Are You Reaping All the Benefits of Big Data for the Insurance Industry?
By Sophia Van, Principal CTO, Mercer Marsh Benefits
Sophia Van, Principal CTO, Mercer Marsh Benefits
Though many believe the insurance industry is still in the Stone Age in terms of technology, we have witnessed increasing technology investments. With a market size of US$46 billion and US$187 billion forecast in 2019, Big Data is a key enabler driving disruptions and creating hard-to-copy competitive advantages.
Moving Beyond Traditional Business Intelligence
Business intelligence once involved reporting, benchmarking and prescriptive analytics alone on a small set of structured data. But organizations now have unprecedented opportunities to acquire much-deeper understanding of risks and consumer needs courtesy of mass digitization and the evolution of the Internet of Things. Big Data is so complex and large, however, that traditional analytic skills and tools are inadequate, necessitating strategic investments in three key areas:
Business models and applications
To start a Big Data initiative, a common mistake is to immediately hire data scientists or to build Hadoop clusters without identifying strategic positions where organizations should play a pipeline of feasible pilots to test the fit. Instead, organizations should first understand how existing business models could be transformed or disrupted with relevant Big Data applications. For example, risk scoring might be a strategic focus for reinsurers, who could invest in machine-learning scoring models. Conversely, customer interaction could be game-changing for primary insurers, who may prefer to look into using chat bots or virtual sales assistants. Similarly, individual fraud detection may be more important to primary insurers than reinsurers, because reinsurers deal more with portfolios than individual policies.
Data
Also important is having a data strategy and a deep understanding of data capabilities in terms of availability, accuracy, adequacy, reliability and compliance. Almost all organizations will have problems with their data, whether the problem is fragmentation, inaccuracies or inadequacies. But it is extremely costly to try to fix all the problems at once. Instead, identify quick wins from analytics pilots that can be done with the status quo or little improvement to the data on hand. And throughout pilots, quickly learn and develop your strategic data roadmap—what additional data is needed; how data should be accumulated; and how and when strategic partnerships should be formed to gain early access to critical external data sources.
Big Data is a key enabler driving disruptions and creating hard-to-copy competitive advantages
Skills, tools and infrastructure
Big Data is like a diamond mine, with 80 percent of data unstructured; it needs the right tools and skills to mine the diamond. The number of tools available to analyze the data and extract useful insights is growing rapidly. Popular technologies include voice mining, text mining, machine learning and cognitive computing.
Potential Applications in the Insurance Industry
Big Data is most applicable in insurance in five key areas: claims and fraud management; customer relationship management; pricing and risk assessment; distribution management; and product development. Particularly in pricing and risk assessment, new tools and data sources (such as genomics, wearables, sensors and satellite data) will revolutionize underwriting because insurers now collect actual behavior data at much more granular levels in real time.
For example, traditional car-insurance underwriting looks at vehicle make and model along with driver profiles (such as age and gender).This non-real-time data tells us nothing about driving behavior. With telematics, we could collect (almost) real-time data that tells us more accurately about the risks. GPS and sensor data can tell us the frequency and magnitude of hard-breaking and acceleration events; combined with contextual data, this information can tell us whether such driving behavior is risky within the given context—that is, speed relative to traffic, road and weather conditions.
Similarly, although lifestyle has significant impact on health risks, lifestyle was not traditionally an underwriting factor for life and health (L&H) insurance. The recent leapfrog in wearables and analytics has removed barriers for insurers to track and incorporate lifestyle factors into L&H underwriting models because it becomes increasingly easy and inexpensive to collect this kind of behavior data. This leapfrog enables personalized, automated and disruptive underwriting for an innovative type of product, “wellness insurance.” Similar to the objective of many wellness insurance programs—to help people make positive changes to become healthier—Mercer Harmonise, a digital employee benefits platform, is applying analytics by providing employees personalized health and financial wellness guidance.
How to succeed with Big Data?
With increasing analytics spending, the question globally is no longer whether Big Data is relevant but rather how organizations can leverage and monetize Big Data effectively.
To succeed, Big Data analytics must be in top management’s agenda, because a clear strategic vision drives how organizations deal with data as a valuable asset, how appropriate short-term and long-term investments should be planned, and how the right innovation culture should be cultivated.
A culture of innovation must encourage controlled experiments and accept that failure is necessary. This requires a significant mindset change, especially in risk-averse cultures. I personally encourage my team to “Fail Smart, Learn Fast.”
Starting small is an effective way to experiment with controlled damages. This approach will also help organizations be agile by quickly learning from failures and moving on.
Success needs to be commercialized and scaled up.
Predictive models have no business values until being integrated into products or applications. For example, Benefits Forecaster by Mercer Marsh Benefits is a diagnostic app that helps companies to predict their future benefits cost with greater accuracy to uncover possible savings opportunities and assess sustainability of benefit programs.
