The evolution of data processing—with cloud computing, software-as-a-service and the hybrid cloud — and the advent of advanced analytics, and their penetration into the world of insurance, is shaping the future of the insurance industry. The cumulative changes in technology add up to a cocktail that is going to make more data available where it is needed, faster, for better risk mitigation, business improvement, improved policy formulation, timely fraud detection and enhanced customer satisfaction.
It’s worth remembering that the foundation of the insurance industry has already been built on data to a large extent, looking back at patterns of information from the past to determine the future. But until recently, this business view has been built on a descriptive model, falling short in providing accurate, relevant and organized data, rather having inadequate data or ‘dirty’ data. The industry is now shifting towards advanced predictive models that hinge on quality data and better analytics.
Here are several ways that data technology is helping the insurance industry take a leap forward to a better future, being in a position to predict customer preferences and customer behaviour.
Assessment of risk
Insurance is an industry that is heavily dependent on the accuracy of predicting risk. At its most basic understanding, insurance exists to protect the customer from a potential loss, and to predict the likelihood of that loss, or claim, occurring.
The correct assessment of risk determines the policy premium that’s payable by the insured, upon which the insurer’s losses or gains are dependent. Previously, risk assessment was conducted mainly using a mix of intuitive understanding, historical data and demographic data by the underwriters. But today, risks that were previously considered complex, requiring the personal judgement of the underwriter, are increasingly able to be modelled, and suitable for applying standard—though complex —products and solutions.
With the advancement in data-driven technology, guesswork is being managed out of the insurance equation. But all of this depends on having ever-greater availability of risk data, and granularity of understanding the risk of an individual person or company.
Increasingly insurance risk modelling can use data on an individual’s real behaviour, for example in relation to their health or driving risk. This can come into an insurer’s core IT systems from wearable technology (such as medical monitoring devices, a smart watch or fitness band) or from telematics devices installed in vehicles. In property insurance, some interesting data sources for the future would be around active protection with devices for escape-of-water, photoelectric and temperature sensors for detecting fire and smart locks, especially for commercial buildings.
Such real-time data on the behaviour of the insured brings a new world of data into the insurance underwriting and pricing decision, ensuring better prediction and calculation of risks.
In motor insurance for example, predictive modelling helps in identifying whether a driver is more or less likely to have a vehicle stolen (indicated by the areas where the vehicle is parked and other factors) or to be involved in a crash (indicated by the driver’s speed, acceleration, hard-braking events, night-time driving and so on). This is done by monitoring driver behaviour, road conditions, vehicle condition, safety in the neighbourhood and other data inputs such as crime records for an area. Similarly life and health insurers are also using new data sources to improve risk assessment.
Detection of fraud
Predictive modelling and data analytics can also help in the timely detection of insurance fraud and false claims. Using the principle of contributory data or pooled data across the industry, we at LexisNexis Risk Solutions have been working with insurers on building fraud scoring models by detecting anomalous behaviour. This fraud behaviour can be uncovered when insurers cooperate together for a greater benefit to identify cases such as:
- Vehicles written off or stolen and settled in a claim, then appearing as a fresh vehicle on another policy
- Vehicles with multiple claims for the same accident on different policies
- Vehicles subject to total loss (and supposed to disclose the engine/chassis number) then subsequently appearing on another policy
- In third party personal injury claims using personal identity data to identify drivers or claimants appearing across multiple cases, for example where names and identities are artificially planted.
These are just some examples of how pooled data can help uncover frauds through collusion in the motor claims process.
Suspicious claims identified through risk scoring can then be subjected to further investigation, thus reducing losses and focusing the investigative resources on the most important cases, and those bearing highest legal costs. The accuracy and speed with which fraud can be detected with data analytics is not achievable with human effort. Need of the hour is to supplement human effort with lot more data points and focus it only on the most critical, complex and valuable cases while letting the algorithms handle the regular and easy cases in a straight-through fashion.
Predicting customer patterns
Gaining intelligent insights into customer behaviour, their preferences and tastes, habits and needs can help insurers deliver the most appropriate policy for a particular consumer, appropriate to their life-stage. For example this can take the form of identifying those customers who are most likely to show loyalty, and being able to price their premium accordingly. At LexisNexis Risk Solutions we use our LexID identifier to create a single view of the customer across products and attach all information regarding the individual to this ID.
For example, as customers have many transparent choices when it comes to coverage, it’s not unusual for them to switch life insurance providers several times in their lifetimes.
The proliferation of these choices puts ever-increasing pressure on insurers to attract customers more likely to remain with them for the long term. We commented in a previous blog article on the challenges of improving retention in the life insurance sector in India, which has a persistency rate of just 61% by the 13th month (meaning over four in ten policyholders lapse after the first year).
This has a significant impact on an insurer’s profitability as they are not able to recoup the acquisition cost of that customer.
Add to this the non-monetary cost of losing policyholders, including:
- Loss of opportunity to cross-sell or up-sell additional coverage
- Loss of additional revenue through customer referral
- Lost feedback with no opportunity to improve
Identifying, attracting and retaining high-quality policyholders is a challenge that insurers across the industry face. To solve this, insurers must amass and utilise highly predictive, insurance-specific intelligence while prospecting. Doing so allows you to more accurately determine the potential for a prospect’s retention.
Smart customer handling and servicing
Customer servicing can be looked at from two angles: internal efficiency gains and the external, creating a ‘wow’ factor for the customer. Big data and technology are helping insurers take a closer look at their own data quality and understanding what processes can be improved for better data collection that drives subsequent profits in underwriting and servicing.
Over the years, data quality and data technology have typically been compromised to improve operational efficiencies. It is now that insurers are realising the gain from improving data quality, and focusing on better data technology and analytics, is much more profitable compared to the already lean and efficient operations that are run.
Real time analytics and data prefill tools will help create a memorable experience for customers. As long as data is used by consent for a specific purpose in a manner that helps a customer meet his objective in an effortless manner, we have created an overall good experience for the customer.
Improving marketing strategies
Understanding customer patterns using data analytics increases opportunities for insurers to cross-sell and up-sell policies. Services can be personalised, product portfolios can be improved (by removing under-performing products and defining new ones), premiums can be adjusted, and targeted marketing strategies can be applied by insurers to improve business.
Life, health, home and vehicle insurance products can be tailored to suit the needs of the customer, improving customer satisfaction and loyalty. Long term relationships with customers can be built and churn rates can be reduced. Customer acquisition, which is generally the biggest product cost for insurers, can be made more effective using smart and targeted marketing with a higher probability of conversion to a retained customer.
Advances in data technology are bringing improved decision-making processes, more automation and decision-trees for processing in insurance, enhanced operations and improved business insight. Analytics solutions speedily detect previously undiscovered areas of high risk and open up fresh opportunities in coverage and underwriting.
From its very foundation, the insurance industry has been about the prediction of future events and the assessment of risks and potential losses arising from such events. All of this in the age of big data is still going to rely on organising, mapping, cleaning and governing the relevant data in the insurer’s system. But the potential is going to be very exciting considering the variety, volume, velocity and veracity of huge sets of data. The proper use of this resource is going to help the insurance industry improve in some unprecedented ways, delivering products that are most relevant to the insured, and most likely to trigger loyalty and retention.
Follow the link to the LexisNexis Risk Solutions India website to find out more about how we support insurers. Follow these links to read more about specific solutions for life insurance, health insurance and motor insurance.