The world is awash with data, but it's still surprisingly hard to find reliable statistics on what is happening in and around insurance. If you’re creating a business plan, sizing up a market opportunity, creating content or assessing an investment, finding the data to make your case can be hard. Or expensive.

I’ve been building and reviewing business plans for years and come across some great free resources to help me along the way. It was time to catalogue the sources, and having done that I thought I may as well share them with the wider world -- so here they are.

Some of the information is a few years old -- and, as companies increasingly become aware of the value of the data, some of these resources are disappearing behind paywalls. Use with care, but grab the data while you can. Imperfect facts are a better guide to decision making than foggy or fantasy forecasts. I hope some of this data will help you make better decisions -- and convince others to follow your advice.

Health warning: this is quite a long article -- over 2,500 words and lots of links -- if you have a low attention span I suggest skimming until you find something interesting rather than reading every word....

Premium Information

Insurance premium information is one of your data essentials. Understanding how much premium a company is writing, or is being generated by a class of business, is essential for market sizing. If you are building technology or selling data, looking for guidance on how to set prices or on how much your product is worth, then premium is often a good proxy for what you might be able to charge. Using a percentage of premium to figure out what your product may be worth to your buyer is a great sanity check. For example, in my prior life in catastrophe modelling, 1% of reinsurance premium was a good way of validating fees for an individual prospective client. At a country level, it's a good guide as to whether to invest in building models at all.

The most detailed premium information is provided by A.M. Best or S&P, but prices start at around $10,000. If you work for an established insurer or broker, your employer may already be licensing this data, so ask around. Both companies do provide summary and aggregate data to third-party publishers for free, so, if you can’t get the real deal, then the free versions may be enough. I cover these below.

For broad market sizing, try the Insurance Information Institute (III); it has useful headline data, showing for example that U.S. insurance industry net premiums totaled $1.32 trillion in 2019. You will also find the top 10 companies by different insurance sectors here. If you want to go deeper, try the searchable database of the top 100 U.S. insurers at reinsurancene.ws. It has the added benefit of links to recent news about the insurers themselves.

For who is who in global reinsurers, Atlas has a list of the top 50 reinsurers extracted from data published by A.M. Best. Interested in more obscure or specialist data, such as the top 30 insurance companies in Saudi Arabia? Atlas is also a good resource.

Statista is a subscription service but offers a taster dish of free data, such as its pie chart of global P&C premiums by region. A nice added touch is the ability to download the data as a spreadsheet. By the way, keep an eye on Statista's chart of leading insurance brokers, one chart that changes significantly every couple of years as consolidation continues. A Statista license isn’t expensive, but the data is limited compared with what you get from A.M. Best or S&P.

For a U.K. focus, Insurance Post has also just published a comprehensive table of the top 100 U.K. insurers for 2020. The data is behind a paywall, but Insurance Post offers a free trial, so maybe you can try before you buy -- or if libraries open again and you're in London, pop into the City Business Library and get December's print version. NSinsurance has some summary data on the top nine insurance companies in the U.K., which may be sufficient.

Unpicking the maze of the Lloyd’s market is a full-time job in its own right. Lloyd’s publishes some excellent detailed information on each of its 110 syndicates but charges £3,000 to non-members. At one time, this data was also available in business libraries. Atlas has a list of the top 20 syndicates from 2018, which covers those over $600 million annual premium, a good starting point if you are looking for potential clients or wondering who is driving the performance at Lloyd’s.

Comprehensive list by company type

Many of the insurance market associations have full lists of their members. They tend not to have financial metrics, so it's not possible to separate the big from the small, but the lists are useful nonetheless. The MGAA association in the U.K. links to members, as does the American Association of Managing General Agents. Insurance Post published a list of the top 50 brokers in the U.K. in 2016 – there’s quite a bit of consolidation going on in that market, but this is a good starting point. Head over to Insurance Age for the top 100 independent brokers from 2020 by revenue bands. A bit clunky in the way it is presented, but good enough for many purposes. For the U.S., Business Insurance offers you 2018 top U.S. broker firms in a table (no need to squint, click to read).

