A rapidly developing consensus in business, education, law, medicine, and even professional sports is that insight into information can assist business leaders in making accurate, objective, and cost-effective decisions. Recently, significant attention has been given to the expanding use of analytics in the insurance industry that spans property and casualty, life, and health insurance coverage. The use of analytical models by actuaries to underwrite and price policies is not new to the insurance industry. However, the amount and variety of data available to insurance companies today provide a wealth of new opportunities to control costs, counter competitive initiatives, grow revenue, manage risks, and prevent fraud.
The information management challenge is to use analytics to optimize business processes, improve efficiency, streamline costs, and respond quickly to an ever-changing business environment. This challenge needs to be handled as a business problem and not as an IT problem. Insurers want to offer products that fit the specific needs of customers.
In certain complex insurance business models such as group health insurance and disability insurance, the complexity of the data within various possible combinations can create an extensive coverage mix. Much of the data related to claims is captured in conventional claims applications by making textual notations. Claims adjusters manually adjudicate each claim, and the process can take from weeks to months before closing the claim—resulting in high overhead and cost. Over time, millions of textual notes and remarks are captured, creating huge natural language data sets—big data.
Strategic decision data resources
Like oil, data is a resource that has always existed. In the era of big data, having the capability to refine it by applying big data technologies is now possible. As technology has evolved, it has provided a sophisticated platform for organizations to process a variety of data sets. Applying analytical tools to these new huge volumes of data requires a distinctly different infrastructure from traditional database architectures and query workload platforms. Once the exclusive domain of scientific research, national security, and oil and gas exploration, today’s big data applications can be run on hundreds or thousands of clustered servers instead of supercomputers. And some large insurance companies worldwide are capitalizing on big data analytical applications to make strategic decisions, enhance the customer experience, and achieve successful business results.
For example, organizations have been using predictive analytics for quite some time, and the endgame has been insight discovery. That insight is then put into action in business operations, and applying analytics to operational systems has taken a lot of time. Today, there is a continually growing amount of data to work with. There is also a lot more variety than before, because in addition to internal data sources, much of the data is coming from a wide range of external sources.
Instead of relying on only internal data sources such as loss histories from claims notes, for example, auto insurers began incorporating behavior-based credit scores from credit bureaus into their analyses. This tactic was based on an emerging awareness of empirical evidence that people who pay their bills on time tend to be safer drivers than those who have less-than-stellar credit histories. While using credit scores in private auto insurance underwriting has been an ongoing challenge between the industry and consumer groups, adding behavioral and third-party sources marked a significant leap forward from claims histories, demographics, and physical data in past analyses.1
The proliferation of third-party data sources is reducing insurers’ dependence on internal data. Data exhaust from computers, consumer and industrial devices, smartphones, social media, and multimedia—used within privacy guidelines and assured anonymity—has become a rich source of behavioral insight for insurance companies, as it has for organizations in other industries. Life insurance providers in particular have historically leveraged some form of analytics in various areas, including market research.
Today, the push for analytics is intensifying; it is driven in part by competition, consumer expectations, financial pressures, and information availability. For example, one large insurance company uses big data applications to look at hundreds of terabytes of data for patterns to gauge how well the company is doing to minimize risk, understand how various products are performing, and recognize trends. Other companies providing property and casualty insurance are applying big data applications to rationalize product lines from new acquisitions and to understand the risks from global geopolitical developments. And other leading-edge insurance and banking organizations conduct experiments by using big data to segment customers and tailor products and special offers based on these customer profiles. Several other use cases are quite relevant to the insurance industry and have high return-on-investment (ROI) potential for organizations:
- Advertising and campaign management
- Agents analysis
- Call detail records
- Catastrophic planning
- Customer sentiment analysis
- Customer value management
- Fraud detection and analysis
- Loyalty management
- Personalized pricing
- Product personalization
- Risk avoidance
- Social media analytics
- Underwriting and loss modeling
Insight from the power of analytics
For years, IBM has been working toward more pervasive adoption of analytics, and a well-suited method to accomplish this goal is to weave analytics into the fabric of business to help drive successful outcomes. As a result, analytics needs to become an inherent part of key business processes across all organizational departments regardless of industry. Client organizations are maturing as their analytics journeys progress. Tools, approaches, and challenges differ depending on the industry, but there is an across-the-board phenomenon of organizational interest in leveraging the power of analytics, and often with a big data use case.
For example, diversified financial and investment institutions are increasingly taking advantage of analytics to help manage risk and drive compliance against growing sets of regulations such as Basel III and the Dodd-Frank Wall Street Reform and Consumer Act. In the early phases of their analytics journeys, organizations tended to focus on connecting their analyses to their internal data repositories only. In recent phases, they are actively looking at connecting their analyses to external data sources such as social media for customer information. And they are using social media analytics in all kinds of ways to understand, sell to, and build credit profiles for their customers. IBM has been a partner with many organizations on these journeys.
