Data is a powerful tool for insurers when analyzed and used effectively. While the pandemic and changing regulations have shifted some of the ways insurers can collect data, leveraging—and effectively governing—in-house and third-party data is still helping carriers stay competitive.
At Novarica’s recent Virtual Town Hall on business intelligence, data, and analytics, Novarica VPs Eric Weisburg and Deb Zawisza joined me in exploring how these issues are impacting the insurance industry.
Third-Party Data Providers
On the life side, traditional fluid-based medical underwriting tools are being impacted by the pandemic—home visits for drawing blood and taking swabs are currently even less appealing to prospective insureds than they were before. Deb Zawisza observed that physical inspections of property for claims or underwriting have been impacted on the property/casualty side of the business as well. She went on to note that, pandemic or not, insurers without new types of data for their underwriting models could be at a competitive disadvantage.
Eric Weisburg mentioned that he was seeing ever more interest in the concept of continuous underwriting. Regular reviews of risks can apply to a range of product types, which can have impacts for margins already being squeezed by a variety of market forces. It might have made sense in an earlier era to find out 30+ years later whether something you wrote was a good risk. That will not work now. As Deb noted, “Insurance is one of the few products sold where you don’t know the cost of goods sold for years, so analytics are absolutely vital.”
Eric also pointed out that the California Consumer Privacy Act (CCPA) and other emerging regulations have implications for the types of data carriers can have, how it can be stored, when it must be deleted or kept, and more. Moreover, we may see these rules augmented or replaced by federal regulations in the coming years as well. Good data governance will be key to maintaining compliance as the regulatory landscape continues to shift.
External Data Challenges
As vital as good data is for the modern insurer, leveraging it comes with challenges. Deb underscored the need to seek out the “right” data, rather than the “cool” data. She advised insurers to look for a good use case and then find the data to support it—not the other way around. At the same time, effectively using data is an iterative process: It involves validating new data, analytics, and hypotheses; tweaking any of those three elements; and then potentially discovering new use cases and data sources before beginning the process anew.
I asked Eric and Deb if the notion of the pandemic having broken models overnight was still a concern. They agreed that it certainly was but cautioned that we still do not have enough data to create new models in many cases, and depending on the nature of the recovery, the models may be self-healing. It’ll be important to look for places where risks may have changed and monitor them closely since reacting too quickly could lead to uncertain results from overcorrection.
That need for oversight brought us to the next question: As underwriting and claims decisions are increasingly made by algorithms, should there always be a human-controlled trump card to override these decisions? Deb advised insurers to have expertise review predictions to help ensure the models are tuned appropriately when implementing analytics. She also observed that sometimes reviewing the models can also help expand the perspective of an underwriter or adjuster to factors they may have been aware of. Sometimes the analytics are wrong, but sometimes they’re right in ways we hadn’t anticipated, and those insights can be valuable. Nevertheless, it’s also important to have humans come back to the process to validate it periodically as well.
Emerging Technology and InsureTech
When asked if it is appropriate to mix and match a cloud provider and third-party analytic tools, the resounding answer was “it depends.” Eric explained that some want to be in the cloud but remain vendor-agnostic and thus able to “lift and shift.” The downside? “You’re less able to take advantage of the cloud-native capabilities the providers have to offer.” Deb brought the focus back to business value: “If you’re already using AWS or Azure for cloud, it might be quickest and easiest to leverage some of their analytic products,” but every company’s needs and philosophy are different.
Finally, Deb had some guiding principles for working with InsureTechs: “Don’t start at scale.” Moving directly from a limited POC to a full rollout across several lines may be too risky. She also advised insurers to be open to learning and pivoting, noting that some InsureTechs may not have a clear understanding of the most worthwhile use cases for their products. Working with them to discover what they are could open up new possibilities, but this should always be driven by business value rather than “coolness.” Novarica’s InsureTech report can be a helpful guide on solutions emerging in the market.
This week, our Virtual Town Hall focuses on evolving talent strategies. If you haven’t already, register here.
Add new comment