The last few years have seen the use of predictive analytics move from a niche technology employed sparingly by mathematically adept power users in certain operational sectors to a common tool employed by business users throughout the enterprise. Insurance Networking News asked Steven Callahan, practice director at Robert E. Nolan Co., how insurers are faring at employing predictive tools at the operational level.
INN: What are the new trends in analytics?
SC: An extremely interesting trend is the integration of external data into company data to enhance predictive capabilities. Consider the aggregators that are mining data about individuals from their social network pages, websites, rewards programs and various other sources of information about individuals. By adding this data into the analytics repositories, correlations to purchasing patterns, eating habits, hobbies or any number of other variables are integrated to determine predictive patterns.
Already, the use of social network mining is becoming a common practice in the claims fraud arena as investigations incorporate specialists to harvest the information. Add in grocery store discount programs, credit card purchases, magazine subscriptions, etc., and a rich array of data can be consolidated with company data to generate predictive profiles, define market segments, screen agents, determine lifetime value of a customer set and other value propositions.
INN: What management challenges does wider adoption of analytics present?
SC: The first challenge is achieving an executive-level understanding of what analytics means to their business. Surveys indicate a gap in seeing the potential ROI of analytics. An IDC study indicated a 145-percent median ROI on predictive analytic projects and an 89-percent median ROI on business intelligence projects. The results have to be put in the context of a company's strategic priorities.
Second, organizational silos have to be organized to represent a consolidated view. This requires overcoming cultural resistance to change and territorial ownership of data. Gaining the buy-in and participation needed requires strong executive sponsorship and persistent leadership.
Third is establishing a governance structure. Today, solutions are fragmented and heavily dependent upon IT. An overriding oversight is necessary in order to arbitrate conflicting approaches, maintain consistency of definitions and methodologies, and ensure integrity and quality of the integrated data.
INN: What are some common barriers to implementing advanced analytical systems and solutions?
SC: The most common barrier is overcoming the challenge of organizing and validating the fragmented data that exists across disparate systems. Effective analytics requires a holistic view of data throughout time to be able to bring the predictive and correlating values to the forefront. Many organizations use this as a reason not to invest in analytics until systems are replaced and the data scrubbed, despite significant advances made in programmatic data reconciliation and recognition tools. In addition, advanced analytic programs have a greater tolerance for error, reducing the need for complete data scrubbing and integrity checking.
The other most common barrier is resource constraints. Technology resources are fully engaged in long lists of tier 1 and tier 2 program priorities, and there is a lack of business resources with the necessary expertise. Complete outsourcing is difficult given each company's unique data sources, structures and definitions. This requires business experts and technical resources in order to implement a solution. Many companies solve this gap by contracting analytics expertise to collaboratively leverage company expertise, taking fewer resources than a dedicated team.
INN: What other challenges might analytics bring that have not been fully considered?
SC: As companies identify discrete risks within their portfolios, regulators may require them to disclose this to consumers. What impact will this transparency have on consumer perception of company stability? Will rating agencies define new categories for more discrete risk profiles defined by investment analytics?
The convergence of technological advances in analytics tools and an increasingly rich supply of consumer data bring both opportunities and challenges. It should be evident that this transition will be debated heavily in the public and regulatory arena-yet it is already underway. Finding the balance between legitimate uses of predictive capabilities against a data rich world, the privacy rights of individuals, and the inevitable need for greater transparency will bring greater complexity and a trial-and-error learning process.
