February 14, 2013

Warranty Keynotes:

The typical warranty manager is going to transform from a data processor into a data communicator, according to the predictions of the keynote speakers at next month's WCM Conference. They're going to become the eyes and ears of the corporation, advising other departments how their products are being used and what's causing them to break.


When the Warranty Chain Management Conference convenes next month for the ninth time, warranty professionals are going to be wondering what's new.

We've been to eight WCM Conferences, of which two have been in the same hotel in San Diego where we're headed again in March. So as we begin our previews of the proceedings in this week's newsletter, the top question is going to be: what's new and different this year?

The keynote speeches will be a big part of that. Paul Sonderegger, the senior director of analytics at Oracle Corp., delivers the opening day keynote speech on the morning of Wednesday, March 13. Joe Barkai, the research vice president of IDC Manufacturing Insights, delivers the keynote on Thursday, March 14.

Encore Performances

Both are familiar faces at past WCM gatherings. But this year, both have a provocative new message to deliver: that warranty has the potential of becoming a new kind of communications and intelligence hub for manufacturing companies, connecting and collaborating to distribute the valuable data that comes from claims processing, and combining that data with other sources to create even more valuable new metrics.

This will be the second time Sonderegger has kicked off the proceedings with a keynote speech. But it will be the first time he's done so since Endeca Technologies Inc. was acquired by Oracle in October 2011.

Endeca has always been involved in business intelligence, and in regards to warranty, in the provision of analytics that help a company and discover patterns of failure in their claims data. Since the acquisition, those enterprise search and discovery capabilities have been pointed towards some very large databases operated by Oracle clients.

Paul Sonderegger

Sonderegger told Warranty Week that he plans to make his speech not so specific to Oracle or to databases as it will be to big collections of data in general. He said he thinks that the harnessing of big data is going to transform the warranty department from one that merely processes data into one that interprets and communicates that data.

"Oracle sees warranty as one of these industries that will be transformed by the advent of big data, and aims to do everything it can to serve it," Sonderegger said.

"It's not just the database," he added. "Oracle is squarely in the middle of this big data revolution," he said. For instance, Oracle is deeply involved in not only the storage of data, but also in reducing the cost of analytics and enabling the exploration and discovery of data.

"But the big story in big data is big variety," he said. "It's not big volume. That's going to happen, and Oracle and other large vendors already know how to solve that problem. There are plenty of data warehouses out there that are already petabytes in size [a petabyte is a million gigabytes]. So that's not the big story.

"The big story is in variety. And the reason it's a big story is because the world is making a digital copy of itself. When that happens, companies can expect to have data about just about anything they want to know about, at acceptable prices, virtually immediately.

"One of the things that will become obvious is that all of the data we have captured so far is a very thin strip of the information that is available about, say, warranty issues, or quality issues, or product design. And we're about to hit this point where not just the amount of information but the variety of it is going to multiply," Sonderegger said.

Interpreting Data Correctly

Of course, there's also a big problem: The world's ability to produce data has outstripped most organizations' ability to put it to good use. "You can see this all around us," he said.

For instance, he said, during the 2012 U.S. Presidential election, there were so many polls that touted such diverse outcomes, that most of the media concluded the race was too close to call. However, Nate Silver, who runs the FiveThirtyEight blog for the New York times, looked at the same data as everybody else, and concluded that the outcome was never in doubt.

In fact, he predicted all 51 of the electoral college outcomes correctly, and was only 0.3% off with his final popular vote projections. For the U.S. Senate, he picked the winner in 31 of the 33 contests. That degree of accuracy was better than any other newspaper, network, or website has ever managed to do. And for Silver, it was his second time besting them all (although in 2008 he got only 49 states right).

"By combining this variety of data in new ways, he was able to call the Presidential election," Sonderegger said of Silver. "And the way that he did it is really instructive: by combining polls. Each of those pollsters, they didn't intend for their polls to be combined with anybody else's. That wasn't their plan. But it was Silver's plan. He realized that if you take a whole bunch of different polls and put them together, you can actually reduce the margin of error on what they're telling you."

Another example Sonderegger cites is the most recent financial crisis that became a full-scale panic in September 2008. Michael Lewis, in his book, The Big Short: Inside the Doomsday Machine, profiled several of the biggest winners and losers of that era, such as Oppenheimer bank analyst Meredith Whitney and money manager Steve Eisman, who saw it coming, and Joseph Cassano of AIG Financial Products and Howie Hubler of Morgan Stanley, who evidently did not.

