Trusting your technology partner is important in any softwaredeal. However, that trust is even more critical in predictiveanalytics solutions because carriers are placing their rating andunderwriting future--the lifeblood of the company--into the handsof a third party. Insurers must have faith in their vendor and needto connect with their partner on a cultural level, according toKaren Pauli, research director in the insurance practice atTowerGroup. "Does the vendor really understand what you want to bedoing?" she asks. "You need to look for someone with priorexperience in the same market segment you are in and someone tohelp walk you through the process and give you some businessexperience."

Island Insurance Companies is just getting started in theprocess of predictive analytics, relates Jeff Fabry, vice presidentand CIO. The carrier has been working on a personal auto projectwith its vendor, Fair Isaac. The situation Island finds itself inis it has three different companies that write personal autocoverage in Hawaii, and each one has a single rate, explains Fabry.When competing with national carriers, Island finds the competitionhas a much wider range of rating possibilities. "Right now, we haveonly three slots, and we're trying to increase that to six or 10slots so people who have worse loss ratios have a little higherrate and the people with better loss ratios get a better rate," hesays. "You are able to be more competitive on the betterrisks."

Pinnacol Assurance performed extensive due diligence beforeselecting Valen Technologies to perform its predictive analytics.One of the steps in that process involved sharing confidential datawith the vendor to see what the technology and modelingcapabilities could do for Pinnacol from a business perspective."That was crucial for us to test drive predictive analytics becauseit was all new to us," says Mark Isakson, associate vice presidentand pricing committee chair for Pinnacol. He admits there weredifficult moments in the process. "At the end of the day, becauseit's so new, how do you get your arms around [analytics] and absorbit before you jump all the way in? It can be a scary thing,especially for an insurance company," he says, adding Pinnacol sawthe possibilities during the proof-of-concept process.

Building the models was where Isakson believes the relationshipwith Valen was most important because the development leveragedPinnacol's expertise in workers' compensation. Since the vendor wasbuilding a new application, it needed to work with the carrier'sexisting applications and processes. "It was a fairly extensiveengagement--about nine months," recalls Isakson. "We provided Valenwith our data, and it would go back, do the analysis and modeling,and show us what it had. We kept refining the approach as Valen wasdoing the analysis, and once we had what we thought was a prettygood model--something that would work with our business model--wesat down and basically went through a testing and validatingprocess."

Pinnacol ran the model back in time against data from prioryears and evaluated how it performed against what the carrier didwith its older approach to assess what the performance of the modelwould be like in a simulated production environment. Even moreimportant, according to Isakson, the carrier wanted to see whatsome of the challenges would be in the implementation phase thatwere going to change dramatically the way Pinnacol did business."We needed to go through the change management process so we couldhave communication strategies built and establish expectations,particularly with our underwriters, who would be using the model toprice," says Isakson.

The change management piece is crucial when introducing a newpricing model and the underwriters are accustomed to what they hadbeen doing for years, Isakson believes. "We would learn only byseeing it in a production environment rather than learning it in alive environment," he says.

The pleasant payback of the whole project was the model led tomore predictable outcomes for Pinnacol because the carrier had sucha huge amount of involvement in customizing the model to itsbusiness needs, culture, and philosophy. "Most of the things welearned in the process were things we anticipated along the way,"says Isakson. "I don't know that would have been the case if youpurchased a box product off the shelf."

Another factor in selecting the right partner is finding onethat will give you the tools to allow you to run strategies, notesPauli. Carriers need to avoid coming up with a model that willvalidate old assumptions. Pauli points out insurers need to seewhat the data tells them and then use the analytics tool to runstrategies and discover the outcome. "Not all vendors have thetools to run strategies, so that's a real key for any company tolook at--can you do simulations of strategies so you know what theresult will be?" she asks.

The partnership between Island and Fair Isaac is vital becausethe vendor is doing most of the modeling. "It is important tochoose a vendor with a lot of experience and that has worked withother carriers," says Fabry. "We knew who the big players were, soit wasn't that difficult a decision. For us, it came down tocost-effectiveness and things like that. Fair Isaac has a prettygood track record."

Carriers need to have a marketing and an underwriting objectivein mind when analyzing their data, advises Pauli. In her priorinsurance company experience, the first time Pauli ever did amodel, the data suggested to her company that people who chosehigher limits and added endorsements or additional coverages totheir policies were the carrier's most profitable customers. "Wewouldn't have necessarily known that, but that's what bubbled upfrom the model," she says. "To some degree, the data will suggestthings to you, but you need to have an objective in mind as to whatyou want the ultimate outcome to be."

