Is the Insurance Industry about to Change Radically?

By May 6, 2015February 28th, 2017Blog

There is no doubt that insurance is a massive industry. Net premiums for life and property & casualty insurance in the US alone are estimated to be worth around $1,000,000,000,000 (yes that is twelve zeros). But this industry has barely been affected by the massive software-driven changes that other industries are experiencing.

At its core, the insurance industry is impacted by the same underlying trends as the banking industry:


Online price comparison sites have brought unheard of transparency in pricing. In the UK market, 56% of consumers declared having used a comparison website in the last 2 years. Google’s move into the insurance arena therefore comes as no surprise. In the US, a growing number of companies are looking to play this role, positioning themselves as brokers. Notable examples include Coverhound and Policy Genius.

That transparency however, is starting to expand beyond just pricing to also encompass how an insurance policy covers an individual’s risk at granular level. In the mass market, insurance products, while relevant to the risk profile of the individual from the perspective of the insurer, tend to lack detailed, granular-level information from the perspective of the customer. Companies such as Trov (disclosure: Anthemis is an investor) are helping to create customisation in home insurance by allowing coverage of key items, vs. having a policy with a coverage set with averaged amounts. Metromile is another example in the motor insurance industry.

Impact of Digital on Distribution

The Insurance Industry, due to its specificities, has evolved to a reseller model (whether captive or independent). However with online distribution channels increasingly becoming the primary channels, will the insurance agent suffer the same fate as the bank branch? As digital distribution channels prefer scale, it can be expected that markets with a large number of small broker players will face intense consolidation going forward (just look at the rapid consolidation of the UK insurance market). The success of direct distribution models in some countries, including China, can also be be a hint to what is coming next. Will the brokers be disintermediated? Or will they find new value and add activities to justify their existence?

Source: MCKinsey Global Insurance Industry Insights

As we have seen in other industries, it is very difficult to turn a legacy company into a digital-first one. It may sound cliché, but the innovator’s dilemma effect is at play when most of your current revenue base depends on traditional distribution and management methods. We are just starting to see the emergence of new digital-first carriers. Companies like Oscar are leading the pack. Oscar’s mobile experience not only reduces its operational cost but also helps redefine the link between carrier, customer and physician (not to mention affecting the levels of risk that are potentially attached). Oscar also provides a flexible platform that can adapt to emerging new technologies such as wearables.

 From a venture capital perspective it is an expensive proposition with an important part of capital required for regulatory requirement. Taking the long view however, one could argue that these balance sheet companies may become more attractive in a regularised rate environment.

Beyond that, if you think of insurance as a class of exotic financial products, one has to wonder: could we see a larger scale disintermediation (think banks and alternative lending) that involves traditional carriers and alternative insurance funds?

Big Data

Insurance is in many ways historically the big data business and in trendier places, one might be able to call actuaries data scientists. One of the historical assets of insurance is in its past loss data. However with the exponential growth of sensors in our world, new sources of data are emerging (see Metromile as well).

Source: KPCB

Additionally companies such as Google, Facebook or Amazon are more likely to have attracted top data talents compared to traditional carriers. A good example is the success of Climate Corporation (a former Anthemis portfolio company), founded by an ex-Google employee and acquired in 2013 by Monsanto. By leveraging publicly available weather data including the large network of weather stations maintained by NOAA in the US and combining it with a modern big data infrastructure built from the ground up, Climate Corporation was able to provide instant weather insurance quotes.

In the B2B insurance world, for the most part, Lloyds of London still relies heavily on email and excel. When databases exist, they are often siloed and sit in technology that makes live manipulation of data very difficult or impossible. 

Companies such as Quantemplate (disclosure: I am a non executive director and Anthemis is an investor) bring data efficiency to players that have been struggling with it. The ability to manipulate complex data has the potential to reinforce an algorithmic approach on risk in these markets.

However, perhaps even more than banking, I believe the insurance industry will be facing fundamental challenges in the near future. One way to think about it is perhaps to start again from the definition of insurance itself.

