The Path to Analytics is Paved with Good Data
Carriers are implementing data warehousing and master data management initiatives to build a better foundation for analytics.
Insurance Networking News, 07/01/2010
Insurers face a deluge of datacoming in from all directions that, if properly mined, could provide valuable insights about customers and product preferences. However, in many cases, the business structures of insurers-based on multiple business lines-do not make it easy to get at this data and turn it onto actionable information. "Many insurance companies run individual insurance products as fiefdoms, as individual divisions," says Claudia Imhoff, president of Boulder, Colo.-based Intelligent Solutions, Inc. "You have auto insurance, home insurance and everything else, all run in separate little companies."
However, in today's hyper-competitive business environment, insurers can no longer afford to keep data locked up within these silos. As insurers strive to compete on analytics derived from their data, the quality and ability to integrate that data becomes critical. "There are fewer physical connections with customers-it's more an electronic relationship," says Paul Luongo, president and CIO of Shared Technology Services Group, Inc., the IT service company for the Plymouth Rock Group of Companies, based in Boston. "We need to know as much as we can about that person and what their needs might be."
Advertisement
Many insurers are discovering the additional market power their data may deliver-for better customer insights, enhanced product segmentation, and fraud detection. For example, Private Client Group of New York-based Chartis recently took a hard look at its existing homeowner policy data stores, and was able to put together an algorithm to better predict the likelihood that certain properties will have a higher propensity of claims. "We developed a capability to better understand our catastrophe peril exposure by geography and construction quality," says Karl Uphoff, SVP and CIO for Chartis. "This assists underwriters with impact analysis on how a new piece of business would affect our existing book."
WHY DATA QUALITY MATTERS
However, for many insurers, an ability to get to the right data-or poor data quality-is hampering their ability to achieve such capabilities. "The implication on data quality is significant," explains Perry Rotella, SVP and CIO for Jersey City, N.J.-based Verisk Analytics. "More and more, insurers are using data in advanced analytics and predictive modeling for competitive advantage and better loss control. As analytics become even more sophisticated, the ability to correlate seemingly unrelated data elements across a multitude of sources will create better business solutions. In turn, a strong data management function is required because quality data is core to an organization's ability to apply analytics and predictive modeling."
However, little support from management seems to be forthcoming. A survey of 193 executives from a variety of industries by London-based consultancy The Information Difference Ltd. found that one third of respondents rate their data quality as poor at best and only 4% considered their data to be of excellent quality. What's more, 63% of respondents have not attempted to calculate the cost of data errors, and only about 37% currently have some form of data quality initiative in place. Information Difference also found that the top two barriers to achieving quality data were the fact that management does not see data quality as an imperative, and difficulties presenting business cases.
Forward-looking insurers recognize the urgency of effective data management and governance for delivering strong enterprise analytic capabilities. In the absence of quality data, insurers face a number of challenges. "Incomplete data could lead to higher costs for insurance organizations," says Pankaj Sinha, practice lead for Tata Consultancy Services' Insurance Information Management and Technology Excellence Group. Bad data leads directly to "adjustment of higher claims, setting lower premiums for customers, and lack of proper risk identification. Inaccurate data could lead to an adverse impact on downstream business processes and a ripple effect on operations across the organization. For example, capturing inaccurate customer information leads to bad decisions on policy levels, terms, premiums, loss of additional selling opportunities and wrong customer preferences."
Insurers are taking a number of paths to tackle this problem. One course is through more effective governance, along with initiatives that take data out of the silos and makes it an enterprise resource, such as the related approaches of data warehousing and master data management (MDM).
Enterprise-level data governance is an important priority at Columbus, Ga.-based Aflac, which has been building up the capabilities of its data warehouse environment. "Over time, our overall data volume continues to go up as our business continues to grow and succeed," says John Keddy, VP of IT application services for Aflac. "We've been making some modest investments in our enterprise data warehouse, and those investments have been very well received by our organization." For example, Keddy says, his department has been promoting greater self-service for business users querying information from the data warehouse. "So far in 2010, my department has completed 300 user queries, but at the same point in time, our users have successfully run 300,000 data warehouse queries on their own."
