Carrier Management Magazine - Winter 2014

Page 22

Technology & Analytics continued from page 21

Big Data Sources

New data is proliferating for insurers to tap into, according to Martina Conlon, principal for Novarica. Some types of big data available to insurers include: • Credit scores and details • Consumer and business data • Social networking analysis • Geo-spatial information • Weather data • Satellite photos • Purchasing history • Health and prescription data • Industry specific policy and claims data • Competitive filings • Public records (properties, competitive filings, lawsuits, crime rates, census data, etc.) • News services During a presentation at the SAP Financial Services conference in September 2013, Conlon noted that Novarica uncovered more than 90 vendors offering different types of information for insurers when conducting some research two years ago. better decisions in agent appointment, management and compensation with insights gleaned from big data analysis. Demographics, business data, population and business growth statistics, purchasing data, and public property records can be

analyzed to determine the potential risk, demand and profitability for the location the agent services. Likewise, personal lines writers that are trending toward direct sales are using big data to better engage prospects. Web usage,

social media and third-party data can be used to anticipate life events or purchases that accurately predict insurance needs, such as auto comparison and real estate sites, thereby enabling early prospect engagement. After the sale, consumer, claims, location and risk data can be leveraged to design loyalty and retention programs that deliver the right information to policyholders at the right time via the preferred communication channel. Finally, big data types can prove invaluable in basic risk assessment and pricing. Consider the ocean marine insurer that covers container ships where underwriters commonly price accounts based on historical pricing and loss history. What if the underwriter also has access to the weather conditions and piracy activity along the route, or crime and political statistics for the ports of call? With the competitive pressures in commercial lines today, price accuracy can have a big impact on the close ratio of an insurer. Big data can provide enhanced perspective to help ensure optimized pricing, and big data analysis can uncover dependencies and insights to define better

Netflix’s ‘House of Cards’ and Other Big Data Case Studies

H

ow did program developers at Netflix know that viewers of the BBC television series “House of Cards” also enjoyed movies starring Kevin Spacey and directed by David Finch? They analyzed big data to identify the overlaps, Martina Conlon reported during a presentation at the SAP Financial Services conference last year, going on to describe more of the big data insights Netflix gathered to produce its own online remake of “House of Cards.” Conlon said that examples of big data applications are harder to come by in the insurance industry. “The insurance industry is a little bit slow on the uptake except for the very, very large organizations in both commercial and personal lines.” In

22 | Winter 2014

addition, “there aren’t many insurers willing to share their successes. They view it as their special sauce,” she said, explaining why she looked outside the industry for big data case studies. Continuing her Netflix example, she noted that behavioral analysis of the big data gathered on viewing habits revealed that the likely viewers of “House of Cards” are addicted to viewing marathons—a data tidbit that prompted Netflix to release all the episodes at once. Conlon also drew an example of an underwriting application of big data from outside the insurance industry, noting that Zest Finance, a small consumer loan company, used big data to determine that the way in which a loan applicant fills out the application really matters. All small letters or all capitals may be bad, while

those using mixed letters may have a lower likelihood of default. Within the insurance industry, Climate Corp. (discussed in the accompanying article) provided Conlon’s only example of using big data for product development, but she offered two examples related to underwriting and risk assessment—one from American International Group and the other from Florida-based commercial lines writer FCCI. AIG, she reported, bought all the data on lawsuits against executives across the country since 1996 and analyzed it to find six predictive factors of the likelihood that an account in the executive liability line will be unprofitable. At FCCI, predictive models developed from data mining challenged the existing assumptions of underwriters and revealed

www.carriermanagement.com


Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.