“The worst place to develop a new business model is from within the existing one.”

Clayton Christensen

At its most conceptual level, a business model is no more than a strategic configuration of the various components constituting a business and the logic of how this business intends to operate and make money. The application of technologies permeates all aspects of each business model and holds the potential to reconfigure existing business models or create entirely new ones, in a quick efficient manner leveraging the context of the consumer. In order to provide a sense of where the focus is across innovation models, we categorize them according to Nagji and Tuff’s¹ three-tier innovation ambition framework (see Figure 1):

1. Core – optimize existing products for existing customers (not the focus of this chapter);
2. Adjacent – entering adjacent markets and customer segments;
3. Transformative/Disruptive – creating new markets, and targeting new needs.

McKinsey’s three horizons framework, if you are familiar with this, is equally interchangeable.² Most companies anecdotally strive for the “golden balance” of 70-20-10 split for core, adjacent, and transformative initiatives in terms of innovation asset allocation. However, research has consistently identified that the financial returns are broadly the inverse of this, with core innovations returning around 10% of the long-term cumulative return on innovation investment.³ Given this, it’s important to highlight the current state of play in insurance innovation to ensure that we are channelling innovation efforts appropriately.

Core (or incremental) innovations are really focused on delivering product innovation improvements to existing markets and customers. A lot of the innovation coming from InsurTech during the course of the past two formative years has been targeting the optimization of online distribution (including price comparison and aggregators) and the digitization of inefficient processes, with startups building digital capabilities to augment the existing value chain. There is already a well-versed body of research on this area (value chain unbundling and digital distribution plays). As a result we won’t be revisiting this specific topic within this chapter.

Adjacent innovation is about stretching existing capabilities/products to serve adjacent markets or consumers. Notable business model trends in the adjacent innovation space include:

• Product disaggregation – InsurTech businesses such as Trov are tackling the short-term/on-demand insurance opportunity through a combination of new pricing strategies and new product features to create a great user experience and frictionless way to create a digital asset vault. This storage of digital assets is a powerful tool for client retention, which helps explain the emergence of startups focused on owning the digital vault, such as CBien or AssetVault. This presents adjacent opportunities like asset (digital and physical) life-cycle management. Qover are taking product customization further through personalization of both product term and specific non-life coverage options, white-labelled to commercial clients. Creating these insurance building blocks will be critical to driving Insurance-as-a-Service and could rapidly open up new non-insurance distribution channels as companies looking to cross-sell insurance can be as granular as designing insurance propositions for the customer “segment of one”. For on-demand propositions, a key consideration is to build sufficiently sophisticated techniques to avoid fraud as it is difficult to track/prove that users haven’t “turned on” insurance coverage after the event. Fundamentally, however, the core product is the same.
• “Robo broker” – insurance advice and broad risk assessment is a clear extension of price comparison and automation of the role of the broker (e.g. Wrisk or Brolly). In effect it enables a digital consolidation of siloed insurance products into a single orchestrated view. Over time this should extend to catch up with the robo investment business model innovations that are heading towards generating insights that are timely, contextual, relevant, and personalized. It’s not difficult to envisage even predictive insights about an individual’s insurance requirements or risk profile. Artificial intelligence (AI) and neuro linguistic programming techniques can be used to understand consumers’ risk appetite and compare policy wording, which is critical to identifying and appropriately communicating risk overlaps/gaps. This baseline understanding sets the stage for robo insurance advice and appropriate product recommendations. This is already happening in the commercial space with offerings from RiskGenius, leveraging technology to identify wording gaps, and much more. Many consumer-facing robo investment models have struggled to gain sufficient initial traction and so have pivoted to business-to-business distribution, creating a hybrid robo platform that enables, rather than replaces, the broker to profitably serve the long-tail of clients – important given the customer acquisition costs, churn rates, and combined ratios of insurers today.
• Embedded “pull” insurance – Adjacent distribution/product bundling – these models are starting to emerge across the InsurTech landscape.  Commercial partnerships with banks, telcos, and ecommerce providers have existed for some time – good old-fashioned affinity models. The innovation opportunity is in deeply embedded insurance, effectively as a business-to-business insurance API or application programming interface, invisible to the end-user. For example, Tesla are embedding auto insurance in their cars in partnership with AXA, packaged as an extended warranty. Tikkr are seeking to embed on-demand sports activity insurance within the various sports apps on the market. These embedded models shift insurance from a mandatory purchase/sales push effect to an invisible or embedded purchase or sales pull effect – part of the package that creates a frictionless end-to-end experience. Creating this frictionless integration and a seamless user experience for the distribution channel partner to sell to their end-users are quickly becoming hygiene factors. Successful propositions will be those that can offer deep-embedded individualized offerings with minimal manual data inputs, leveraging AI/machine learning (ML) techniques and bringing together relevant open accessible data, as they establish higher switching costs, margin protection, and improved customer loyalty.
• Preventative risk management – these models are about creating user engagement or monitoring services proactively before risks arise. This opportunity is only now becoming possible at scale and in a cost-effective manner by leveraging data, the Internet of Things (IoT), and ML – bringing all of these together to add material value to the customer and reduce the insurer’s risk. Great initial examples are emerging such as automated vehicle or driving behaviour diagnosis (e.g. Amodo), tracking health and activity status (e.g. LifeQ, Quealth, Fitsense), monitoring household water leakages (e.g. Leakbot, by Homeserve and Aviva), or industrial machine monitoring with Hartford Steam Boiler. This model requires an adjustment to the existing value chain as you build out the ecosystem of relevant service providers and adopt a preventative operating and underwriting model – one that is based on real-time/forward-looking, and potentially even predictive, data; a very different world from most incumbents today. MMI Holdings’⁴ “Multiply” or Discovery’s “Vitality” products are well-established insurance propositions in South Africa that have pioneered preventative models in the life/health insurance markets. For the incumbent, such business models drive user engagement and a positive impact on customer acquisition, retention, and cross-sell rates, but also could dramatically change the risk pools as we know them today. The challenge for startups is that replicating these ecosystems often requires a critical mass to bring together various service providers so the entry point may be to enable incumbents to adopt preventative models, as Sureify is doing, for example. A key consideration for the insurer is to avoid too stringent risk screening that leads to adverse selection. Abusing the asymmetric information advantage that the insurer holds on the basis of client-permissioned data sharing (i.e. unknowingly using client data against them for ulterior motives), while balancing the insurers’ growth trajectory, is a difficult tension and conflict that will need to be actively managed. It will be a cost-benefit trade-off between acceptable balance sheet risk and overheads associated with proactive risk management actions. Equally, adverse selection would have a regulatory or political risk too if you create pools of uninsurable, or risks that are out of access stopping people being able to get to work, or cover their homes.

