Analytics in Insurance Overview
- Operational Versus Traditional Analytics
- Key Insurance Analytics
- Challenges
- Insurance Analytics Evolution
- Summary
- Key Terms
- Additional Resources/Reading
This chapter is from the book
This chapter is from the book
This chapter is from the book
Insurance is an industry that runs on data. Regardless of the industry segment—Property and Casualty, Life, or Health—insurers provide a service or financial product versus a physical product to policyholders, and most often a promise to pay to or to indemnify on behalf of a policyholder or to invest on behalf of. From the initial marketing and customer insurance shopping and buying experience, through to a claim payment or investment payout, the entire insurance value chain, the end-to-end link business processes, are data driven. Data is literally the life blood of the insurance industry and analytics drive overall industry business performance through increased revenue or decreased expenses. It relies on data and analysis to improve financial performance and services to shareholders and policyholders.
Analytics is the discovery and communication of meaningful patterns in data. Business intelligence (BI) strictly speaking is a subset of analytics. BI is generally understood to be more narrowly focused on reporting and data visualization technologies such as dashboards. Another area of analytics is predictive analytics, which includes data mining, text mining, and predictive modeling. Advanced visualization is yet another area of analytics which includes maps, three-dimensional graphics and more sophisticated visual representations beyond tables and simple graphics such as those found in Microsoft Excel. BI and analytics are used interchangeably through this book from a practical perspective.
Analytics improve business processes, decision making, and overall business performance and profitability through insights gleaned and actions taken based on these insights. Analytics rely on the simultaneous application of statistics, computer programming, and operations research to quantify performance. Analytics often use data visualization to communicate insights.
The overall analytics process takes raw data from multiple operational systems, transforms that data so that it is “normalized,” (or in the same format regardless of its source), often augments it with external or third-party data, and turns it into information. It then enables analysis and produces insights or observations from the information and, lastly, guides decisions or actions based on the insights made.
Operational Versus Traditional Analytics
Analytics take two major forms: operational analysis and traditional analysis. Operational analytics, often referred to as embedded analytics, are embedded or built right into the business processes or application systems such as marketing, underwriting, claims adjusting, and so on. Operational analytics are more real time because immediate access to the data brings higher value for certain functions before an action is taken, for example, claim fraud detected during a first notice of loss reporting. It is more effective to identify and prevent paying a fraudulent claim than to pursue the payment recovery.
Traditional analytics such as loss development trend analysis or emerging loss exposure analysis occur after the processing transactions take place and are based on more aggregated analysis. Analytics in both these areas are increasingly using predictive analytics to go beyond understanding historical trends about what happened and why, by using leading indicators and correlation metrics to forecast what will happen and even to optimize future business performance
External Data
It’s hard to imagine the industry running without analytics as analytics have been so hard-coded into the industry DNA in marketing, underwriting, pricing, and claims with a wealth of data available at the micro or tactical operational level. Yet at the macro or strategic level, when deciding on new products to offer or markets to enter, insurers have limited and sometimes not any internal data. In these cases, insurers can turn to external data from third parties such as state insurance department filings, industry composite databases from rating agencies, or public records for information needed to develop new market offerings. Insurers can also use psychographic data (psychological or behavioral data plus demographic) as proxies for target marketing and even underwriting characteristics, for example, income, occupation, and so on. Even when insurers have data, they often use external data to augment and enrich their internal data.
Table 1-1 shows common business functions where third-party data is used, how it is used (use case), and the type of data used.
Table 1-1 External Data Sources
|
Function |
Use Case |
Third-Party Data Type/Source |
|
Marketing |
Marketing Campaigns |
Psychographic |
|
Product Management |
New Product Development, Pricing |
State Insurance Rate Filings |
|
Claims |
Subrogation Recoveries |
Warranty Data, Product Recalls |
|
Underwriting |
Risk Profile Enhancement, UW Risk Assessment |
Motor Vehicle Records (MVRs) Credit Reports |
|
Sales |
Lead Generation |
College Alumni Records |
|
Medical Management |
Health and Wellness Management |
Prescription |
The following descriptions provide more detail on the above Use Cases:
- Marketing has used psychographic data to augment existing customer data or as proxies for customer data in target marketing. Profiling characteristics of existing customers, it looks for similar traits in prospects.
- Product managers often review competitor rate filings to compare products and product pricing, as well as target markets.
- Claims recovery analysts have used warranty data and product recall data to augment claims and underwriting data for claims involving defective products. Two examples are sport utility vehicle rollovers involving faulty tires and kitchen fires that originated in dishwashers or stoves.
- Underwriting has used credit scores as part of its underwriting risk models, where allowable.
- Sales management and producers have used various sources such as college alumni records for lead generation.
- Medical managers are using prescription data and even retail data from drug stores to profile members as part of their wellness and disease management programs.
Insurance Industry Data Flow
To understand and apply analytics effectively, it is essential to understand how data is created and processed, and how it flows throughout the organization. Figure 1-1 traces data from data originators on the left, across the business processes with key metrics by process, how it flows to the accounting systems and general ledger, and ultimately to statistical reporting. Although the example is for Property and Casualty, the flows are similar for Life and Health industry segments.
)
Figure 1-1 Property and Casualty Industry Data Flow
Analytic Maturity
Analytic maturity, or the increasing sophistication in the use of analytics, is based on a number of factors. Following are the four key factors:
- People: How analytically sophisticated the employee area is (and how data driven the organizational culture is)
- Process: How mature the analytic processes within an organization are
- Technology: What tools have been selected, deployed, and made available to employees
- Data: How well data is managed, governed, and made available within the organization
Figure 1-2 shows different analytic maturity levels.
)
Figure 1-2 Analytics Continuum Matrix