Strategic change and future direction
There is a drive for efficiency in insurance markets, accompanied and enabled by changes in the way that data is captured, processed, stored and shared. Digital innovation and data sharing strategies are designed to reduce costs and develop new, innovative products, services, and distribution channels throughout the insurance value chain1. With the competitive threats of new entrants into the insurance industry, i.e., AI start-ups and technology giants2, there is a strong incentive for established firms to respond and adapt to take advantage of advanced digital technologies and Artificial Intelligence (AI) applications3. Our recent survey of industry experts expects transformation to occur in 3-5 years but the panel of experts do not perceive an immediate threat over the next 12 months4. This is a classic example of the innovator’s dilemma for established firms in markets that are in transition, where clear benefits can be seen, but change is risky and the current markets are still profitable5.
Incumbent insurance firms face two key questions. What are the future data sharing and market scenarios, and how soon will these occur? In this article, we explore how data sharing is changing across the insurance value chain, introduce four data sharing models and market scenarios, and identify the imperative strategic choices facing the insurance industry.
Insurance markets are defined by the contracts agreed between the insured and the insurance firm6, and the exchange of information starting from the customer and then along the insurance value chain7, which is comprised of a set of different types of organisations connected together through relationships to form a market network. Figure 1 is a schematic illustration of how retail and business customers, insurance firms, brokers, e-marketplaces, re-insurance firms, capital markets and regulators are connected together in a complex insurance value chain.
Insurance policies require a diversity of data types, which vary between policies dependent on the particular insurance line, e.g., retail automotive insurance or property and casualty, and the purpose of the communication such as to share customer data, fraud prevention, or broker a deal between a customer and an insurance provider. However, some broad categories can be defined for all insurance markets:
It is well established that data sharing between separate companies improves operational efficiency, increases data transparency and accuracy within a value chain, and enables a much more effective, coordinated response to external changes in the marketplace (such as changes in demand, new product designs and regulatory requirements). The recent TECHNGI survey of insurance industry experts confirms that they see numerous significant advantages to better sharing of data in the insurance value chain. However, although there is strong evidence of innovation in bi-lateral agreements between close partners, the experts also confirm that there remain significant barriers to wider data sharing, which can be categorised into three groups4:
This raises the issue of how these barriers will be overcome, how quickly, and by whom.
In addition to the continuation of the current insurance value chain shown in Figure 1, which is a system characterised by the dominance of bi-lateral relationships, experience in other sectors suggests that four market scenarios could emerge: electronic marketplaces; smart business networks; data platforms & ecosystems (these would typically be controlled by a single, large technology firm); and data trusts.
01
Electronic marketplaces are digital platforms that connect multiple buyers and multiple sellers together, with fast and low cost switching between competitors9. In markets where the level of inter-dependency between the customer and the supplier is low, and where a strong ongoing relationship is not crucial to delivery of the service, then electronic markets are economically attractive. This is evident in the widespread use of price comparison websites for home and car insurance in the US and Europe, where a handful of dominant price comparison engines significantly influence new customer acquisition and competitor switching, which takes place at the point of renewal. A Business-to-Business (B2B) example is the insurance broker market, where Polaris’ ‘imarket’ is an electronic marketplace that connects broker and insurer systems to facilitate real-time risk assessment and price quotes for a range of commercial insurance products from competing suppliers10.
02
A Smart Business Network is a network of separate organisations connected together through a common set of strategic objectives and facilitated through digital connectivity and advanced data sharing11. This is a network of members, who choose to cooperate closely with each other. A key role in an SBN is the 'orchestrator' or coordinating organisation, which operates across the smart business network and determines the overall structure and membership of the network.
Data standards, such as Polaris’ e-trading standards and ACORD’s electronic standards, forms and software tools, facilitate the successful evolution of smart networks by improving operational efficiency, reducing the cost of sharing data with economic partners and define common business processes along the value chain. Importantly, SBNs are more flexible than vertical integration in supply chains, because new organisations can be switched into the network and those organisations that are no longer needed or choose to leave, can be disconnected. They offer more stability than an electronic market, so trust and continuity can be built into the relationships. New technologies such as blockchain could be used as part of the digital infrastructure to build a smart business network and there is early evidence of this happening with companies such as b3i facilitating close ties between insurance and re-insurance firms for the secure exchange of risk data, which is based on a common language.
