Insurance has always been a data-driven business. The more information an insurer has about the risk at issue, the more accurately the insurer can underwrite that risk.

In the past, insurers’ access to data was limited largely by human capabilities. Insurers could understand and underwrite risk based on their own experts’ capacities for analyzing and understanding data. The data used was limited by other humans’ abilities to gather and record information.

Today, those limits still apply — but in a new way. Digital disruption makes it possible to create, store, access and analyze vast sets of data. Analysis can occur in one instant and the results made available in the next.

Data analytics provides an exciting new world for insurance. Yet this new world poses challenges as well.

It’s a Data-Driven World

Information has always been valuable. Data about weather events helps insurers better understand and predict risks. Information about customer behavior helps retailers decide whether, when and how much inventory to order and how to display it. Circulation logs give librarians the information they need to curate media collections effectively.

The explosion of data-driven devices in recent years, however, has generated its own baby boom in the form of data. Approximately one trillion connected devices are expected to be in use by 2025, generating an “avalanche of new data,” write Ramnath Balasubramanian, Ari Libarikian and Doug McElhaney at McKinsey.

Insurers see the (digital) writing on the wall. In an EY Data Science in Insurance survey, every responding insurer said that data analysis will be very important to insurers’ continued success, write EY’s Corina Gruenenfelder and Sabine Betz. Yet not every insurer has taken strides toward adapting to this digital future. Only 25 percent of respondents have advanced data analytics methods in use for their core business. The other 75 percent hoped to improve and expand upon their existing analytics methods with better use of digital analytics tools.

Most insurers have begun the transition to robust digital data analysis by focusing on analytical tools to apply to their own collections of data. Insurers have focused on gleaning insights from their own stored customer data, information about their internal processes and data collected from agents and brokers. Yet insurers also have access to a rapidly-growing pool of data from outside sources, including public data sets and third-party aggregators, writes Kirill Pankratov, head of group UW transformation in commercial insurance at Zurich Insurance.

Tools that allow insurers to analyze this data for insights and patterns will provide information that insurance companies have not historically been able to access — and that will profoundly change how insurers underwrite risk and distribute coverage.

Applying Data for Better Insurance Decisions

Applying data offers insurers new ways to conceptualize, address and solve old problems. Still, data analysis poses challenges of its own.

Better data analysis can help insurers manage costs, says Tim Brockett, executive vice president and head of specialty lines at Munich Re US. Brockett notes that with the right tools, the insurance industry “can enhance our pricing, our underwriting, and our analytics around claims and customer segmentation to improve our profitability, and understand our risk.”

Underwriting and distribution offer fertile ground for insurance data analytics. “Actuarial science as traditionally practiced bears many similarities to data analytics,” writes Rachel Hastings at Emeritus. Both actuarial science and today’s data analytics depend on high-quality data inputs for their success.

For insurers, however, success with data analysis will depend on how data is understood and used.

Shifting risk patterns have started to create two distinct categories: Risks that are more measurable, more frequent and less severe versus risks that are less measurable, less frequent and more severe. For example, the risk of basement water seepage during heavy spring rains falls into the first category, while the risk of a catastrophic flood washing out the entire basement of the same house falls into the second.

Data analytics can address both types of risk, although the challenges of applying data to each type differ, write Tanguy Catlin and fellow members of McKinsey’s Insurance and Organization practices.

For example, more measurable and frequent risks involve much larger quantities of data, especially when individual real-time data points from smart sensors and similar devices are involved. Insurers will need tools that allow them to aggregate and synthesize massive data sets and to track existing patterns as the basis of predictive analytics.

By contrast, less measurable and frequent risks come with smaller data sets, but the intensity of these risks puts greater pressure on insurers to predict and price accordingly. The ability to model uncertainties reliably and provide solutions beyond risk transfer will be essential, explain Catlin, et al.

Data offers new ways to tackle insurance challenges. It even offers a way to stay on top of rapidly-changing risk patterns. To seize these opportunities, insurers will need to approach their data tools thoughtfully.

Responsibilities and Challenges in Data Analytics

“Data and analytics capabilities are becoming table stakes in the P&C sector in Europe and North America,” write Kia Javanmardian and fellow researchers at McKinsey. By investing in data and analytics, particularly for improved underwriting, insurers have seen loss ratios, new business premiums and retention increase.

As advanced data analytics become commonplace in insurance, three major shifts will take place, says McKinsey’s Violet Chung:

• Insurance premiums and benefits will adapt based on in-the-moment customer behavior, such as in pay-as-you-drive auto coverage.
• Seismic shifts in risk profiles, and access to unprecedented data about those risks, will require insurers to reconsider risks and associated premiums.
• Automated analysis of real-time data will allow insurers to automatically underwrite more types of risk.

These changes in turn will rebuild other fundamentals of insurance, including customer relationships and agents’ roles in distribution. To put data to work, insurers will need to face certain challenges — and accept the responsibility that comes with embracing those challenges.

“To seize these opportunities, insurance companies need to reshape their organizations and transform into data-driven enterprises,” write Marck Timmermans and fellow researchers at Compact Magazine. The excitement of seizing new opportunities must be balanced with the need for security, control of the process and awareness of social expectations.

Other challenges that lie ahead in the use of data analytics include:

• Entrenchment of bias existing in available data sets, especially when data sets are limited.
• Creation of incomplete or misleading customer profiles from public data.
• The risk that autonomous decision-making by algorithms will lead to costly errors.

Ethical concerns about data analytics in insurance focus on the impact of data use on human lives. Yet ethics are not the only reason insurers should consider how their data tools interact with human efforts.

Despite recent advances in technology, the insurance industry continues to rely heavily on the expertise and skills of its human workforce. Data analytics and other technological tools can enhance insurance professionals’ expertise and skills, but it cannot always replace them. Pairing digital tools with human experts can bring out the best in each, leading to further growth and more resilience in the face of future digital challenges.

Images by: skracht/©123RF.com, woodysphotos/©123RF.com, pavelvozmischev/©123RF.com