Does Analytics Mark the End of Underwriter Influence?

Free Preview

This is a free preview of of our exclusive, member only content.
If you enjoy this article, please consider becoming a member.

Underwriting is the secret sauce of the insurance industry. The process of selecting and pricing risk is essential to the profitability of a carrier, and the deep expertise necessary for this skill set is among the most valuable assets an organization has. Insurers traditionally have competed largely based on the individual judgment of highly experienced underwriters. Insights captured in manuals of procedures and underwriting guidelines are carefully taught to succeeding generations.

Over the last few years, carriers have been heavily engaged in replacing core policy admin systems and increasing the automation of their underwriting processes. This is enabling a fundamental transformation of the underwriting process. Gone are the days of identical products across the industry and standard rating algorithms used by all carriers.

Carriers are using new technology capabilities to create dramatically different products and to develop innovative processes that drive efficiency, improve decisions and transform the customer experience. This underwriting transformation is enabled by the ability to use business rules to drive automated workflow, but even more important, this is a story about the transformation of insurance through the application of data.

Almost every aspect of the underwriting and product management functions is being fundamentally transformed as carriers find new ways of utilizing and applying data. It’s no wonder that investments in core system replacement are currently outpaced by investments in master data management, data governance, predictive modeling and business intelligence tools.

There is no shortage of remarkable assertions around the power of data to transform the enterprise. Predictive analytics are being used to better assess risk quality and assure price adequacy, as well as control costs by assessing which types of inspections are warranted or when to send a physical premium auditor or to purchase third-party data.

While analytics provides a powerful tool that can help organizations find eye-opening opportunities to improve underwriting, it’s not as simple as just buying a model. Companies should consider four major components as they move down this path:

  • Data and tools: Companies need to assure their existing data is cleansed, consolidated and linked where possible. Third-party data is widely available, and companies should think through which data has the potential to provide the greatest value. To best utilize the analysis, companies find it most helpful to have tools that allow them to easily access, report, visualize and drill down on the data.
  • Use cases: There are a number of use cases in underwriting analytics. The predominant use cases focus on risk assessment, matching pricing to that risk assessment and book profiling with a focus on taking action to shift the book to a more profitable composition. Companies also need to determine which approach they’ll use toward modeling these different use cases. While some have data scientists on staff, most institutions don’t have the skill set or resources to use their own personnel for this topic. Companies do have other choices: They can hire external consultants to build bespoke models, utilize scores that address very specific aspects or work with firms that have embedded models with technology focused in this area.
  • Operations: Simply having a model isn’t enough. The next task is to assure the use of the model—making sure that when certain conditions exist, underwriters take the appropriate actions. Carriers that have replaced their core systems with modern solutions that can use business rules to drive tasks have an easier time with this. But it’s not just about the technology itself. Standing in the way of adopting an analytic approach may be a corporate culture that encourages individuals to rely on their intuition and experience when making decisions. Almost 40 percent of carriers report that a resistance to understanding data issues creates difficulties in finding value in data. And underwriters may worry about being replaced by a model instead of seeing it as a helpful tool for making a more informed, objective decision. Gaining benefits from predictive analytics requires a devotion to cultural, organizational and procedural changes in addition to the technology aspects.
  • Governance: Once these models are implemented, companies can’t forget the importance of governance. They should constantly measure outcomes and refine ideas, tweak the models as they learn from them, and assure the models are implemented consistently and appropriately. Savvy carriers are creating a road map for building capabilities that will allow them to take advantage of future innovations. Key capabilities include both technologies and skill sets. The jobs of the future require a different type of skill than ever before. Technologies need to be extremely flexible. Data governance needs to be a discipline, and rules management needs to be a core competency.

Putting Analytics Into Action

The notion that a company understands the key drivers behind its use of data and how that directly affects the specific business challenges within the organization is the guiding principle of applied analytics, a term that must become second nature for insurance carriers leveraging data in their business. Analytics provides the competitive edge, but the sustainability of a project and the degree of success it achieves is contingent on tying together the predictive analytics strategy, data, predictive model, implementation and training into one cohesive and adaptable plan. Carriers that excel in these areas have found tremendous success.

Strategy & Goals

The first step in a predictive analytics project should establish what specific problem a carrier is attempting to solve and how it plans to measure progress. If a goal is to align price to risk and establish a market leadership position, what metrics will be sufficient so the C-suite can gauge the results of their investment? In many cases, loss ratio improvement is the preferred metric. However, if growth into new markets is a core goal, the carrier may elect to focus on underwriting consistency as a metric of choice to gauge the predicted performance of new business being written compared to existing markets. Aligning to strategic goals is the backbone for a successful analytics initiative and requires involvement from multiple stakeholders. C-suite commitment to a data analytics approach to achieving a well-defined objective is key.

Implementation

After the strategy and goals have been accepted and agreed upon, proper implementation is the next key driver for analytics success. Each predictive model is built unique to a carrier, and the way the model is baked into the corporate culture should be catered to the individual company as well. In underwriting analytics, decision rules and guidelines are typically used to differentiate implementation. Analytics provide valuable insight, but it’s the people who must use them correctly to achieve success. How much free rein does the carrier wish to give to their underwriters with the model? For some of the lowest-risk policies the model identifies, carriers may want to bypass the underwriter entirely, saving their expertise for more complex risks. Other carriers want underwriters to review every policy, with the model’s risk score used to guide how much leeway on debiting or crediting the policy. The overarching theme is to position analytics internally as a tool to be used in combination with underwriter expertise, not as a replacement. The implementation plan is often tied directly to the organizational culture of the company, and this component cannot be overstated.

valenOrganizational Buy-in

For analytics to work successfully it must be a business decision, first and foremost, with a common understanding of the end goal. Oftentimes for commercial lines insurers, it’s both underwriters and agents who are the most sensitive to new analytics projects. For agents, a carrier doesn’t want to send the wrong message that they are shifting from a customer service culture to straight-through processing. For underwriters, it’s important they don’t feel the company is sidelining their expertise in favor of a predictive model.

A solid on-boarding and communication plan can make all the difference to avoid these issues. Training sessions with specific rules of engagement are necessary to assure underwriters that they still own their decisions and the predictive scores will only make them better at their craft. Underwriters should be trained to interpret individual policy scores with the model in place and partake in frequent discussions with the underwriting team when they disagree with a score. It helps everyone become comfortable leveraging analytics in their decision-making. In this type of situation, it helps to start small and in phases. The first phase might be implementing analytics into new business only, followed by a second phase to include renewal business. Clear decision-making, supervisor review guidelines and easy-to-follow documentation processes help keep all stakeholders on the same page.

Why is it so important to successfully merge predictive analytics with human expertise? It has proven to produce the best lift for carriers, which is a metric to gauge the overall success of improving the risk assessment for individual policies. According to Valen research, the combination of underwriting decisions and the model produces the best lift at 185 percent versus 125 percent for the predictive model alone and 50 percent for underwriter expertise alone.

In order to become a data-driven insurance carrier, change must begin from the top. Without proper management of the data that goes into the model and strategic alignment between the IT and business sides of a company there can be no sustainable, long-term success in analytics initiatives.