In insurance markets, the candidates for the role of orchestrator are primary insurers, brokers and re-insurance firms. The role requires a combination of influence and prestige arising from an organisation's existing position in the insurance value chain, combined with a high degree of expertise and capacity for digital innovation and leadership.
03
A data platform shifts the centre of gravity for insurance data to a data platform, which might be managed by an ecommerce or automotive company, or a technology giant such as Tencent, Google or Apple. In this scenario, insurance data would be a part of a much larger ecosystem12, and the insurance service would be subsidiary to other services such as transportation and mobility, health, property services or e-commerce. Technology giants potentially have significant data and analytics advantages over incumbent insurance firms and could use them to embrace and integrate InsurTech companies into their platform, which would offer the specialised industry expertise.
04
A data trust is a legal and technological construct that enables the compliant, ethical and secure sharing of sensitive data among a network of data providers. Willis Towers Watson is actively exploring this concept and has piloted its use in the insurance sector13. A data trust can remove much of the friction from the commercial, legal and technological barriers identified in the TECHNGI survey. This is achieved through privacy-by-design as well as by identifying use cases with compelling commercial and social advantages from data sharing, e.g., the sharing of industry-wide claims data for fraud prevention and the sharing of loss data from natural catastrophe to build more robust risk models. The insurance industry is clearly advancing towards digital, data-driven products and services that personalize and disintermediate, while reducing cost and increasing expediency and value to customers. One of the biggest challenges on this digitalization journey are the scarcity of interoperable, high-quality data, and the technology skills of data science and machine learning skills. The Willis Towers Watson pilot demonstrated that a Data Trust incentivises members in a “Minimum Viable Consortium” (MVC) of data providers to share data and resources to solve a common problem, as long as there are clear policies for data governance, technologies that protect privacy (such as federated learning and differential privacy), as well as commercial agreements between MVC members that ensure the equitable sharing of value generated through collaboration.
In the TECHNGI survey, the panel of industry experts were asked about how and when data sharing in re(insurance) would change and over what time frame. The majority expected significant change in a time horizon of 3-5 years, but not within 12 months. They felt that the outcome is likely to be a mix of the scenarios outlined above. It is not clear which, if any, will dominate though the participants favoured electronic marketplaces. The time-lag of 3-5 years is significant in illustrating some reticence towards change, supported by a notable minority of survey respondents still seeing the status quo in 3-5 years. Our interpretation, however, is that there is a much higher level of urgency to change and that the inertia of incumbent firms is typical of the innovator’s dilemma5 and understates the disruptive effects of data platforms and ecosystems.
Existing bi-lateral data sharing agreements are valuable but do not fully exploit the strategic value of data from a market-level perspective. Vertical integration, i.e., common ownership of the insurance value chain, is not a viable option for insurance firms because it is too expensive and inflexible. Insurance firms should therefore embrace new market scenarios and develop strategies that take advantage of their existing knowledge and expertise, relationships, and data – this will entail a re-evaluation of core competencies and a new approach to business models that takes an external or market network perspective rather than being internally focused. The strategic focus for an individual insurance firm should place much more emphasis on building product differentiation through analytics, rather than from the proprietary ownership of raw data. This implies a need for increased data sharing and investments into data science and Artificial Intelligence (AI) skills.
Electronic markets work effectively in those areas where products are standardised and there is no need for close collaboration between buyers and sellers. Electronic markets will therefore continue to be successful in consumer markets where insurance services are standardized and well understood, e.g., comparison websites for automotive and house insurance. There are also standardised marketplaces in B2B markets where the concept of open data is likely to thrive. This is where the industry as a whole agrees to share information and not use data ownership as the basis for competition. For example, OASIS is an open hub that facilitates the free sharing of environmental and risk data, supported by a common modelling language. The common good and the need for a better global understanding is more important than the competitive advantage of a single firm.
In specialised consumer markets such as health, and B2B relationships between insurance firms and reinsurance, and in the capital markets, smart business networks will come to the fore. SBNs require close alignment between their constituent members, and the role of the ‘orchestrator’ is therefore crucial. This is a principal organisation that builds the network around a set of common strategic objectives and provides leadership in areas such as technology strategy, innovation and the continuing evolution of the network, including its membership. Incumbent insurance firms are the natural contenders for this role though reinsurance firms and brokers could also have legitimacy in this role. SBNs offer an interesting opportunity for InsurTech firms because SBNs are inherently flexible in terms of new relationships, including technology partnerships, and are therefore more likely to embrace InsurTech services.