Those who saw it coming shorted the market and made billions. Those who did not lost billions. And what Sonderegger said is the most amazing is that both sides had access to pretty much the same data on the housing market, on interest rates, and on the prices of credit default swaps, derivatives, and index futures. The difference was that some interpreted the data correctly while others did not.

Those who saw it coming, Sonderegger said, did the same thing as Silver did with his polls: they put data together in new ways to sharpen their ability to predict future events. "They took the time to combine diverse sets of data about these complex financial instruments and the mortgages underneath them in order to get a completely different understanding of where the risk was in the housing market."

In other words, it's not the data. It's the data processing.

Product Usage Data

"The reason that big data is such a big deal for warranty is because warranty is such an information-intensive activity," Sonderegger said. "When you start looking at root causes, for example, in order to shorten detection to correction time and even to improve the design of future versions of the product, this big variety is a gift. It provides not just a greater amount of data, but data about more aspects of the problem -- data about how that product is actually being used in the world, the interaction among its components, or the environment it ends up operating in."

The sheer variety of data will enable warranty professionals to get to the root causes faster and with fewer resources, he said. But they'll still need to choose what data they want, and what to ignore. "What do you wish you had data on, that you don't have data on today?"

And let's not lose sight of the fact that while half of a warranty professional's job is to pay the costs and process the paperwork for past failures, the other half is all about predicting the future accurately. Warranty professionals are expected to set accruals at precisely the right level, so there are no surprises to either the upside or the downside. There's a little Nate Silver in all of us.

However, it's not just the numbers. It's the relationships and the behaviors too. The warranty department of the future, Sonderegger said, will really begin to function as a business intelligence hub for the company.

"Warranty departments can expect to play a new role in not just processing warranty claims and resolving warranty issues, but also a new role, which is this greater authority about how products are being used in the real world, potentially increasing their profile and their role in design," he said.

Such a change is going to require a huge amount of collaboration, co-operation and communication, Sonderegger noted. "Not just in different divisions within a manufacturer," he added, "but up and down a supply chain. So a Tier One auto supplier that would like to reduce its warranty cost can potentially use the data generated by the OEMs' vehicles to better understand the causes of breakage and failure, and then to improve the design of the next generation of the product, not only reducing their own warranty cost, but improving the quality of the products they provide to the OEMs."

Big data technology also will reduce the cost of combining different data streams, Sonderegger predicted. And just like Silver did with his polling data, or as he did even earlier in his career with professional baseball statistics, the value of combining existing data in new ways will be enormous.

Advice for Warranty Professionals

"There are three pieces of advice that I would give to warranty managers," Sonderegger said. "First, think in terms of data market share. Imagine, every single part, every single component, every single product they have responsibility for -- to imagine those things out in the world producing streams of data."

"That is the total market of data. And the question for the managers is what share of that data market do they have immediate access to? Now, some of that data might have to be bought. Some of it can be acquired through contract. Some of it can be acquired by adding instrumentation to the products. But that's the first thing to think about.

"The second thing to think about is creating proprietary data assets. Obviously, if you're thinking about data market share across all the players in your value system, whether they're downstream or upstream -- well a lot of that data is going to be available not just to you, but also to your competitors.

"So what do you do? Here's the funny thing about combining multiple sources of data: When you bring in data from a third party, and combine it with data only you have, that whole collection becomes proprietary. It's like a drop of milk in a cup of coffee. The whole thing becomes a new drink.

"So that's the second thing for warranty managers, to think about creating proprietary data assets. Even when they're bringing in data from the outside, combine it with proprietary data only they have on the inside, and that will actually give them a proprietary perspective on that market data.

"And then the third thing is to use data to make data. If you can get a better understanding of the true performance and use of your products out in the world, and use that to refine the performance of those products, you may be able to capture a greater share of the customers out there, improving your ability to generate proprietary data that your competitors don't have," Sonderegger said.

He compares it to the way Google operates. When you search for a topic, Google notes the way you respond to the results it offers. Then it uses your responses to improve the results offered to the next person who searches for that same topic. Amazon.com and many other websites do the same thing: use data about you to make data for others.

Warranty Communications

Joe Barkai

The next day, Joe Barkai will present a somewhat different take on the same topic: How warranty departments can be transformed from data processors into business intelligence centers.

Barkai, who joined IDC shortly before the first Warranty Chain Management Conference back in 2005 and has spoken at every event since then, said he's seen the beginnings of a turn from warranty as transaction management to warranty as a data analysis hub. The proper interpretation of current warranty data can have a positive effect on future warranty costs, he said, as with the automotive industry's early warning efforts and the implementations of warranty analytics packages by top manufacturers in other industries.