Insurance coverage

Most of the time, you will want to know what types of insurance different insurers offer. Getting consistent data, by line of business, by premium, by insurer, is hard or impossible without paying for it. In some industry sectors, particularly where there are only a small set of participants, such as marine or cyber, it can be easier to get information. Sometimes you need to come at this sideways, though. Lloyd’s list, for example, provides a list of the top 100 influential people in shipping, so it's possible to extract companies from that. There’s a lot of other free information on that site if you are interested in marine.

If you have good eyesight, the Bank of England’s List of U.K.-authorized insurers is available, as compiled at January 2019. This gives a split by different classes -- albeit at a high level.

Lloyd’s, of course, writes a wide variety of classes of business. You can find out which managing agent writes what business free from the Lloyd’s website. It’s ordered by managing agent, and with about 30 classes in each case requires a bit of data wrangling to figure who does what. If you are wondering who is going to license your cyber tool, for example, then this is not a bad place to start. It's possible to get deeper into the individual Lloyd's managing agents and their syndicates with the report and accounts from Alpha Insurance Analysts. The most current forecast performance of individual syndicates against their business plans is provided by Hampdens.

Insurance spending

To learn how we consumers are spending money, you can drill deeper into insurance spending in the U.K. by downloading the excellent free data from the ABI in Excel – this provides information on average insurance spending, with variation by region, age, coverage type. This is where you will find useful insights such as that the average insurance premium of structure and contents per family, with a head of household age 30 to 50 years old, is £382. Bear that in mind if you are hoping to sell insurers data to improve their underwriting -- or bundle in an IOT sensor. There is not much margin in these numbers to spend on risk selection, pricing or hardware. You'll also find out that in 2018 U.K. insurers wrote £2.7 billion of commercial premium compared with £4.6 billion of domestic insurance premium, and how much they paid out in claims. Fascinating stuff.

Company-by-company level

To really understand what is going on at some point you will want to go deep into the leading insurance companies. S&P or A.M. Best will provide that if you pay. Alternatively, you’ll need to unpick the annual report and accounts. The good news is that these, such as the Aviva annual report, are getting more comprehensive each year as companies are required to provide additional information on regulation and risks. The downside is that also means more work to extract the information you need, and every company report is slightly different so comparing like with like can be difficult.

Country-specific

Munich Re provides some excellent research, with good graphics showing trends and data for countries. This chart is a good example, showing projected premium growth by countries over the next 10 years (and notice that China is forecast to triple in size). Swiss Re through its Sigma research is probably the most comprehensive research into themes and markets with great data and more excellent graphics. Axco provides detailed data on each country, and is well-respected for its research and access in local markets around the world. It is currently offering a free trial of its Market Intelligence database.

Emerging markets and emerging risks

Opportunities for tapping into new markets can open up potential both for insurers looking for new lines and entrepreneurs looking to provide new analytics. The “insurance protection gap” totaled $84 billion in uninsured losses in 2019, according to Swiss Re, so there is a lot of untapped potential. The reason there is a gap, though, is because pricing is tough, aggregation potential is high and premium levels may not reflect the true risk. Five years ago, Allianz forecast cyber premiums of $25 billion by 2020 (from $5 billion at the time), but today they are still lower than $10 billion. In June of last year, A.M. Best published a summary of coverage types, market growth, claims and the top 20 cyber insurers along with information on losses and numbers of policies in force, reproduced here gratis. This article by SpringerLink goes into detail around cyber insurance and is so cheap as to be effectively free.

New industries such as cannabis could hold great potential. According to New Dawn Risk, the legal U.S. cannabis industry would pay about $1 billion in annual premiums if it were insured to levels normal for other businesses, but insurers are generally keeping clear. For a wide range of reports and reliable data on emerging markets, dig around the Geneva Association website.

