Challenging economic conditions, reduced returns on investment, in-creases in large claims, and increased competition have all combined to drive approximately one-third of U.S. insurers out of the market over the last 15 years. Reductions in top-line revenue concurrent with shrinking profitability have had their impact. Today, fewer, generally speaking larger, and definitely more intensely focused insurers are vying for a larger share of an increasingly well-informed and demanding market, one that has shifted demographically, behaviorally, and intellectually.
It has been said that the insurance industry is rich in data and poor in information. Given the constantly evolving market dynamics and competitive landscape, a strategic imperative for surviving is the translation of this wealth of internal and externally available data into meaningful decision-making criteria. Being a services industry, insurers know profitability rests on understanding consumer expectations, effectively delivering products that meet these expectations, efficiently servicing the resulting customers in a manner consistent with their desired methods and timelines, and profitably managing the risk pools.
While many companies have attempted to cut themselves to profitability by way of expense reductions, the real leverage in this industry is on the marketing—or revenue—and claims sides. Unlike the third major component, underwriting, which has long based its activities upon predicting future events based on multiple discrete variables, both marketing and claims are recent entrants into the predictive world of analytics. For the most part, the focus with these areas has been a more retrospective trending and projection process until recently, when the more advanced companies have realized the potential and started investing in the predictive aspects of analytics.
Customer segmentation and differentiation, selective lead generation, agent productivity, and lifetime value of customers all represent areas of rapid progress within the marketing world of predictive analytics, generating revenue from better understanding of how to profitably grow specific market shares. While these opportunities are significant and worthwhile, it can be argued that a quicker path to profitability rests with operationalizing improvements in claims practices and processes.
By leveraging the wealth of information available on the claims front, even relatively small one percent improvements in losses can result in significant increases in profitability as well as releasing reserves for improved investment returns. These improvements are achievable on existing business with relatively quick time to implementation. Given market and shareholder pressures, the opportunity seems both timely and material. Companies would be well served to invest in improving the use of analytics within their claims operations in order to realize these benefits.
How are predictive analytics any different from the long-standing and well-developed practices of claims analysis? In effect, predictive analytics bring the added dimension of adaptive learning from experience to the modeling process, which contrasts to the more traditional trending or univariate analyses. Instead of using linear expressions of probability based upon reviewing historical data to specify risk identifying parameters, properly constructed predictive models massage the data in search of generalizations that identify common patterns associated with higher risks. The patterns are discovered usually via a process known as data mining, which is a technique for working with broad descriptive attributes of a claim to find correlations.
Here is where predictive models differentiate themselves the most, by continuously undergoing the equivalent of a learning process that integrates newly found correlations regardless of whether or not there is a causal connection. For example, the review of a large pool of claims may indicate that a specific and narrow age range at a particular ZIP code has a higher propensity to litigate even the smallest of claims. This information can be used as a pricing consideration, to determine initial reserves, as well as fine-tune the process so that these claims are fast-tracked for legal review. The relationship between age, ZIP code, and risk, or the correlation, is integrated into the model even though being that age and in that ZIP code may not necessarily cause higher litigation—there is a relationship even if it is not causal.
By leveraging advanced mathematical techniques like logistics regression analysis, decision trees, Naïve Bayes, and neural networks, predictive models are able to handle a much larger number of variables. They are able to transform seemingly random pieces of data into decision criteria that aids in identifying how to best distribute and process a new claim differently than would otherwise have been done. As a result, the overall business value of successfully implementing predictive analytics in a claims operation is wide-ranging. Some of the most common benefits include:
• More accurately priced risks at a finer level of distinction;
• Closer management of case reserves, freeing up capital and increasing investment income;
• Lower claims-handling, supplier, and investigation costs;
• Less claims leakage;
• More extensive insights into claims unit performance;
• Greater claims processing efficiency, especially in the areas of predicted fraud, subrogation, and litigation;
• Faster transaction cycle time resulting in improved levels of customer satisfaction.
Consistent with this wide range of possible benefits, recent surveys of business and technical staff have indicated that ROI on predictive analytics projects exceed 20 percent on over 75 percent of the best projects and exceed 20 percent on over 30 percent of the worst projects, and in all cases less than 10 percent of respondents claiming no ROI. According to an IDC report, the median ROI for predictive analytics initiatives is 145 percent compared to an ROI median of 89 percent for non-predictive business intelligence initiatives. Similarly, according to a white paper by the Aberdeen Group entitled “Predictive Analytics: The Right Tool for Tough Times””
“Users of predictive analytics…have achieved a one percent improvement in operating profit margins over the last year, and a year over year increase in customer retention of six percent. Survey respondents that have not yet adopted predictive technologies experienced a two percent decline in profit margins, and a one percent drop in their customer retention rate.”
The value to a company, especially in today’s market, should be apparent. Which begs the question, what are some of the more common areas where predictive analytics is able to make a difference within a claims operation?
Not surprisingly, one of the most common and impactful areas using analytics is identifying fraud, estimated by the Insurance Information Institute to cost insurers over $30 billion annually. Claims fraud adds approximately 10 percent to loss and loss adjustment expenses.
The main challenge is in identifying the true instances of fraudulent activity among the millions of claims filed each year—a veritable search for a needle in a haystack. This effort is made more difficult by the fact that besides the typical types of fraud, there are the more complex issues of staged accidents and padding or inflating the value of a claim that occur. In fact, according to the Insurance Research Council, up to one in five claims received by a company may appear to be fraudulent and require investigation, with padding occurring at a 2:1 frequency over staged accidents.