But he still doesn't see manufacturers paying attention to the ways that changes in purchasing, design, or service can have an adverse impact on warranty. Top executives are still not paying that much attention to what's happening downstream. It's not the software. It's the people.

For his Thursday morning keynote speech, entitled, "The Quest for Zero Warranty: Breaking the Warranty Vicious Cycle," Barkai told Warranty Week he plans to talk about the product development lifecycle and its effect on warranty. Or perhaps it's really about warranty's effect on product development?

Decisions Affecting Warranty

Barkai said he sees warranty generally being at the end of the process, and having to clean up the problems caused by earlier steps in the process. For instance, purchasing finds a new source for parts at rock-bottom prices. But they all fail and need to be replaced under warranty. The company's warranty costs soar. But purchasing is cutting costs and has no clue about the rest of the story.

And then there are the design people, always looking for an edge over the competition. It's not just more cup-holders. It's all about new and improved systems and features, some of which don't end up working out exactly as intended. Once again, it's the warranty department that has to clean up the mess.

"As product development becomes faster," Barkai said, "we tend to invent and innovate more so than we tend to reuse." And so, designers tend to lose the accumulated value they gained through their familiarity with existing parts over time.

"You have the best practices, you have documentation, you have the tooling, plus, you also have prior experience in regards to quality, service, and warranty." But its value is lost when the design changes.

On the other hand, sometimes designers reuse parts and modules that could have benefited from some innovation or refinement. The same flaws and defects are passed on from one generation to the next. The root cause, if you'll pardon the pun, is a lack of communication. Nobody ever told them what a problem a given part was causing.

"Why are we repeating the same mistakes again and again?" Barkai asked. "We're not learning. So while I understand that the reality is it's impractical to build zero-defect products and to get everybody to perform perfectly, we're not leveraging the experience and information that we have."

Warranty Data Driving Design Changes

The key is for manufacturers to put that failure analysis and warranty claims data to work, and using it to affect future product design. "We really don't pay enough attention to other activities that in the end manifest themselves as lower quality, more failures, poor service, and therefore higher-than-needed warranty expenses," Barkai said.

In other words, it's not the data processing. It's the data communications.

"I think the main reason is inadequate internal processes," he said. "And it's the proverbial silos in the organization. Design doesn't always understand manufacturing. Design usually does not pay enough attention to service."

In previous WCM presentations, Barkai talked about designing for warranty, a subset of the topic of designing for serviceability. Frequently-replaced parts should be easier to replace. For instance, desktop computer manufacturers used to secure their disk drives with three or four small screws, some of which were on the inward-facing side of the unit and were therefore hard to reach. Now, most disk drives snap into place using brackets, and it's so easy to do that many customers can perform the repair themselves. And that lowers warranty labor costs.

"But if we don't have adequately-trained service technicians, he or she may replace two parts instead of one. Warranty costs go up, the no trouble found rate goes up, and so on and so forth," he said.

In the automotive industry, the way technicians are paid for warranty work makes them tend to do less diagnostics and more parts replacements, Barkai suggested. "Technicians in the U.S. work on a flat-rate basis. So they get paid the same amount for a job, no matter how much time they spend doing the work. Time for adequate diagnostics is not included in the flat rate. It's much more focused on wrench time. So a technician is motivated to do a fast job that will get the car out of the service bay. And this is likely to increase warranty costs, because you may replace higher-level parts, or two parts instead of one, etc."

Worse With Electric Vehicles?

With electric vehicles, it may actually get worse before it gets better, Barkai suggested. Technicians are already telling IDC Manufacturing Insights that they don't believe they fully understand the new systems, or they worry about their own personal safety while working with them. "And we found there's quite a gap between the tools they have and how they feel about their ability to service the vehicle properly," he added.

But even with combustion engine vehicles, there's a tendency to innovate through new software designs, especially with systems that control entertainment, communications, and passenger comfort. And while not meaning to insult automotive software designers too much, Barkai notes that they tend to look at bugs as mere pests that can be fixed later at low cost, not, as with hardware, as important safety issues that could be costly to recall and replace. We'll fix it in version 1.1, they seem to think.

"The rush to innovate -- the rush to put all this innovation in software -- the tendency to cut corners and say we'll fix it later -- is not helping here," Barkai said. Also not helping: the tendency for consumers and the media to jump on software defects as quickly as they can, using social media and mobile communications to spread the word.

"All these pieces come together," Barkai said. "We call it product lifecycle economics. How do you actually optimize decisions, understanding the entire value chain or the entire product lifecycle?"

