Authors: Claudio Stadelmann, Ueli Konrad, Ladina Hostettler
The Internet of Things (IoT) offers the insurance industry disruptive and transformative potential. Nowadays, devices using IoT technology are omnipresent and generate a wide array of data. Leveraging this huge amount of data presents insurance companies considerable business opportunities in terms of customer relationship, data insights and individualization (see Figure 1). On the basis of three use cases, this article provides insights into how IoT can be used to generate value for insurance companies. Further, we show how value-generating business models can be created by tapping into the opportunities. To conclude, we provide guidelines for insures keen to benefit from IoT.
Motor insurance policies offer numerous areas of application for IoT which will lead to a rapid transformation of the market for this kind of insurance. During this transformation, it is crucial to get a competitive edge by quickly launching products and to start collecting, analyzing and exploiting data.
- IoT-Based Motor Insurance
Let us conduct a thought experiment on how an insurer which offers a data-driven, IoT-based motor insurance policy can derive a superior risk model and will become a leading motor insurance provider within a short timespan. An evident way to collect data from the policyholder’s vehicle is to provide a dongle which connects to the vehicle’s OBD II1 interface. The dongle contains a set of sensors, such as accelerometers, positioning services, such as GPS, as well as mobile network connectivity and sends collected data periodically to the insurance company. Dextra uses such a solution in Switzerland – however, only for a mileage-based pricing model (Pay-As-You-Drive). An alternative is to use an app on the policyholder’s smartphone. Swiss Re follows this approach in their white-labelled product Coloride.
- It is All about Data
Once collected, the insurer can analyze the data in order to gain insights about a customer’s driving behavior. For this purpose, the collected data can be enriched by additional information about the weather, the road traffic or the like. With the help of wearables (e.g., fitness trackers or smart watches), even information about the driver’s personal condition, such as sleeping patterns, pulse and physical activity can be used. In this context, strategic partnerships with telco providers or healthtech companies in order to monitor a driver’s vitals remotely are imaginable. On the basis of the aforementioned information, it is possible to create an individual driving profile for each policyholder. Next, the insurance company calculates the correlation between the driving profiles and their associated insurance claims. After collecting the data over a certain time period and thus having a sufficiently large data basis, patterns will emerge such that the insurance company is able to determine an individual premium for each customer.
- Rewarding Good Risks, Pricing Out Bad Ones
With the help of the enriched dataset, the insurer can identify the risk quality within its customer base. That provides the possibility to calculate individual premiums based on granular risk profiles. Consequentially, the insurer can offer attractive rates to good risks, whereas bad risks will receive more expensive quotes over time. For new insurance applicants, novel underwriting approache have to be introduced since driver profiles do not yet exist. They can leverage existing data from the applicant’s mobile devices and the insured vehicles in order to make the underwriting process as accurate as possible.
- The First-Mover Advantage
The bigger the available data pool, the better insurers can determine their customers’ risk quality. Over time, bad risks are priced out of the insurer‘s customer base, whereas good risks are kept due to low premiums. This will lead to bad risks switching to competitors which are not (yet) able to accurately differentiate between good and bad risks. Ultimately, the first mover in the market of dynamically priced motor insurance policies will decrease its claims volume due to amassing good risks in its portfolio. The competitors’ claims volume, on the other side, will subsequently increase because of the shift in bad risks. That leads to a self-reinforcing effect.
- Beyond Being a Commodity Provider
In addition to an appropriate insurance coverage, the customers’ goal is to prevent an accident in the first place. By providing safety tips derived from the collected data, insurance companies can move from being a mere commodity provider towards becoming a true partner for their customers. Tips can cover how, when and where to drive. For instance, navigation applications can be used in order to suggest routes for which the accident risk is lower. By following such tips, the customer not only improves his safety, but also decreases his or her dynamic premium. Positioning itself as a partner allows the insurer to create stronger lock-in, increase the interaction frequency and improve customer loyalty.
- Be Fast and Know the Data
For this use case, we see two key capabilities which are needed to benefit from opportunities generated by IoT: be fast and know the data. First, there are strong early-mover advantages as the insurer that collects data first is the one which can create a better risk model first. Second, the capabilities to analyze and enrich data in a meaningful way are key to create accurate risk and pricing models.
Insurers are rarely known to communicate their fraud fighting activities openly. The IoT-savvy insurer, however, may change this practice and improves the customer relationship by disclosing how the company uses IoT to protect the whole insurance collective and to keep premiums as low as possible. Moreover, data insights gained from IoT fraud detection can improve claims processes and thus reduce the effort needed by the customers. But we are getting ahead of ourselves – let us first introduce how IoT will impact fraud prevention and detection in insurance.
- Fraud as a Major Financial Lever
In 2018, the global insurance fraud detection market size was valued at an annual turnover of USD 4 billion and is expected to show a compound annual growth rate (CAGR) of 13.7 percent from 2019 to 2025 (Grand View Research,2019). An old rule of thumb is that insurance fraud accounts for at least ten percent of all insurance premiums (Business Wire, 2019) – a share which has remained relatively stable over the past two decades. In response to this threat to their business model, insurers have long invested considerable resources into fraud prevention and detection. In general, two types of fraud can be distinguished: Soft fraud, such as exaggerated claims or deliberate damage events, and hard fraud, committed mostly by organized crime, whereby the fraudulent offence is a planned endeavor, usually aiming at gaining large sums.