Funding and acquisitions

We all know that venture funding alone is not the perfect indicator of the quality of a company, but, if you are like me, I’m sure the mega investments get your attention. Investment funding is a good guide as to whether a company can start to pay for your services if you are selling, pay you properly if you are looking for your next job or will be around for a few years if you are a buyer. The go-to source for most companies is Yahoo-owned Crunchbase – much of the information is still free, including funding rounds (where these are disclosed), investors, a handy summary of what the company does and more. Check out, for example, how are our friends at Friss are doing, in the Crunchbase news items, including a link to my recent podcast with CEO Jereon Morrenhof. Be warned, though, that some of the Crunchbase data is out of date or incomplete.

The two best sources I’ve found to track recent funding news, and acquisitions, are the quarterly reviews by FT Partners and the quarterly insurtech reports by Willis (led by Andrew Johnston, my guest on podcast 97). The Willis report is simpler to read but has less information about individual companies. Both are in PDF format. It's good to see a U.K. company leading the field for 2019 Q3 funding (Brit-owned Ki at $500 million) – my podcast with CEO Mark Allan and James Birch of Ki is episode 84 on the Instech London podcast and one of the most popular.

You'll come across various lists of top 100 insurtech companies. Some are of dubious quality (listing companies that have closed is a common indicator of sloppy journalism). Most are useful for identifying what companies do. Don't pay too much attention to the order of the listing; many great companies, particularly those that bootstrap and spend less time marketing, operate below the radar of these listicles.

Insurance rates, up and down

Each of the brokers, Aon, Marsh and Willis, provide its own assessment of the state of the market and how it's changing year to year. Aon Thought Leadership provides a comprehensive library here, and Guy Carpenter offers its Risk Benchmark report.

Other interesting sources

Catastrophe bond issuance is becoming a significant part of the global reinsurance market, with around $100 billion of bonds outstanding. Artemis deal directory provides an excellent directory of all the bonds that have been issued with details of the issuer, size and coverage. If you want a comprehensive view of parametric insurance, and the companies to watch, download our free InsTech London Parametric insurance report. Look out for more of these reports from us in other key areas in 2021, starting with Robin Merttens on e-placing platforms, followed by Mathew Grant (that's me) on location intelligence.

The Cambridge University Risk Centre creates an annual 100 cities risk index, assessing the impact of 24 different natural and man-made hazards. More comprehensive reports are provided on the site for individual areas, such as cyber. The university has strong links with insurance, risk managers in corporations and modelers, and the content is usually excellent. The 2019 report had pandemic as the fourth largest loss, with $49.9 billion GDP at risk.

Coverager used to provide a lot of great free data, but much of it is now available to subscribers only – fair enough, as Shefi and Avi need to get paid for all their great research. There is still some decent information on the website, though, and maybe some of the free content is still in there somewhere. London-based Oxbow Partners do a good job with their annual insurtech 25 companies and have useful data here, but beware of the sell-by date given how fast things change in this space.

By the way, if anyone from VentureScanner is reading this, can you please update your insurance technology map -- the reinsurance category, in particular, is about five years out of date and still lumps in multibillion-dollar global reinsurers with zombie startups. Not all free data is good data.

Going, going, gone…

I suspect we’ll have less choice in a year’s time, so grab this data while you can. This list is far from definitive, so please, if there are any other sources you know. let us all know by adding them in the comments.

Finally, at InsTech London we have over 160 sources of information, currently all still free at www.instech.london with podcasts, interviews, reports and insights. To keep track of what we are discovering each week -- including my weekly "Free Report Worth Reading" and "Other People's Podcasts" -- then sign up here for our hand-crafted newsletter delivered to wherever you are at 7.00am GMT every Wednesday.

