefits. The benefits seem so obvious that it's a wonder they even need to be detailed. However, the problem isn't with the benefits. It's with the cost -- specifically the high cost of the software.

If a warranty claims processing or analytics package costs a million dollars to implement, will it result in a million dollars in savings? What exactly will be the return on investment? Will the software pay for itself in a month? In a year? Ever?

Despite the aggregate for manufacturers' warranty claims adding up to more than $25 billion last year, the truth is that only 473 companies reported more than $1 million in claims, and only 187 reported more than $10 million in claims. Therefore, for most manufacturers the sad fact is that at current prices the cost of these benefits can never be justified in terms of savings. And it's not just the cost of licensing the software that needs to be considered. Even if the software were given away for free, it would still cost a considerable sum to install and customize. There's no such thing as a shrink-wrap warranty application that installs in a few minutes and pays for itself in a couple of weeks.

The VisiCalc Revolution

More than 25 years ago, VisiCalc revolutionized the world of numerical calculations. VisiCalc, the spreadsheet predecessor of Lotus 1-2-3 and Excel, sold for only $100 a copy, yet it allowed number crunchers to do in 15 minutes what used to take some of them 20 hours to do by hand. As of yet, there is no VisiCalc of warranty software.

Many companies in fact are still stuck in the VisiCalc stage. They find that one Excel package and maybe one good graph-drawing package in the hands of one clever employee is enough to allow them to look after their warranty claims. Of course, the million-dollar warranty software packages are much better than those simple homegrown spreadsheet systems, but are they a million times better?

At a manufacturing industry conference hosted recently by the SAS Institute at their Cary, North Carolina headquarters, what emerged as a common theme was that warranty analytics packages are highly beneficial to those who have the resources to implement them. But the decision to implement them doesn't always come down to simply a question of cost versus benefits. Many times, there's another reason behind their decision that has more to do with an opportunity that presents itself.

Chuck Wodke from Allison Transmission told the gathering that his company decided to replace both its warranty claims processing and its warranty analytics capabilities at the same time. But, as was recounted in a Feb. 23 article in Warranty Week, what the company was really replacing was a career's worth of personal knowledge accumulated by Dave Ellis, Allison's about-to-retire warranty expert.

Ellis did his job well, but as he announced his retirement the company realized it was relying very heavily upon one person to be a walking, talking data warehouse. Whenever anybody had a question or needed a report, they came to Dave. Therefore, the computer applications that would take his place would have to be accessible to a wider group of users, so they could make their own reports.

Merger Benefits

Dwight McAlexander, the director of dealer systems and processes at Volvo Trucks, said his company seized upon the opportunity that presented itself as it acquired Mack Trucks four years ago and the two began to integrate their respective data processing systems. "We decided at that time that we needed an analytical system that we could use globally," he said. Rather than replacing one business unit's warranty system with the other's, the merging companies decided to replace both systems with something new.

At Whirlpool, there was no lack of computer systems or databases. John Kerr, the company's general manager of quality and operational excellence, said the problem was that those systems couldn't take Whirlpool where it needed to go. There were lots of early warning systems all over the company, but they used different formats and didn't communicate well.

Kerr told the audience that he was training dozens of engineers at a time in the science of six sigma -- how to reduce variation and increase reliability and how to make histograms and pareto diagrams out of all the statistical data at their fingertips. The problem was, the data wasn't always available in a usable form. He said sometimes the engineers would have to re-key the data because it was stored on a legacy system. So the engineers were learning techniques their data processing infrastructure couldn't always support.

Kerr said he was also training his people to look at the big picture: the total cost of quality. Warranty, he said, is only a small portion of that cost. There's also the cost of repairs made outside of warranty, the cost of retooling and reworking manufacturing processes, the cost of scrappage and waste, and much more. In sum, the total cost of quality can be three to five times the cost of warranty alone, he said.

"Any company can come back and say 'here is my cost of warranty,'" he said. "But there is so much more out there."

The Cost of Lost Sales

Not even considered is the cost of lost sales, as when a dissatisfied customer warns a dozen of their friends and neighbors to avoid a certain brand and/or a certain retailer. "I'm in lost sales prevention," Kerr said. "It's all about loyalty. My job isn't just to reduce the defects. It's to make sure customers continue to buy my products."

With this in mind, he went from one department to another within the company, talking about the value of not just better transaction processing, but better business intelligence. "When an appliance fails, help me figure out why it failed," he said. That's not just a matter of statistics, but it has to begin with some number crunching. "It's not just better data but better information."

Kerr said his chief financial officer was all in favor of better intelligence, but he at first said the company's SAP system could do it. It wasn't until SAP's own analysts said they couldn't do it that he was cleared to look elsewhere.

The problem is that better intelligence requires better data, or more precisely, bigger databases. Whirlpool is manufacturing something on the order of ten million units per year -- everything from washing machines and refrigerators to air purifiers and water coolers. Many of those units contain 400 or more components. So to answer that simple question of why something failed, Whirlpool would have to look after something on the order of four billion parts, many of them provided by its suppliers.

The SAS warranty analysis system Whirlpool is implementing was 2.5 terabytes in size even before it was turned on. The SAP transaction processing system that contains all the customer records and sales data is only slightly larger at 3 terabytes. "So just my quality information is almost the same size," Kerr noted.

The Sea of White

Kerr said that in a business such as major appliances, a manufacturer must differentiate itself from its competition. Otherwise, it's going to get lost in what he called a "Sea of White."

"You walk into Lowe's, Best Buy, Sears, anybody -- we call it the Sea of White. There's all the washing machines lined up; there's all the dryers lined up; the refrigerators. Everybody has them, and they're kind of common," he said. In such a situation, customers can be forgiven for shopping on price alone.

"Innovation to us is the way we have to break out of that commodity industry -- to break out of this Sea of White," Kerr said. So Whirlpool is pushing ahead with innovations such as a refrigerator/oven combo that keeps a dish cold during the day, heating it up just before its owner gets home from work. Delayed at work or a change of plans? After keeping the food warm for a while, the unit will go back to keeping it cool.

Whirlpool also is developing machines that can take the place of dry cleaning, using chemicals that freshen and take the wrinkles out of clothes. And it's selling small dishwashers built into kitchen sinks, and washer/dryer combos that finish their cycles at exactly the same time.

Sadly, with all innovations comes an increased chance of product failure. Consumer Reports still tells its readers to avoid buying automatic ice makers built into refrigerators. It's not that they're against the concept of ice; it's just that they believe there are more reliable ways to make it. It will be interesting to see what they think of dishwashers built into sinks or units that can both cook and chill food. And it will also be interesting to see if claims rates go up or down company-wide as a result.

It's still too early to tell what kind of a return on investment these analytics systems will provide. Most of them haven't been in use long enough to allow for meaningful calculations. But one thing is for sure: the investments the major manufacturers are now making in early warning and business intelligence systems are so large that the returns will have to be even larger.