Transformative (or Disruptive) Innovations are about creating new markets and solving (sometimes latent) customer needs. Typically, such innovations are difficult to achieve as the dominant business model design is not predetermined so it is a highly risky form of innovation. It will eventually cannibalize revenues and market shares of incumbents as many disruptive innovation strategies start by gaining a toehold in an adjacent or underserved customer segment deemed unattractive to the large incumbent insurers. However, they quickly move into prime customer segments with product extensions. These sorts of innovations may also appear too small or too slow to value to meet investment criteria of traditional (larger) businesses, which are often more interested in innovations that hold the potential of a bigger financial contribution to the bottom line faster (and at apparently less risk). This thinking often closes the door to the consideration of low or new market disruptive ideas. Examples include:

• P2P or “digital mutual” models. These are deemed to be disruptive as they remove insurance incumbents from the equation in many cases and democratize access to insurance, potentially acting as an investment asset class over time for institutional and retail investors. The irony with P2P models that they have existed for many years in an offline version called ROSCAs (Rotating Savings and Credit Associations) and have been used to great effect in Africa. However, technology has empowered these models to seek scale. Various digital models have developed, such as relying on shared insurance requirements of similar groups of people to negotiate better terms (e.g. Bought By Many) or platforms that are self-governing, self-selecting risk pools (aka risk pool unbundling) of individuals who typically rely on affinity groups and some form of user engagement to make the unit economics work in terms of customer acquisition costs or keeping fraudulent claims down (e.g. Insure-a-thing, Friendsurance, Guevara, So-sure, etc.). Users typically pay monthly into a risk pool and are rewarded through some form of incentive (commonly cashback) if no/minimal claims are processed. An enabler for this trend is the granularization of risks (enabled by IoT/smartphone data). A key hurdle is the inherent self-selection bias as many of these platforms rely on affinity groups and trust between members to help keep fraudulent claims to a minimum, which is difficult to do at scale. These also require some level of user engagement, which is hard to achieve outside of customer segments that are extremely price-sensitive as insurance is not an engagement-driven product today.
• Preventative models 2.0. This evolution sees a transition to real-time data and proactive suggestions/remedial action in a mobile-first distribution channel. The shift is from an IoT infrastructure “connectivity” focus (“connected home”, “connected car”) to a focus on “outcomes” via platform-enabled ecosystems. However, what if we imagine a broader preventative model, for example encompassing all forms of transport in an ecosystem model? A consumer could buy one travel insurance policy, “Citymapper v2.0”, with an insurer orchestrating an ecosystem of transport options (car/bike rental, taxi, public transport) in the event that your primary mode of travel isn’t available to get you to your destination. You are automatically presented with a pre-funded alternative transportation option based on the best risk-reward ratio or your policy tier. So, in effect, the user is buying peace of mind that they’ll reach their destination. Companies that can help enable this model include Concirrus.
• Two-sided platforms. Health insurers already adopt the structure of a two-sided platform model, albeit not in a digitally-enabled manner. Platform models work most effectively when they impact at least two variables positively – typically price and availability/variability of an offering. There is wider applicability of two-sided markets in general insurance as there is a market-wide under-utilization of balance sheet capacity that creates an opportunity to aggregate the supply of capital on the “producer” side, from a wide range of institutional investors. Global digital platforms (Facebook, Google, Amazon, etc.) can also compete given their massive online distribution and reach. Reinsurers are also actively experimenting with digital to customer propositions through a managing general agent (MGA)⁵ proxy. The differentiation is mostly needed on the consumer side of the platform to address poor customer satisfaction and product innovation efforts, otherwise consumers will remain price-sensitive and such a market with one variable (price) will drive commoditization and margin erosion. To limit commoditization risk on the platform supply-side (producer), a critical success factor is exposing the buy-side (consumer) to differentiated solutions, perhaps with advisory layer/personalized recommendations to avoid becoming an aggregation play and just competing on price.
