Insurers have always sought to analyze customer metrics. Understanding the potential for loss or damage and the propensity to claim are fundamental to risk management. In recent years, however, modern software, modeling and data analysis techniques have transformed the landscape. The most successful insurers are those who fully exploit these cutting-edge capabilities.
Like many in the financial services sector, insurance companies face a major challenge in generating profitable and sustainable growth, especially in current economic circumstances.
However, insurers often face potential costs stretching far into the future. This means that sustainability and profitability depend on identifying the specific characteristics of policyholders that contribute to differential loss potentials as well as developing lasting and deep relationships with clients. Modeling and analysis of customer data are, therefore, fundamental to a successful business model.
Insurers have access to vast amounts of data and associated statistics, particularly historical data on customers, policies, claims, etc. In addition, increasing amounts of collateral information are available on many aspects of customer behavior and experience. This is especially true for insurers owned by banks.
Whereas dedicated insurers may have relatively few interactions with customers between an initial sale and a subsequent claim, a parent bank may have access to daily information about a client’s lifestyle, behavior and habits. Third-party data sources, such as credit rating agencies and consumer research companies, can also deliver substantial quantities of potentially useful data at comparatively little cost. When this information is properly harvested and analyzed in a responsible and customer-centric manner, it can unlock enormous potential within an insurer’s existing and new customer footprint.
The critical issue, therefore, is not the availability of data, but how to analyze it effectively and apply the results to the business.
Much more sophisticated software has become available in recent years. Deployed hand-in-hand with more powerful computer hardware, this allows companies to routinely undertake analyses that would have been prohibitively time consuming or expensive only a few years ago. Hard figures (e.g. costs, claims, dates, times and other customer details) can be trawled automatically to extract significant correlations that can yield valuable insights into core business parameters.
In addition, dramatic developments in the analysis of unstructured data, or text mining, can allow large volumes of text based material held in-house to be scanned to extract potentially significant information. Furthermore, publicly available material can complement existing consumer behavior data. Internet technology and social media now allow customers and potential customers to chronicle their lifestyles and consumer experiences online. Many people also use social media networks to evaluate products, services and providers. It is well known that certain key life events, such as buying a house, having a baby or retiring, are associated with insurance purchases. Increasing numbers of analysts and modelers are experienced in the necessary actuarial and statistical analysis. Identifying these key life events can make targeted marketing much more effective.
The critical issue, therefore, is not the availability of the data but how to analyze it effectively and apply the results to the business.
The key benefit of these new capabilities is that they can improve the success rate of prediction. This applies not simply to the probability of an eventual claim, but by extension to all other stages in the business cycle, including targeted marketing, customer segmentation, policy risk profiling and fraud detection. Predictive analytics allows insurers to understand their customers in new and deeper ways, both in aggregate and as individuals or members of more closely defined segments. Personalized and tailored, this can be developed to satisfy client needs in deeper and a more meaningful way. While this type of analysis was once typically the domain of large personal line (consumer) insurers, the changes we are concerned with enable these techniques to apply equally to commercial, specialty, life and health insurance.
Finer and more discriminating customer and claim segmentation brings additional benefits to many aspects of insurers’ business processes. For example:
Predictive analytics also supports the development of complementary perspectives on the business. Where insurers have typically focused on risk, its quantification and management, effective modeling and prediction can facilitate a more customer focused approach. And it is clear that the more that is known about customers, their needs, behaviors and predilections, the easier it is to deepen an insurer’s relationship with his or her customers and sell more products to them over a longer period.
A similar approach can be applied to improving distribution channels. Predictive analytics can help attract and retain the right sales and marketing people as well as the most effective distribution partners. By casting light on how much value particular individuals or channels are actually generating, it allows for the fine-tuning of rewards and incentives or, alternatively, helps identify undesirable behavior.