Quicksilver LB
Follow Us



Insurers Look Into The Eyes of Their Policyholders

The use of analytics to look into the hearts and minds of policyholders has provided insurers a tool that can change the way they look at their customers and their own business.

By Robert Regis Hyle

Insurers have different ways of going about the use of data analytics, but it is apparent that the future of insurance hinges on how well carriers use the data they have been collecting for decades and match it with third-party data to draw meaningful conclusions on risks, policyholders, and even agents.

Miguel Edwards, director of technology, Allstate Ivantage Select and a member of the Insurance Technology Association advisory board, believes analytics is changing the way insurers think about their business, particularly through the prevention of risks. If carriers think about coverage from a prevention aspect insurers can use data feeds from companies that digitize applications and cross reference against their own database.

“Is a customer using a licensed and bonded contractor for a construction project?” asks Edwards. “Is the contractor doing the work himself and need a builders’ risk policy? I know of a company that entered into a partnership with an alarm company. How is capturing alarm events impacting the protection they offer? Those are a few examples of openness. There are a handful of large carriers capitalizing on this, but what are the smaller carriers going to do?”

Novarica conducted a survey early in 2013 and asked insurers what they were doing with analytics. Martina Conlon, a partner with Novarica, reports that risk score was one of the top areas to identify the score around a property or a driver in order to determine if it is going to be a profitable risk for the insurer or whether there will be significant claims.

Another area for analytics involves claims, although Conlon expressed surprise that insurers are not using it as significantly there as they are in underwriting. Insurers are using claims fraud scores and claims severity scores, but the survey showed only about one-third of insurers are doing that.

“There really are big benefits [in claims scoring],” said Conlon. “Insurers are operationalized in terms of the core systems they are using as well as enabling rules engines or third-party services in order to execute the models. Risk scores and claims related scores are the most commonly used.”

Tools Improving

Analytics tools are getting better, according to Bryan Fowler, CIO of Oregon Mutual Insurance, but he hasn’t found them to be as intuitive as business users would like.

“Some are more user friendly for non-data scientists. Some aren’t insurer centric, but it does put power in the user’s hands,” he says

Fowler doesn’t believe insurers should focus strictly on industry-specific tools.

“If you need industry-specific tools you are trying to confine the uses of data by defining industry values,” he says. “It’s like big data vs. relational data bases. You spend months getting data exactly how you want it, but if you ask a question outside the boundaries, you are toast.”

For the last seven years, carriers utilizing modern core systems have been able to integrate with a predictive model that sits outside the system and returns a score, explains Conlon. A few carriers also have rules engines embedded that allow the insurer to develop a full predictive model inside.

“Carriers may put some basic underwriting rules in the system, but if they have a complex predictive model based on lots of different information and may refer to other databases, you do see them building them out into a separate system, which could be a rules engine,” she says. “Sometimes you see them develop predictive models in their rating systems, but it’s not as common given some of the complexity of the models. The system can support fairly complex models, but most carriers do develop them externally.”

Conlon believes the bigger issue is not necessarily the core system technology, but the challenge in coming up with a predictive model to answer questions such as: What are the steps and the rules? How do you know how to score?

“The vendors are stepping up in that space in the form of consulting services and actuarial services that are available for the mid-size carriers to lean on to develop models that are customized to their business,” she says.

Some core system providers are partnering with analytics firms to get the tools in the hands of the carriers. Others are buying traditional tools and embedding them, according to Karen Pauli, research director for CEB Tower Group.

“Vendors are all over the place. There’s not necessarily a trend,” she says. “The Guidewire situation is an important capability. (Guidewire bought the analytics provider Millbrook in 2013.) The core system is what comes first and the analytics comes later. I don’t think people are buying core systems for the analytic tools.”

Outside the Insurance World

Pauli believes there is much the industry can learn about practice, procedure, and attitude perspective from retail businesses.

“Retail uses analytics and models literally across everything,” says Pauli. “As an industry, insurers tend to use analytics on personal lines coverage, primarily auto, but there is no one in the retail business that would make that kind of a limiting decision. There’s something to be learned everywhere.”

Pauli believes a second area insurers need to study is what retailers do with a particular product or product line: They look at it and try to determine what can be done better.

“The insurance industry looks at [products] differently to figure out why something isn’t working,” she says. “There’s value in that, but the retail attitude is in optimizing things. Let’s get better. Retail looks at how to make things even better than they were last week or last month and insurers can learn a lot from that view of getting more out of the things that are working for us.”

Some insurers have hired from the retail industry to help change processes.

