Harnessing Big Data:
As more "big data" sources become available, extended warranty administrators can expand from break/fix into highly customized offerings, preventative maintenance and product monitoring services. But first they will have to decide what data is both meaningful and reliable.
For decades, the extended warranty business has been centered around one value proposition: What are the odds of a product needing a repair, and what is the average cost likely to be?
Multiply those together, and you have an estimate of the break-even price of an extended warranty for both the consumer and the provider. Then, if a repair is eventually needed, the consumer wins the bet and the provider spreads his loses over all the service contracts that expire unused.
Big data is going to radically change this equation, according to Aleem Lakhani, executive vice president at AMT Warranty, a unit of AmTrust Financial Services Inc., and a longtime sponsor of this newsletter. Big data is going to drive providers into deeper relationships with consumers, based not only around the odds of needing break/fix services, but also into services in general.
For instance, once a washing machine gains an operating system and a Wi-Fi connection to the Internet, it can be monitored remotely and breakdowns can be predicted in advance. Or they could be prevented entirely, through a cleaning of the hardware or a modification to the software.
The administrator of such a service could also monitor usage, and could feed that data back to the manufacturer for use by product designers. Or it could be fed to retailers, who would know when it's time to pitch a replacement, or perhaps merely when it's time to buy more detergent.
Total Change of Focus
It's a total change of focus for the third party administrator, from managing the sale of the contract and then the break/fix episodes, to managing the relationships between the buyer, seller, and manufacturer of the product, and all the data that results from those relationships. Lakhani explains the concept with a Venn diagram consisting of four overlapping circles:
Harnessing Big Data
The Big Picture
"Traditionally, as a TPA or as a carrier, we have been focused on transactional-level information, or ERP information. So it's very one to one," Lakhani said. "Something's sold, and there's a warranty attached to it. And then we look at the data across the lifecycle, initially for purposes of sales reporting and actuarial analysis."
Then there was more data, and more reporting, and eventually more analysis. But it was still aimed at that central question: Is this the right price for that extended warranty, or are we losing money in the long run?
"We were just churning the same data with a different view," Lakhani said. "We weren't digging into it, asking, what does this illuminate? What does it tell us? We didn't get into major trending or observations."
Frequency and Severity
Then the next evolution in our industry came when the actuaries looked at the data, probing it in terms of frequency and severity, on the age of the product and the month in which the contract was sold. Then they took it to the next level.
"They began to ask," he said, "'What does this mean for mean time to failure, for how we provision partnerships with parts distributors, so that the logistics were most optimally configured, and so that our partners had as much of our data, so they could plan inventory, service capabilities, service level agreements, and scope of territorial coverage for parts and brands?'
"We started using that data in a more extensive fashion. It was still transactional, but it was helping us to design a better business model to manage costs, deliver services better, and even plan, in terms of expected losses over a particular lifecycle of that product."
In the diagram below, Lakhani explained, the industry remained inside the light blue circle, where the incoming data revolved primarily around transactional information such as products, prices, time, date, location, payment type, returns, cancellations, and other items. But it was all focused on the ERP or the transactional level information.
Harnessing Big Data
Step 1: ERP Data
"I think the opportunity for us, and the change that has to occur, is two-fold. First, the mindset has to change," Lakhani said. "We have to think differently about data. It's not just a simple transactional event, which is relevant just for us to process and deliver services. It is much more than that. It is a bigger commodity, it can be monetized, and it can be shared with more partners to yield even bigger and better decisions that affect not only merchandising, but also customer experiences, product reliability, and even pricing and product positioning.
"I think where we're going now is we've taken that information, and now we're going to partners and saying, here's how you can use this data from a merchandising point of view, because we're giving you that product lifecycle behavior. Here is what it's going to look like over a three- or a five-year lifecycle, so that you can go back to your OEMs and now, whether it be negotiating a better price, discounting, better service, better warranties, and better product reliability.
"We have to have the mindset that this data is meaningful, and we want to go after it. Because if we don't embrace that, then I think we're just a call center and a service provider," he said. "We haven't gone further in the evolution of our business. We're just doing what we were doing 20 years ago."
Harnessing Big Data
Step 2: CRM Data
The next step will cause service providers to cross the line from transactional data into three more types of data: customer relationship management, web, and collaborative data. At each stage, the data gets more complex and more difficult to interpret correctly.
"When you get into CRM, now you're looking at more customer information," Lakhani said. "So you're trying to understand the customer, not from just that singular event, but from a multiplicity of events, whether they be within your own website and your own channels -- be it brick-and-mortar point-of-sale, catalog, online, or browsing.
"All of these data points that you have, you need to start determining what is meaningful. You have massive amounts of data, so the other challenge we have is to crunch it into something that's meaningful. There's so much data. It's happening so fast. It's from varied sources, and some are valid but some are not. So what is of real value that we need to extract out of this and start measuring, so that it augments our understanding of an individual," he said.