• Freemium models. The insurance model is effectively a reverse freemium model as everyone pays for the benefit of the few (who claim). An interesting example is the Tigo and Bima partnership in Ghana, scaling to over 1 million users before offering a paid upgrade option to double their coverage. A user’s mobile phone credit was the proxy for their assessment of risk. With an initial 60% conversion rate, this has proved to be an attractive model. Low acquisition costs make this an attractive growth model but a key consideration is ensuring you have the balance sheet capacity to cover the claims risk exposure and well-defined upsell/cross-sell strategy.
• Open platform business models. Insurers can expose parts of their business to the open market to benefit from talent, specialized skills, or IP (outside-in) or expose any world-class capabilities to the market for them to collaborate or build upon (inside-out). A good area of opportunity for startup collaboration is to enable insurers to conduct an assessment and quantification of new forms of cyber risk threats and incorporating new forms of data into existing risk and pricing models. EigenRisk are an established independent software vendor in this space with startups AdaptReady and ThreatInformer offering useful risk assessments already.

In summary, there are broader trends that will influence future business models:

• Asset sharing. As we move to a world of asset sharing, the complexity of managing risk between parties can become more complex, specifically who and what you are sharing and when.
• Data ubiquity. Proliferation of real-time will eventually shift focus to forward-looking risk models and will enable more companies to move from mono to multi-line insurance. Ultimately the most important thing here for insurers is the shift from product-centric to customer-centric cover. Insurers will likely need help making sense of how to assess/score risk from unstructured data sets/IoT.
• Alternative risk models. Insurance industry barriers to entry fall as the dependency on long-term historical risk models declines and the focus moves to forward-looking, data-driven models. When we reach this inflection point, insurance becomes an investable asset class for many institutional investors seeking alpha. Hedge funds are well placed, as modelling risk-adjusted returns on a diversified portfolio of policies based on geography and peril bears similarities to hedge fund portfolio construction. A recent partnership between quantitative hedge fund Two Sigma and insurers AIF and Hamilton suggests we’re already seeing the line blur between credit risk and insurance risk pricing.
• Future of mobility. Self-driving cars/fleets remove human error and will make auto insurance largely redundant. Business models will need to evolve to create B2B commercial insurance-warranty hybrid or microinsurance products. The nature of risk also evolves to one that is systemic and down to technology system errors.
• Future of work. The gig economy continues to grow, placing stress on credit products, insurance, and pension schemes. Future products need to reflect a blend of personal and commercial liabilities for these “businesses of one” – a massive opportunity to get SME right.
• AI and automation. The opportunity is the automation of increasingly complex tasks, e.g. underwriting, which provides potential for cost savings and lower risks. The threat takes the form of AI-enabled cyber risks, coupled with IoT, that present macro-systemic risks for which we as society are not prepared. As an example, IoT and automation technologies applied to smart city use cases (e.g. monitoring the structural health of city buildings/bridges, providing access control to restricted areas, and controlling traffic flow) offer great benefit to society but present a real threat if visibility of this data and control of automation protocols fell into the hands of individuals with malicious intent.

No matter what the future holds, there are some fundamental changes we can be certain of – given what we’ve seen in other consumer-focused industries. Insurance will be data-driven and there will always be a need to manage risk even if the characteristics and bearer of the risk evolve. As digital weaves its way into insurance, adjacent industries will converge and value chains reconfigure. The pace of change enabled by technology means the biggest global insurance companies and new business models are unlikely to exist as they do today. Ultimately, whichever companies and dominant business models win, they will need to delight the consumer, influence behaviour, and solve for the original intent of insurance – peace of mind and financial resiliency. Whether we call it insurance or not is a different debate!