- Using IoT for Fraud Detection
By using IoT applications for fraud detection, insurers’ efforts to reduce fraud no longer depend on claims agents relying on few facts and a large amount of intuition. As smart home applications, telematics and wearables make their way into everyday life, insurers are keen on evaluating possible fraud detection and prevention methods based on this new sources of data. At first glance, potential applications, such as the use of GPS data to track stolen and acclaimed lost goods or the identification of unlawfully obtained daily insurance benefits using step counter information, seem endless. Some of them are more refined and are capitalizing on data compilation (e.g., combining information from smart home devices with weather data in order to detect fraudulent water damage claims). From a technical point of view there are few – if any – limits to the use of IoT in insurance fraud detection which is impressively illustrated by the following example of a convicted Ohio arsonist: The man claimed he had fallen asleep before a fire destroyed his home. However, a review of the man’s pacemaker data not only revealed him to be awake long before the fire broke out, but also showed that his heart rate was unusually high, indicating elevated levels of stress. That created suspicion and in the end proved the attempted insurance fraud (Johnson, 2017).
- Self-Censorship and Proof
IoT’s greatest impact will be on soft fraud, which accounts for the majority of fraud cases. For one thing, there is clear evidence of increased self-censorship (Johnson, 2017). Policyholders whose driving behavior is tracked by telematics devices are much less likely to try to report falsely on car accidents (Hynd and McCarthy, 2014). Soft fraud has always been hardest to prove for insurance companies. With the help of more accurate data on policyholder behavior, this problem is substantially lessened. However, technological advancement has also paved the way for criminals to attempt cyber attacks and created new ways of fraudulent behavior. Above all, the manipulation of IoT data used by insurance companies to verify claims or to identify fraud – thus, ringing in the next round of insurance companies against fraudsters.
- Impact through Innovation and Target Orientation
Hence, to improve the customer relationship, insurers should focus their IoT efforts on areas in which will be most to gain through the effects of self-censorship and easy to prove discrepancies. That applies particularly to the B2C market. However, the traditional insurance policies in this market have not yet leveraged IoT possibilities. Consequently, insurers are well-advised to seize the chance to develop new products and offers with the help of IoT. Namely, offerings based on data insights, which require the necessary data sharing and tap into the advantages of individualized insurance.
- Enriching Existing Data
At first glance, the link between predictive maintenance and insurance products may not be the most obvious. However, insurers can greatly benefit from enriched sensor data improving their underwriting and claims handling capabilities. Predictive maintenance uses sensor data to evaluate the condition of equipment in service in order to determine the right time for maintenance work such that the possible downtime can be reduced. Since sensor data allows for better forecasting of future machine breakdowns, accidents can be prevented, increasing the overall plant safety. In addition to these advantages to the industry, sensor data provides insurers with the opportunity to understand the risks they are insuring much better.
- Automated Claims Handling
Connected sensors continuously provide information on the equipment’s condition. In case of damage, sensors immediately send the necessary information to the insurance company which autonomously triggers the claims handling process. This significantly improves the customer experience since tedious paperwork is no longer necessary. In addition, claims can be processed automatically, since comprehensive information is provided by the sensors. At the same time, insurance companies will gain an indepth understanding of the clients and their risk profile. This in turn can be used to handle the portfolio of insurance policies more efficiently (e.g., through a bonus-malus system).
- IoT as a Data Multiplier
Sensor data has far greater potential than just simplifying the claims process. Used area-wide, sensors can provide much more than isolated information on the condition of a single machine or a whole factory. For example, sensors collecting information on weather, traffic, usage profiles or capacity utilization can be used to enrich insurance-specific data. This is particularly interesting for insurers with many policyholders using IoT products since they are able to collect a huge amount of data across companies and industries. Applied comprehensively, data patterns will allow insurance companies a better understanding of the circumstances leading to the occurrence of risk events. That in turn enables insurers to better advise their clients on how to minimize their risks. However, to take this business opportunity, insurers need the capability to analyze large amounts of data. Without this prerequisite, it is impossible to combine sensor data and existing knowledge to accurately identify risk patterns.
- Get Access and Enrich Smartly
This use case illustrates how insurers can use IoT to change their business model and become a partner for their clients rather than being a mere commodity provider. However, this transition is a step by step process which requires two key elements: get access and enrich smartly. The fact that sensors collect data does not automatically mean that insurers have access to this data. Hence, insurers should collaborate with their clients to ensure access and the rights to use the data. Once this is established, it is crucial to build up additional competencies to process the new information and draw the right conclusions to further enrich the existing data.
The use cases presented in this article illustrate how insurers could benefit from IoT. However, in order to exploit the full potential of this technology, insurers will not only have to adapt their business models but also have to invest in new capabilities and adjust organizational processes. To make this possible, we identified three key capabilities: Be competent, be fast and be secure.
- Be Competent
Data competency is key. Not only will insurers have to attract extraordinary talents in the field of data collection and interpretation. Becoming a data competent organization also requires a shift in the managerial mindset, serving as foundation for any subsequent changes in the operating model and the technological infrastructure.
- Be Fast
Speed is of high relevance in the IoT market. First-mover advantages can be generated if a company collects data first and thus is perceived as a pioneer regarding IoT. Therefore, it is crucial for established insurers without in-house IoT knowledge to early join forces with smaller, faster and more flexible partners (e.g., startups) in order to create sustain-able solutions. Equally relevant as partnerships with IoT knowledge carriers are partnerships with companies owning data (e.g., telecom companies) and making customers to agree with sharing their data.
- Be Secure
Especially in today’s uncertain legal context (e.g., because of the upcoming Swiss data protection law), it is essential to handle personal data in line with data protection requirements. As shown, insurers must act now but within the boundaries of the legal system. Moreover, moral factors have to be considered as studies illustrated that even if the data collection is legal, it can be perceived as immoral which could cause reputational damage. In summary, IoT has the disruptive potential for comprehensive innovations in the insurance industry. Therefore, to take advantage of these opportunities, our approach suggests a strategy consisting of analysis, design and implementation including the recommended capabilities as basic prerequisites to get ready for your IoT business (see Figure 2).