Open to Sharing Data
While some of the administrators, carriers, retailers, and manufacturers will be open to sharing their data, others will not. Lakhani said the more reluctant entities will have to realize that they have to share their data in order to see the data of others, and that this sharing process will become the best way to accurately forecast costs and risks. And that, in turn, will lead to more realistic pricing. "Everyone is becoming more educated and more aware," he said.
The key will be the partnerships between data collectors, and how and when they exchange data. "You have to be much more open, and more willing to partner with all sorts of people in the domain, to gain access to their expertise and skills, and also to their information, so it can be integrated into a data value chain," he said.
"I'm not suggesting that all of this is easy. But it's a journey we have to go through: Outside of the retailer, how do we extract a better understanding? I think that puts us in position where we are challenged now to look beyond just our partnership with the retailer. We have to start partnering with other intermediaries in the data gathering and big data marketplace.
"Some people have web apps, whether they be product search, comparative analysis, or as a matter of fact they could be very specific to a particular product. And we need to start employing and partnering with all of them and their wealth of data, so that we bring more value -- we bring more information to the game," he said.
Harnessing Big Data
Step 3: Web Data
With the web, all the world's information is supposedly just a few clicks away. Consumers can use the web to research a product before they buy, and to rate the product after they buy. Other consumers follow behind, adding their paragraph to the discussion, supporting or refuting the opinions left earlier. With the right keywords and comments, consumers become researchers and critics.
Lakhani said TPAs and carriers are already seeing the results of this trend. "Increasingly, we have seen the consumer become king," he said. "More and more information is becoming available. We now have consumers who utilize the web, whether it be mobile or social, or just Internet, to garner information about products, product performance, and companies, before they undertake the purchase process. It's phenomenal how much consumers use technology to first explore before they buy."
The Price of Privacy
In addition, he said, while most of the privacy discussion now revolves around how much protection consumers deserve from the data collectors, Lakhani said this is increasingly going to turn to a question of how much that privacy is worth. "Just as much as enterprises have learned to monetize information, I believe that at some point consumers will learn that, and will want to control their data and be able to gain value from it," he said.
What can they get out of it? Most importantly, he said, they can gain control. "That's probably one of the biggest fears we have is that our data is roaming uncontrollably in the marketplace, and I think this consumer primacy is going to result in more of a sense of comfort and security in terms of what is done with my information. And I think this is going to create a mindset for the consumer that they have a commodity of value that can yield tangible monetary benefits."
Harnessing Big Data
Step 4: Collaborative Data
"This is the Internet of things," he said, "and things like RFID sensors. Our appliances are intelligent. So now they're giving us data. It could be at a personal level. When I know the frequency of a washer's usage, and its capacity load -- things that I can measure quantitatively -- I now know all the factors that can help me with respect to the durability and longevity of that product in that household. And, based on aggregated data on that model's performance, I can look at product reliability considerations to see the probability and the likelihood of what kind of failure.
"So now I can be proactive in my interventions. And preventative maintenance becomes a much bigger gain than just a reactive breakdown. That's where we can develop a whole new practice for third party administrators and others around how we can develop preventative maintenance programs. And that's all back into using predictive analytical models to start driving these intervening efforts into the game."
Manufacturers, Lakhani said, increasingly want to remain connected to their customers after a product sale. Keeping customers in their ecosystem ensures uniformity in brand experience, excellent customer service, and a relationship with the customer. All this yields significantly greater customer lifetime value for the manufacturer. Monitoring services are one way of providing that ongoing connection between the manufacturer and the consumer, even if there are intermediaries providing the actual monitoring services. Connected Living (inclusive of Home and Car) is an excellent case in point to deliver highly personalized, valued and actionable information, he suggested.
"It also creates opportunities for monetization, because now that I know your frequency of use, and what kind of needs you may have, I can determine your product usage," he said. For instance, if you buy enough detergent to do 50 loads, he can predict when you're likely to run out and when you might be amenable to receiving a discount offer for more.
Big Brother Is Watching You
The worry is that at some point, in the wrong hands, big data could become Big Brother. "The sinister side of this is you can use data to be extremely manipulative or exploitative," he noted. But not only will there be regulators providing a brake against that, but also consumers themselves will choose to do business elsewhere, if they feel .
The evolution, Lakhani said, is under way now. Extended warranties are becoming less about break/fix and more about ongoing service relationships. The resulting data can be used to make those relationships more valuable to both the service provider and to the consumer. But it won't be an easy journey, he added.
"It takes us out of our comfort range, or our normal range of activities, to challenge us to be more creative, and much more resourceful in harnessing information internally, and in a more collaborative fashion, to be more meaningful to that end user," he said.
Eventually, he said, warranty is going to have to become a more customizable product. "It's going to have to meet the needs of individuals whose perception is what calibrates the kind of warranty you're going to get," Lakhani said.
That kind of customizability is already offered with life insurance, personal automobile insurance, and homeowner's insurance, he noted. Specific coverages can be added or removed. Deductibles can go higher or lower. "But you can't do that in warranty," he said. At least not yet.