Insurance is a form of risk management primarily used to hedge against the risk of a contingent, uncertain loss. (Wikipedia)


The combination of an increased number of sensors and computational abilities has the potential to have a significant impact on existing pools of risk. The car industry is a typical example. Statistics tend to show that the human factor a.k.a the driver is the first (by a large percentage – 94% in the US) cause of accidents. Looking further into the statistics the first two reasons are recognition error and decision error for a total of 75%. While fully automated cars may still be a while in the making, machine assisted driving is on the verge of becoming available in the mass market. Whether via radars or sensors, self-braking mechanisms are being deployed by car manufacturers.

To put this in perspective, motor insurance represent about 30% of overall premiums for P&C insurance in Europe. With lower collision risks, premium would be expected to lower as well (and the additional cost of electronics onboard will not compensate for the difference). With all things considered, one of the underlying questions is whether with lowered premiums and less variability per driver, the insurance could be embedded in the cost of the vehicle itself. At what point do insurance and guarantee start to look more and more the same?

As software is having a greater impact on our everyday lives, the question on risks attached to it, whether digital reputation, or impact of software failure in the real world is becoming increasingly important. Will risk shift towards software providers and away from individuals and traditional companies? To take it a step further, should Google for example or the user be insured for a Google Car?


With more and more sensors accompanying us daily and measuring the way we move, eat and feel, is the level of contingency on an individual’s health shifting? There are several experiments by insurance companies in leveraging wearables for prevention in health plans.  Oscar is running a specific program with Misfit and Vitality’s insurance program has a strong focus on prevention. From a strategic point of view, it makes sense for insurers to run these types of programs as they might benefit from an indirect self selection process, assuming people wearing wearables are more health conscious than the average population and have less exposure to certain risk (an equivalent of the original Progressive strategy in car insurance).

Alongside sensors the progress in genetic testing (speed and costs, as well as the increased ability to run large statistical research) will also ask for a clarification of the role of insurance / and who should provide it (educating myself on this topic so leaving it at that for the moment)


If there is an industry that has no doubt on the reality of climate change, that’s the insurance industry. One of the key topics of the Geneva Association (one of the leading think tank of the insurance industry) is Extreme Event and Climate Risk. For a large part the insurance industry is relying on their historical data to derive their model around risk. However if conditions are changing rapidly, these models have a risk of becoming less relevant. This is forcing the insurance industry to reevaluate the way the work, for example fostering the creation of Open Source models and platforms. The OASIS loss modelling framework, fostered by Climate-KIC and the UK Knowledge Transfer Network, is an example of this type of approach. More generally, the emergence of cross industry data sources, models and the infrastructure to support it bear the promise to massively affect the way the brokerage and insurance industry is operating.

Finally, some of the early models of insurance were based on the sharing of risks and rewards by a community of people (in the mutual insurance model, policyholders co-own their insurance company; Benjamin Franklin “launched” this model in the US). In the last years we have seen, via the spread of the Internet and mobile technology, that online communities have now reached an unparalleled scale. Combining the two to reinvent the mutual insurance is therefore bound to happen.

The early players in this field for example Friendsurance and Guevara, have both started from the brokerage spectrum of the industry (as it is a much easier entry point from a regulatory / capital point of view). My way of framing what they do is that it is a form of arbitrage on deductibles. By pooling an equivalent of a low deductible insurance subscription but effectively subscribing to a high deductible insurance product, they are creating a reserve pool of cash for the group to manage first claims. Pool creation and distribution of risk among pools, effective claim management, customer acquisition journey are the type of challenges these models are facing.

The emergence of blockchain / distributed application technologies has the potential to massively increase innovation in this space. A blockchain-based mutual insurance could not only distribute risks and ownership among members but also some of its infrastructure and logic including capital contribution. If you are thinking of doing something in that field, Eris Industries (disclosure: I am a non executive director and Anthemis is an investor) are the people you should talk to!

 The insurance industry is about to change radically and rapidly, creating a massive opportunity for innovative companies to emerge and improve massively on the way insurance is built, bought and experienced.

 Originally published on Tekfin.