Data platforms and insurance ecosystems require huge scale. Ping An’s success in this area is built upon expansion into other market sectors, including healthcare and banking, and a large user base of over 500 million customers14. For European and US insurance firms, which do not have this scale or cross-sector presence, a data platform strategy would require a transformation of the existing insurance industry position and extensive cross-sector collaboration and integration. If a technology giant attempts to use its scale to enter the insurance market with a platform strategy then the incumbents would be forced to respond, either by attempting to block the move through superior offerings or by embracing the initiative and becoming an integral component of a much larger ecosystem.
Data trusts have significant potential to be applied to highly sensitive data such as loss and claims data, or information sharing between direct competitors. For example, in health insurance markets it may be necessary to combine genetics, insurance data and patient information to build AI-powered services. The synthesis of the data is necessary for the development and refinement of the AI algorithms, but the individual data owners may be unwilling or unable to share the data with other organisations, so a separate fiduciary body in the form of a data trust can overcome what may appear to be insurmountable barriers. A simpler example would be the combination of claims data from competing insurance firms to combat fraudulent claims.
The current ad-hoc and incremental approach to data sharing that is based principally on bi-lateral agreements and partnerships impedes the ability of the insurance industry as a whole to improve its profitability. The relatively slow uptake of data standards to overcome the barriers to sharing make insurance increasingly attractive to radical change from a new entrant such as a technology giant and the myriad of AI start-ups launching novel products and services. A combination of smart business networks orchestrated by incumbent insurance firms, electronic markets, and data trusts, combined with leading use of digital technology and data standards, enable a way forward for the insurance industry to successfully evolve from its current position, and build competitive and technology barriers to new entrants.
1 OECD. Technology and innovation in the insurance sector.
https://www.oecd.org/pensions/Technology-and-innovation-in-the-insurance-sector.pdf (2017).
2 Ralph, O. Insurance experts put premium on start-ups reaching the big time. Financial Times (2020).
3 Deloitte. From mystery to mastery: Unlocking the business value of Artificial Intelligence in the insurance industry From mystery to mastery: Unlocking the business value of Artificial Intelligence in the insurance industry.
https://www2.deloitte.com/content/dam/Deloitte/xe/Documents/financial-services/Artificial-Intelligence-in-Insurance.pdf (2020).
4 Timms, P., Hillier, J. & Holland, C. P. Increase data sharing or die? An initial view for natural catastrophe insurance.
https://eartharxiv.org/repository/view/2059/ (2021) doi:https://doi.org/10.31223/X5K313.
5 Christensen, C. M. The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail. The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail (Harvard Business School Press, 1997).
6 Denenberg, H. S. The Legal Definition of Insurance : Insurance Principles in Practice. Am. Risk Insur. Assoc. 30, 319–343 (1963).
7 Eling, M. & Lehmann, M. The Impact of Digitalization on the Insurance Value Chain and the Insurability of Risks. Geneva Pap. Risk Insur. Issues Pract. 43, 359–396 (2018).
8 European Commission. What are the GDPR consent requirements? https://gdpr.eu/gdpr-consent-requirements/(2020).
9 Malone, T. W., Yates, J. & Benjamin, R. I. Electronic markets and electronic hierarchies. Commun. ACM 30, 484–497 (1987).
10 Polaris UK Ltd. Products.
11 Heck, E. V. & Vervest, P. Smart business networks: how the network wins. Commun. ACM 50, 28–37 (2007).
12 Tanguy Catlin, Lorenz, J.-T., Nandan, J., Sharma, S. & Waschto, A. Insurance Beyond Digital: The Rise Of Ecosystems And platforms. McKinsey Co. 16 (2018).
13 Zarkadakis, G. “Data trusts” could be the key to better AI. Harv. Bus. Rev. (2020).
14 Ngai, J. Building a tech-enabled ecosystem: An interview with Ping An’s Jessica Tan. McKinsey Q. (2018).
This research is an output of the Technology and Next Generation Insurance Services TECHNGI project (www.techngi.uk) investigating the opportunities and challenges facing the UK insurance industry arising from new AI, digital technology, and big data. TECHNGI is funded by Innovate UK and the Economic and Social Science Research Council (grant reference ES/S010416/1) as part of the £20M. Next Generation Services Research Challenge.