“We’ve gotten into thinking there are things we can’t do because of regulation,” she says. “Rather than figuring out how to make it work from a regulatory perspective, we are allowing the regulatory environment be the boundary of our horizon, which is the way it’s been done in the past.”

Other than the large direct writers on the P&C side, insurers are not fast followers in picking up most technology, admits Conlon, but most insurance organizations are mid-sized and small and aren’t leading the way as technology innovators.

“[Insurers] tend to follow other financial services and retail businesses,” she says. “There aren’t an awful lot of insurers doing interesting things with big data and so we looked at other industries. A lot of what those businesses are doing is around analytics. Most of the interesting work is related to big data.”

Retail businesses are taking analytics to the individual consumer level, points out Fowler. He recounts a story about how Walmart was looking at their data and noticed that before a major winter storm, many Walmart stores sell out of beer and strawberry Pop-Tarts. So, as storms show up on the radar, they start shipping more beer and strawberry Pop-Tarts. As for insurers, Fowler believes carriers need to see where trends are starting to emerge.

“What kinds of quotes and coverages are being asked for?” asks Fowler “We might be maintaining four or five coverages that no one ever uses, but if I can look at my data I can start to learn what consumers are asking for through their agents.”

By looking at data and transactions as a way to see what people are doing, you can learn what is going on with cross products or in certain geographies and take action on that, explains Fowler.

“I can call an agent and tell them things they might want to know, such as a spike in quotes,” he says.” Maybe it’s not occurring within that agency, but within a 300-mile radius. How can we help them take advantage of that? We can mimic what retail is doing to manage its sales channels. It can make them more successful and make us more successful.”

Edwards maintains that other industries have developed technology-driven strategies, but that is not always the case with insurers.

“The CEO of a bank didn’t just say, 'We need to invest $10 million in a mobile app," there was a technologist that brought the idea to the CEO,” he says. “In this context—and this is where there is a big lapse in the insurance space—we need insurance CIOs to come to the table as business partners. When the CIO has that opportunity, they are going to explain to the CEO that in order to take this company to the next level the IT department needs X amount of dollars to build a data science group to take the business into the next century.”

In insurance, IT departments are order takers and simply respond to what the business side is asking for, according to Edwards

“The business is mired in the day to day stuff because insurance is such a process intensive industry with issues such as rate changes and form filings. "That is what the CIO is forced to focus on,” he says. “If it were up to me, I would make every insurance organization take a closer look at their IT groups. It’s not just business intelligence where help is needed. We’re bad with our use of mobility and social media.”

Edwards believes analytics needs to be used more in sales to predict the next logical sale or product. He cites the book, Competing on Analytics: The New Science of Winning, by Thomas H. Davenport and Jeanne G. Harris, where the authors discuss the success of Netflix, which has developed the capabilioty to suggest the next logical movie for a given customer to keep that chain going based off of data.

“Why wouldn’t an insurer apply that same logic to the way they market and sell to their customers?” Edwards asks. “Insurers need to understand characteristics around their customers' buying habits. You can buy that data or you may already have that data. If you understand characteristics around a risk profile, why wouldn’t you use that information to determine if this customer was a good candidate for another insurance product.”

Mid-tier Users

At CEB Tower Group’s insurance roundtable last fall, Pauli reports that half the people in the room were in roles related to analytics that didn’t exist within their organization 12 to 18 months ago.

“Everybody is trying to figure out the best organizational structure and it is a hard thing to do,” she says. “It will be evolving for a long time. When you look at top 10 carriers, having an enterprise-wide anything is difficult. The thing I think is neat about the mid-tier is the technology providers serving the mid-tier are bringing a common tool across commercial and personal lines.

Analytics definitely is moving down to the mid-tier, agrees Conlon, as they see the benefits big organizations are getting, particularly the direct writers that are able to score claims.

“The mid-tier is just slightly behind because they have limited resources,” she says. “Personnel is part of the issue. Many mid-sized and smaller companies we talk with that have invested in this have a gap in analytics skill sets. That’s where vendors are stepping up to the plate right now. They are great partners for insurers as well as regional actuarial firms are helping smaller carriers get there through predictive models.”

CEB TowerGroup surveyed technology providers to provide a roadmap of the functionalities they will be adding over the next 12 months and all but the smallest vendors are adding predictive analytics in either six to 12 months or 12 to 24 months windows.

“Vendors are bringing competency to the mid-tier folks and that is important,” says Pauli. “The mid-tier insurers have a shorter buying cycle. Once they get [technology] they really get it. You don’t have the political stuff that comes with gigantic insurers. Bringing analytics that to them is really wonderful.”

The bigger issue often comes down to the talent.