Pauli has found discussions between carriers and stateregulators have become common for insurers entering the world ofpredictive analytics. Regulators find what she calls "black boxmodeling" to be distrustful. "[Regulators] can't always tell whatthe carriers are doing in their underwriting. What the regulatorsare suspicious of is underwriting decisions that discriminateagainst people of lower income or on race," she says. "If you aregoing to come up with an underwriting or pricing model, it doesn'thurt to work with the regulator so it can see what you aredoing."

Currently, Island and Fair Isaac are going through the carrier'slast five years of data and are looking for different combinationsof metrics. This is where state regulations can complicate things.Fabry indicates in personal auto underwriting, age of driver is abig factor in many states, but in Hawaii, where Island does itsbusiness, that information can't be used to determine rates. "So,we have to look at other ways to determine the proper loss ratios,"says Fabry. "We're looking at other types of metrics."

Fair Isaac pores through Island's data and examines the losshistories for different metrics. "Some pan out, and some don't,"says Fabry. "Is it location? Is it how many miles you drive towork? A lot of companies on the mainland use credit rating, butwe're prohibited from using that here. It's the same philosophy; wejust have to look at it in a different way."

Island has determined it will not push the envelope with itsstate regulators. "We're not going to do anything that hasn'talready been done here in Hawaii," remarks Fabry. "We pulled thestate filings from [two competitors], and we're not going to doanything other people aren't doing in the state so we don't runinto any issues with regulators."

Analytics has been around the insurance industry for a while,but Isakson believes it is fairly new to workers' compensation,which, he claims, is highly regulated in each state.

One of Pinnacol's major concerns was, with the significantinvestment of resources, whether the models would be palatable toregulatory agents in Colorado, where the carrier operates. "Wewanted to see where the state stood on predictive modeling to seewhat was acceptable and what was not acceptable," he says. "A hugecomponent to success or failure is having a conversation with yourregulators. The regulators may not be comfortable in dealing with[modeling], and your investment [in the solution] already hasbegun."

Pinnacol explained to the regulators what the power ofpredictive analytics could do in its business and where the carriersaw the value. Isakson reports there were two major concerns to theinsurance department. "We needed to avoid factors in the predictivemodel on the pricing side that were either redundant with otherpricing elements already in place or discriminatory factors thatmight be used--such as redlining," he says.

Predictive analytics is not a once-and-done solution forinsurers. "They really do have to stay on top of it," states Pauli.The downfall for some carriers is to put the model in place andthen infrequently refresh the data in the model. "That's a mistakebecause the advantage to modeling is to keep refreshing the data sobusiness analysts can fine-tune pricing and underwriting and not dosomething cataclysmic," she says.

Depending on the size of the carrier, that refreshing could beonce a week or even once a month. If it's a big-volume carrier, therefreshing of the data literally could take place on a daily basis,continues Pauli. "Doing that much refreshing would call for givingthe business analysts a dashboard to look into the data to see whenthey need to start adjusting underwriting or rating based onchanges in the marketplace," she says.

The models don't have to change as often, adds Pauli, butcarriers ought to revisit them yearly or every 18 months, dependingon the number of transactions. "If the models are working and youare transforming your business, somehow you need to see what thattransformation is doing to your results and where you need tofine-tune," she says.

Fair Isaac goes through Island's data and develops a model, thecarrier implements the model into its rating, and then both teamshave to stay on top of the data to maintain that model. "You haveto run reports constantly," says Fabry. "Are our results indicativeof what we predicted they were going to be? Is this really workingout the way we thought it would? If not, we may have to do sometweaks here and there, but that really requires a lot of dataanalysis."

In the past, Island relied on outside actuaries for thatanalysis, but the carrier recently hired an in-house actuary, whois going to be taking over that job. "He'll interface with FairIsaac, do the data analysis, crunch the numbers, and work with FairIsaac to tweak the model," says Fabry.

It is difficult to find business analysts who have the skills tocome up with new business strategies and the ability to look at theinformation and see what it says, according to Pauli. "It's afairly new area," she says. "The business analysts of the futurewill have to be half underwriter and half actuary. Someorganizations have done a good job with that, and some are stillstruggling. Others are using the services of the vendor they areworking with, and that could be a good choice for a carrier."

Part of Pinnacol's plans over the next several years is anannual refresh of the model. "It's an ongoing process of monitoringand using Valen's abilities--Valen is the one with the technologyexpertise--and we're relying on it to stay in front of trends,"says Isakson. "We also internally will go back through the entirevalidation process now that we've had six to eight months ofexperience with the model. We're going to do a more thorough reviewbefore setting up our pricing for 2009, and if there are changes inour underlying business we may see, we need to react--hopefullymore proactively."