Up to now, the basic model of risk understanding has been to observe correlations between claims costs, composed on the one hand of the frequency and severity of insurance claims, and on the other of the characteristics of the insurable assets as described by data. This analysis is the basis for models to select risk and manage risk portfolios. Such information is known as cold data, and primarily entailed a static description of the object being exposed to an insurable risk and the environment expected to influence that risk.
The Internet of Things (IoT) makes available a new type of dynamic information for risk assessment known as hot data. The data relates to the object, its usage, and user behavior, and replaces proxy variables with directly measured variables. Where risk is directly related to the object and the behavior of policyholders, the IoT has the potential to thoroughly transform risk selection, pricing, and monitoring models.
IoT technology is in the growth stage. Usage-based insurance and pay-how-you-drive for car insurance in the US are at the forefront of the development. IoT in a broader sense is expected to impact a range of personal line insurance types that are directly linked to the insurable object, such as automotive, home, and health.
Car insurance is most advanced in applying the IoT, in the form of telematics to monitor driving style and to form pay-as-you-drive pricing or pay-how-you-drive models. Currently, there are more than 150 pilots globally in the telematics field for car insurance, and the number of usage-based car insurance contracts is estimated to be around 20 million. To collect the driving data, insurance companies require drivers to use built-in devices, plug-in-devices, smartphones, or a combination thereof. In some market, telematics has already helped early adopters improve profitability and competitive advantages, but not in some other markets. The different impacts are explained by multiple factors, such as sophistication of pricing and current pricing levels, potential risk of car theft, and existing driving behaviours, etc. More value can be created through adopting telematics in a market, where auto insurance pricing is high, the level of social security is low, or drivers often drive badly.
Similar pilots have emerged in health and life insurance. Wearables monitor the health-relevant behaviour of the insured person, and in the case of health-beneficial behaviour the insurance companies grant discounts. These models require insurers to overcome customers’ data privacy concerns.
The IoT is also expected to impact commercial insurance, with a greater emphasis on risk prevention than on pricing models. We see potential especially in agricultural insurance, building insurance, and insurance for small and medium businesses.
In addition to helping a better understanding of risks, the IoT is enhancing the customer experience. Traditionally, insurance has been a low-interaction business, with most interactions having negative connotations because they concern claims management. The IoT enables insurance companies to evolve from a low-frequency, transaction-based model to one where they work on prevention and provide advice, coaching, and rapid assistance. In the eyes of customers, the transition from “insurer as payer” to “insurer as advisor and protector” is a fundamental change in positioning. The IoT thus gives insurers the opportunity to unlock new revenue streams, reduce claims, reinforce customer relationships, and improve their images. The extent of the value generation will depend on the sophistication of the pricing, fraud, and claims models which are in play today. This can be seen in the degree of penetration of telematics in different markets: Italy, which has a less-mature traditional model; the UK with its skewed pricing; and Germany, which historically already has very sophisticated loss models.