“I’m not sure it is necessary to ask our data those deep questions,” he says. “We don’t have to out-Progressive Progressive, we have to out-Progressive Mutual of Enumclaw.”

Nearly two years ago, Oregon Mutual rolled out a system for its independent agents. Specific data was placed on the iPads so marketing people could have a meeting with the agents to show them how things are going in trends and other areas.

“That information can show agents where the carrier can help them,” says Fowler. “That was a big cultural shift for our agents. If you are a hunter rather than a gatherer that kind of information is invaluable. It is data that is readily available and we can crunch to make it relevant for a conversation.”

Edwards thinks there is contention within carriers with respect to using resources and many are challenged to justify spending money on acquiring and using data.

“The CIO of a mutual might have a 45-year-old policy admin system that needs to be replaced. Is he thinking about that issue or trying to find a place for all the data that he doesn’t have any room to store?” asks Edwards. “Insurers delayed the inevitable in terms of what needed to be done with their systems for so long that they got caught in this trap. They can’t take care of the foundational stuff so they can move into the more innovative stuff.”

The delays get longer because it takes so much time to seek out and find the right platform.

“How long does it take for an insurance company to settle on a core system vendor? You have to hire a consultant and spin off all these resources. The process has become so challenging and time-consuming,” says Edwards.

In 2014, Pauli believes the big data streams will become more normal streams. Instead of Big Data being a project, she believes it will become part of the mainstream data.

“It’s getting easier in part because of technology,” she says. “Cloud technology allows us to leverage technology without the cost of infrastructure projects. That makes it easier and more manageable.”

The Future

The data is going to get better, insists Fowler, and will allow carriers to look at each part of their business—underwriting, claims, marketing—with a holistic view of all the actions and interactions with agents and customers.

“You will find interactions you never knew about or interactions you are missing,” says Fowler. “I guarantee those are out there. I’ll take that to the business unit level and they’ll get amazing results. How to profile the book of business and map that to what we already know about. We can profile a risk or profile the set of activities that underwriters should go through to minimize the time it takes to get the best possible quote out there.”

A second focus for the future is telematics and what is going to happen going forward with the different technologies and data capabilities.

“The type of data that is going to be thrust upon us without even asking for it is going to be something we have to be ready for,” says Fowler. “I wish it was top of mind, but at least it’s in the back of my mind because it is going to be a whole set of issue and you can’t let it pass you by.”

Telematics data can help customers steer clear of situations that might be risks for them and for the insurer.

“We price insurance based on the past, but telematics brings the past so much closer to us,” says Fowler. “It’s not four or five months in the past, but four or five minutes.”

Going forward, Conlon sees more insurers using risk- and claims-related scores but also using different scores for determining whether they need to do a premium audit to determine the likelihood of a positive action. Examples of this involve sending a representative onsite as opposed to accepting a report over the phone.

Additionally, carriers are extending predictive models to determine which claims are candidates for subrogation or litigation—and if they are what claims are prospects for settlement. The other area Novarica has found involves insurers just starting to use predictive models to manage workflow and task assignments to become a more efficient organization, such as using models that score a claim to determine which adjuster should adjudicate the claim, using a risk score to determine which adjuster to assign it to, or to determine the workflow for that particular submission.

“It’s not just quantifying characteristics of the risks or claims, but using models to manage the work process and the assignment of tasks and hopefully contributing to better underwriting and more efficient internal operations,” says Conlon.


“I’m sure you’ve heard the cliché that insurance is data rich and information poor,” says Edwards. “We have all this data yet many carriers can’t make heads or tails of it.”

Edwards does not believe it is a technology problem, though. One of the issues he feels insurers should be concerned with is finding the data scientists--the people who think about data and information and try to find ways to tease out little bits of information.

“One of the perils of a business intelligence project is you put the project together and at the end of the day all you have are reports,” he says. “It became a reporting project. What about the questions around decisions that are needed to drive the business.  When you ask the questions you need all the data points and then it becomes a technology issue to assemble the data points. So often we lack people on the business side to ask questions and look for new data to evaluate the most effective agents and why they are effective. Reporting can tell you who the effective agents are--you can look at hit ratios and retention ratios--but what is it about their business that makes them successful. Someone on the business side has to ask that question. It can’t be driven by IT.”



Featured articles


Clear Lync MR


The Email Chat is a regular feature of the ITA Pro magazine and website. We send a series of questions to an insurance IT leader in search of thought-provoking responses on important issues facing the insurance industry.


April 5th – 7th, 2020
The Diplomat Resort
Hollywood, FL
Become a member today to receive updates – www.itapro.org/MR


only online

Only Online Archive

ITA Pro Buyers' Guide

Vendor Views

Partner News