Sheila Brennan

After Barkai finishes delivering his keynote, some chairs will be shifted around and he will take part in a panel discussion moderated by a member of his team: Sheila Brennan, the program manager of IDC Manufacturing Insights.

This panel discussion wasn't initially designed as an expansion of the themes that Barkai plans to discuss in his keynote speech, but it may turn out that way. Brennan said it was supposed to serve as an update of what's been going on with the Warranty Capabilities Maturity Model, which it will provide.

But she said her subsequent research has been coming to the same conclusion as Barkai's: It's not just a software problem. It's more of a people problem now.

Other members of the panel will be returning participant Kev Yacoubian, director of warranty operations at Volvo Trucks, as well as two new participants: David Christian, the senior manager of dealer support at Mitsubishi Caterpillar Forklift America Inc., and Andrew Kuglitsch, the lead of warranty operations at Harley-Davidson Inc.

"We wanted to give an update on the progress of the Maturity Model," Brennan said. "That's what we were asked to do." But after looking at last year's warranty-related financial metrics and after talking to numerous warranty professionals since the last conference, Brennan said it appears that perhaps the warranty industry is slipping backwards just a little bit.

Changing the Mindset

Therefore, Brennan said she wants to make an attempt to change the mindset a little bit, from viewing warranty as something that happens to you, as if you're a victim, to seeing it as something you can at least partially control. And the way to control it is to let the folks know in purchasing and design, as well as those who work for parts suppliers and repair organizations, how their actions are affecting the warranty process.

"We have to work smarter, not harder," Brennan said. "And in order to do that, we have to reach into other parts of the organization. We have to think about warranty in a different way. We have to think about warranty in context of the whole product lifecycle."

In the past few months, IDC Manufacturing Insights launched an effort to get major manufacturers to participate in a wide-ranging and long-lasting warranty benchmarking study. "This is a real quantitative effort, looking at the numbers and comparing them across companies, seeing the highs and the lows, and being able to see what things are impacting other things," she said.

First, representatives from the participating companies made a list of the metrics they each collected. Then they compared lists, and Brennan said they found that roughly a fifth of the metrics, or at least the labels, were unique to one company. So then they had to decide which metrics were the most important, and how they should be defined.

Then came the hard part: actually collecting the data. "As you can imagine, that can be like pulling teeth, especially for these folks that typically don't have the best seats at the corporate strategy table," Brennan said. "It's a little harder for them to get the information."

In some ways, it's a kind of sequel to the Maturity Model -- an exercise that further sharpens the view a company can get of its own warranty process. But there are some important differences. Unlike the Maturity Model, the exercise will not be handed off to the Institute for Warranty Chain Management -- now renamed as the Global Warranty and Service Contract Association. And while Manufacturing Insights is trying to keep the cost of participation low for its clients, it won't be free.

"Now we're trying to create this long-term vision, and implement it, and set goals," Brennan said. The benchmarking exercise is an initial effort to see where the biggest gaps are, how they can be bridged, and who needs to be involved. And that last item may be the most important of all.

A People Problem?

"Part of that exercise was getting a group together, and a sponsor, and kicking it off. In some cases, that was much more difficult than others. And that just goes to show, that's part of the problem."

Brennan said some clients have implemented claims management software and warranty analytics software, as they were urged to do by not only the Maturity Model but also several market research reports published last year. But they're still not seeing the benefits.

"And you have to wonder, where are the disconnects? Why aren't they getting the benefits? Well, they're still not getting the data. They're not getting the support. They're not getting what they need from the rest of the organization."

Brennan said the benchmarking exercise is not just about gathering operational metrics. It's also asking about the level of support from the IT department, the percentage of claims data that's shared with suppliers, and the amount of collaboration with suppliers regarding warranty issues.

Another question asks what percentage of claims data is being shared internally with purchasing agents and with design and engineering. "You'd think it should be 100% across the board," she said. "But it's not."

In other words, it's not the data communications. It's the communications, period.

It's how companies are looking at the problem, who in the company is involved, and how much they're communicating. It's the mindset that needs to be changed now. The data is available. Now it needs to be shared properly.

"Companies haven't looked at it this way before," she said. "So I think for these folks to be able to bring this back to their executive management and have this kind of information and data presented back in a very real way to a cross-functional team that's beyond just warranty is going to be so powerful, and it's really going to empower the warranty groups."

Mize Warranty Connect
Warranty Chain Management Conference
PCMI - Your technology partner
After Warranty Analytics
Sign up for a free subscription to Warranty Week:
     subscribe     change of address     unsubscribe