When gleaning insight into customers, realizing that life-changing events influence customers’ changing segments is an important consideration. Getting a new job, concluding a divorce, having a child, receiving an inheritance, relocating to a new address, and many other experiences impact customer profiles. When customers call and ask for advice because they have received an inheritance, for example, that additional funding may place them in a new segment. The enterprise data warehouse probably has them in a different segment. Organizations need to recognize these kinds of pivotal events and be able to take action in real time.
Big data innovation for customer profiling
Gaining a comprehensive understanding of customers—what makes them tick, why they buy what they buy, how they prefer to shop, why they switch providers or brands, what they’ll buy next—is strategic for virtually any business. Equally strategic is learning what factors lead customers to recommend a company to others. The IBM Institute for Business Value report, “Analytics: Real-World Use of Big Data,” offers as its top recommendation that organizations need to focus their big data efforts first on customer analytics “to truly understand customer needs and anticipate future behaviors.”2
Big data analytics not only provides strategic insights into customer behavior, but it can also emphasize the importance of the 360-degree view of the customer in how it extends to front-line employees. Forward-looking organizations recognize the need to equip customer-facing professionals with the right information to engage customers, develop trusted relationships, and achieve positive outcomes such as solving customer problems and upselling and cross-selling products. To realize these aims, organizations must navigate large amounts of information quickly to zero in on what’s needed for particular customers.
Big data has the ability to change business models. For example: Who will insure that diamond ring purchase at the point of interaction? Will it be Google, Verizon, Visa, or an insurer making the offer on a valuable items policy? Most likely, the organization making the offer will not be the insurer. So who are the insurance company’s partners? Can someone insure the contents in a single trip aboard a train?
Using radio-frequency identification (RFID) today, knowing what a train or ship is carrying is easily possible. Typically, freight is aboard a ship, train, or truck only 10–25 percent of the time. The remaining time it is at either end of its trip in a warehouse. Is insuring its trip a form of pay-as-you-go insurance? How much location risk is there in the warehouse where it is stored?
Many insurers have some sort of segmentation program. The goal is to identify the segment a customer is in and match the right offer to the customer needs when the customer reaches out through any channel at any time. Big data expands the number of segments when interaction and behavioral data are included. The potential combinations of segmentation rapidly lead to a customer segment of one. In addition, insurers are starting to think about the human genome in insurance. This consideration is not intended for a USD50,000 life insurance policy, but think of wealthy individuals who opt in by providing their DNA for a lifetime product. For example, if a 40-year-old customer knew she had a 90 percent probability of living to age 98, the offer should be to insure her life for the next 20 years and then annuitize her wealth for the next 38 years.
Analytics uplift with big data
Real-time monitoring and visualization are fundamentally changing the relationship between insurers and the insured. By agreeing to let insurance companies monitor their behavior, customers can learn more about themselves, and insurance companies can leverage the data to influence behaviors. In auto insurance, for example, telematics—technology that blends telecommunications and informatics—are being used to monitor in real time the driving habits of the insured and then send data back to the insurer. There is already evidence that this approach is influencing drivers and improving their driving habits. One UK insurance company using telematics reported that enhanced driving habits resulted in a 30 percent reduction in the number of claims. Another UK insurer similarly used telematics to help a large client reduce risky driving maneuvers that have the potential to cause accidents by 53 percent.3
Usage-based insurance is also a widely leveraged approach in the industry. Pay-as-you-drive coverage and translating good driving behavior into premium discounts are leading use cases. This data can be used in many other similar situations. For example, managing commercial fleets can be enhanced. Improved driving behavior can also lead to increased gas mileage and minimized wear and tear on vehicles, which helps reduce maintenance costs.
Data captured at the point of impact in a collision can be used to support claim damage or reduce fraud and abuse. A crash at five miles/hour can cause significant damage to a car, but it is far less likely to cause long-term human pain and suffering. Insurers can defend against these claims. By using streaming analytics, they can push information on traffic congestion or impending storms to drivers. And they can respond to crash sites and effectively handle the claim and manage loss, but also optimize supply chains for repairs.
To achieve high success using big data, organizations need to master their information, gain insight, apply that insight at the point of interaction, and uplift their analytics with big data. Insurers’ capability to move to this type of business and technology strategy can have a big impact on their success as an organization over the next decade. Big data is now part of the information supply chain.
Please share any thoughts or questions in the comments.
1 “Massachusetts Law Bans Credit Scoring for Auto Insurance,” by Chad Hemenway, Property Casualty 360, November 2011.
2 “Real-World Use of Big Data in Insurance,” IBM Institute of Business Value report, IBM Global Business Services, Business Analytics and Optimization, executive report, October 2012.
3 “Unleashing the Value of Advanced Analytics in Insurance,” by Richard Clarke and Ari Libarkian, McKinsey & Company, August 2014.