By applying predictive analytics, a company can improve the probability of identifying fraudulent claims for review. This improvement translates directly into a more efficient use of resources, a reduction in false positives (or Type I errors), and fewer fraudulent cases missed (or Type II errors). Type I errors represent wasted resources as well as customer alienation as legitimate claims are investigated. Type II errors result in higher claims costs which translate into increased premiums and an eroding of competitiveness. Both put the company at a disadvantage.
Advances in the use of automobile fraud detection specifically have improved the fraud detection capacity of companies by as much as 6.5 times. As a result, the more forward-thinking companies are leveraging data mining, pattern recognition, data visualization, and risk scoring to better identify high potential fraudulent cases.
Resource and claims prioritization represent another area of potential for analytics, and comes to fruition via the process of evaluating best next steps based upon specific attributes of the claim. One of the costs of loss adjustment relates to the expediency and type of resources deployed at time of first notice of loss. Proper determination of severity and the relevance of expertise and timeliness assist in selecting what resources to deploy when, whether it is an immediate dispatch of a field adjuster or simply the acceptance of the filed claim with supporting documentation. In addition, an analysis of prior claims results across providers and suppliers can result in presenting least cost alternatives given the characteristics and location of the claim. Alternatively, as part of the claims evaluation process predictive analytics can identify claims that likely will be settled at a higher value and set them as high-priority for internal handling, while the lower-priority claims can be outsourced, put lower in the queue, or sent to a lower-cost skill set for handling.
As a company learns more detail about its claims payments, that information can be used upstream to more accurately market, price, and underwrite coverage. Behavioral information and customer proclivities as well as utilization statistics can be used to individualize and target products to specific market segments, identifying unique needs and offering features that address those needs. The claims segments and use patterns discovered by data mining and validated by the development of a predictive model can also offer pricing differentiations to fine-tune the profitability, competitiveness, and performance of a plan, providing the actuaries with insights at a more granular level than the typical frequency and severity tables.
From an underwriting perspective, greater awareness of the characteristics that may lead to increased claims costs or fraud provide a framework for enhancing the risk review process, reducing exposure where possible while capturing sufficient information to allow for effective management of the risk. In all cases, predictive analytics must first be incorporated from an enterprise and not department level so the marketing, pricing, underwriting, claims cycle is a closed loop providing continuous analytical feedback throughout. This closed loop approach, as opposed to the typical departmental view, allows a company to optimize its profitability across the entire product lifecycle.
Despite the many benefits offered by predictive analytics, there are a number of hurdles that have to be overcome in order to benefit from effective application of the tools. In a recent survey of insurance executives, three of the surprisingly most prominent hurdles were executive sponsorship, adequate ROI, and sufficient prioritization. These are organizational as opposed to operational hurdles, and should be addressed directly as part of the management team’s strategic planning sessions. The breadth of impact and strategic insights offered by a predictive analytics solution warrants top down support and direction setting; it is not a project to be undertaken under the radar or as a sample test in a small corner of the business.
The low profile, low risk approaches do not provide a foundation for the leveraged enterprise-wide potential predictive analytics can realize. As for ROI, this was a surprise response that indicates the need for better awareness, education, and communication, as the returns on analytic projects are both significant and rapid for the companies that have engaged in the efforts. Priorities and resources are always a challenge in today’s busy times; however, the leveraged impact on profits that an effective program offers should more than offset the need to either supplement or reassign the necessary dedicated resources to successfully implement.
Beyond the organizational hurdles, there are also operational issues that most companies face once they decide to implement a predictive analytics solution. The most common, and most difficult to overcome, ones found across the industry typically involve the following:
• Data quality is typically poor and represents one of the hardest issues to overcome, as the scrubbing process must identify and locate missing or incomplete data;
• Models requires a holistic view of the data, difficult to compile across distinct and separate legacy systems;
• Integrating with real-time transactions versus historical data can add complications to implementation;
• Predictive models are sensitive to changes in underlying parameters and therefore must be continuously updated and validated against current data;
• Finding the right talent able to integrate the functionality of predictive analytics into an operation is difficult, as it requires a broad view of the organization as well as a transactional and analytical perspective;
• Organizational resistance is often a challenge as new practices around prioritization and decision-making are installed;
• Transforming analytics into action requires formal procedures and standards that are routinie within the operation;
• Moving beyond point solutions to comprehensive decision-informing claims analysis requires a strategic perspective of the data enterprise-wide starting at new business.
Even given these inherent challenges, the potential for significant returns on the use of predictive analytics that directly translate into greater market competitiveness seem to warrant the investment. Companies that do not attempt to leverage this technology as part of their operation may find themselves at a price and transaction cycle disadvantage.
Famed author of the German play Faust Johann Wolfgang von Goethe once stated: “Knowing is not enough; we must act.” This is particularly applicable to the field of analytics, where retrospective business intelligence like metrics, reports, and dashboards inform but do not necessarily drive action.
Predictive analytics, on the other hand, deal with future events and prescribe the actions to be taken—it is actionable analysis. What priority to give a claim, whether or not to audit or investigate, what resources to assign and when to assign them, are all actionable decisions informed by an effectively deployed predictive analytics solution for claims. Yet beyond the specific actions taken, well developed models also offer insights into customer behaviors, product performance, and market risks. These insights can then be factored into broader strategic decisions that directionally impact a company’s efforts.
Beyond the immediate returns of reduced fraud, increased efficiency, released investment capital, and faster service lay the potential to take a holistic view of how risk is being managed enterprise-wide from product pricing through underwriting to the claims transaction. Companies that leverage the insights that come with this perspective will find themselves better equipped to effectively compete in today’s challenging environment.