Darran Simons at FICO outlines the data sources that could drive hyper-personalised insurance offers – as long as insurers have access to the data analytics expertise.
Insurance providers already draw from a deep well of data. Driving habits, historic claims and weather patterns all deliver essential information for underwriting. But could they go further and base their premium prices on social media activity, refrigerator contents, or educational background?
The short answer is yes, and it’s already happening. The rapidly evolving world of InsurTech and Big Data analytics has opened new opportunities for insurance underwriters to achieve their goals. By embracing rapidly evolving analytics, insurance providers could see improved profitability, respond to new opportunities in real time, better coordinate resources, and drive better customer experiences. However, a critical issue is the lack of skills in the sector to analyse the wealth of data now available.
In April 2023, Forrester published a report on decisioning AI for insurance that uncovered a staggering skills gap obstructing the potential of data analysis and decisioning in the sector. 81% of insurers surveyed reported having the ability to use third-party data but lacked expertise. Only 12% reported having a high level of expertise in data analytics.
New data sources
‘Alternative data’ encompasses data (or other factors) from third-party sources, or not traditionally included within an insurance application or claim submission. Of course, companies’ ability to access, collect, analyse, and take actions based upon these data sources and factors may be subject to local regulations and privacy laws. However, recent developments around the availability of third-party data have significantly diversified and innovative data types have emerged to meet the needs of ever-evolving underwriting practices.
For example, psychographic profiling delves into a consumer’s values, opinions and attitudes, for deeper insights into behaviour and decision-making patterns. In smartphone usage, data can be shared on an opt-in basis, giving insurers the ability to evaluate various user activities and habits, such as exercise routines and screen time, which can be indicative of health and lifestyle risks. And drone-acquired imagery has gained traction as a non-intrusive method of inspecting property locations and underwriting potential hazards like structural defects or flood-prone areas. This is also useful in claims triage during natural disasters.
Looking further afield, insurance providers could make use of data sources such as smart appliances, pet ownership records, and educational background. And if they are looking to improve customer centricity, other data sources could provide a clearer, more colourful view. Crowd-sourced sentiment enables large-scale analysis of online discussions and reviews, fitness trackers record and share information on fitness and exercise levels, and safe driver scores based on driving behaviour can facilitate consistency and fairness.
Applying a data-driven approach
Leveraging a powerful, data-driven approach brings new opportunities for machine learning and other analytic models to inform product development, underwriting decision models, and
robust business outcome simulation. It will also turn contextual insights into reusable commodities within an organisation. And combining these insights with decision modelling capabilities reduces dependencies on IT, fosters greater collaboration across teams and centralises AI decision-making assets.
However, with so much data available to insurance providers, it is essential they also have the right analytical expertise and data platforms. Reaching this level of digital science at scale requires a decision platform (that can readily make use of new insights. According to Forrester, ‘without a clear DDP [digital decisioning platform] strategy, the whole organisation is impacted’. The Forrester research found that 82% of firms agree they would benefit from a centralised platform to support digital decisioning. Yet 38% said they do not know where to begin in creating a data-led strategy, with 62% lacking interoperable data and 51% lacking the right tools.
Accelerating analytics use through a decision platform
As Forrester notes, a decision platform is key to unlocking the potential of new data sources and analytics. This kind of applied intelligence platform enables collaboration between data science or actuarial teams and business users. These platforms integrate multiple vital capabilities, including:
· Operationalise analytical strategies with speed and ease
· Ingesting new data sources with an API-driven approach
· Enabling business users to simulate the impact changes in decision strategies will have on their key performance metrics
We are seeing this approach empower business users to drive change within their lines of business and do so at lightning speed. It has the potential to transform the future of insurance.
Established Alternative Data Sources
These alternative data types are already delivering competitive advantage.
1. Credit bureau-based insurance scores: One of the most predictive, if not the most predictive, rating factors in use in personal lines insurance, public records and electoral role aligning closely to claims propensity, and credit data to support traditional risk rating for premium credit.
2. Property and vehicle data: Can help underwriters identify the risk and profiles of physical assets, such as valuation and undeclared risks. This type data can also be used to reduce the questions Insurers need to ask on applications
3. Social media activity and open-source data: Information from social media data providers can be used to evaluate sentiment, lifestyle, and reputation attributes of a market or entity to provide insurers with a deeper understanding of risk.
4. Vehicle Telematics: Data from connected vehicles on driving behaviour, mileage, and location, used for usage-based insurance and to make more informed decisions about drivers and where and how they operate insured vehicles.
5. Health records: With strict regulation and explicit consent, medical history data, such as electronic health records and prescription histories, can be useful to life insurance and health plan underwriters.
6. Geospatial: This type of location data can be used in the underwriting process to identify location-specific risk factors, such as natural hazards like floods or earthquakes, when providing property or motor insurance premium and coverage amounts.
7. Employment and income: Information about a customer’s professional background, earnings, and employment stability contributes to underwriters’ evaluation of suitability for certain products and coverage amounts.
8. Insurance claims history: Previous claims data, including loss frequency and severity, can offer insights into a client’s past exposure to risks, helping underwriters to better predict future claim probabilities.
9. IoT devices: Data from connected devices is a growing field of new information that can provide real-time insights into an individual’s habits and environment, allowing for more accurate risk underwriting and tailored coverage. These data types are also powering the move towards claims prevention.
10. Customer behaviour: Evaluating data on customer interactions, online behaviour, and service usage can help underwriters better predict policyholder retention and promote loyalty.
11. Consortium data: Pooled intelligence data can be instrumental in combating fraudulent activity.
Be the first to comment