A Breakthrough in AI Paul Carroll, Editor-in-Chief of ITL You may have seen articles last week about a breakthrough for artificial intelligence in medicine that managed to be arcane and exciting at the same time. Google’s DeepMind solved a 50-year-old problem related to protein folding — news only for geeks, right? Not so: The solution opens up all sorts of possibilities for understanding the inner workings of the human body and for rapid development of drugs. What I haven’t yet seen explained — amid all the speculation about just how many Nobel Prizes in Medicine will spring from the work — is that the type of AI that DeepMind developed to solve the protein-folding conundrum should also provide breakthroughs in insurance. This type of AI can take dead aim at some core issues in insurance, especially in underwriting and claims...  continue reading > The Future of Blockchain Series Episode 1: Usage in Personal Lines Blockchain has incredible potential to impact traditional business functions and inspire new innovative opportunities Learn More SIX THINGS Smart Contracts in Insurance by Ivan Kot Smart contracts will likely be used first for simpler insurance processes like underwriting and payouts, then scale as technology and the law allow. Read More Time to Try Being an Entrepreneur? by Kate Terry With businesses cutting back, many are asking that question. But there are huge misconceptions about how to think about the issue. Read More Surging Costs of Cyber Claims by Thomas Kang With home-working widespread because of COVID-19, security around access and authentication points is critical. Read More 4 Stages of Dominance in Performance by Kevin Trokey Chances are, you have natural gifts. However, many of the skills you need must be developed, nurtured and maintained intentionally. Read More Vintage Wine? Sure. But Vintage Tech? by Ian Jeffrey Legacy systems that have evolved over long periods can be bloated and far less efficient and cost-effective than more modern technologies. Read More Do Health Plans Have the Right Data? by Denise Olivares Health plans strive to deliver efficiency and great customer experiences and improve care outcomes. But what data are they missing? Read More GET INVOLVED Write for Us Our authors are what set Insurance Thought Leadership apart. Get Started Partner with Us We’d love to talk to you about how we can improve your marketing ROI. Learn More SPREAD THE WORD Share Share Tweet SUBSCRIBE TO SIX THINGS

Insurers are not big polluters in their own right. Nor do they typically have lots of physical assets at risk, except indirectly through investment portfolios either now or in the future when economic transition raises the possibility of stranded assets.

Yet the impacts of climate on insurance operations are only too evident. Losses from more frequent flood events and other climate-related events, such as the wildfires that have ravaged parts of the U.S. and Australia in recent months; changing attitudes toward insuring and investing in high carbon industries; burgeoning regulation and moves toward mandatory climate risk disclosure; and external ESG (environmental, social, governance) ratings that increasingly reflect assessments of climate risk management - are all changing insurers’ risk landscapes.

With the PRA letter to U.K. insurers also setting the expectation that “firms should have fully embedded their approaches to managing climate-related financial risks by the end of 2021,” it’s relatively unsurprising then that climate change has been rising rapidly up the rankings of the perceived most dangerous risks to an insurance enterprise. In the most recent Willis Towers Watson Dangerous Risks Survey, for example, climate change rose from 53rd position in 2019 to 9th in 2020.

On the other hand, the up-side should not be ignored: Climate risk also brings new insurable opportunities and insurance can often be an enabler of innovation, allowing new technologies to be developed as risks are shared. Insurers that are taking steps now to better understand the risks and opportunities and planning for changes in their mid- to long-term strategies will be better placed to deal with them. These insurers will have built up a body of data, tools, analytical capabilities, processes and frameworks, with experience of learning and refinement, to avoid having to play catch up with the rapidly evolving regulatory environment as our collective knowledge of climate impacts grows.

Climate risk is truly multi-dimensional

Much as loss events grab the headlines, climate risk for insurers is truly multi-dimensional (see Figure 1). Potential ramifications that may not be grabbing the headlines yet could have potentially devastating consequences in years to come, such as sea level rise or threats that destabilize fragile states. Equally, new pathways for mitigating climate risk and resilience that don’t exist now could offer respite from threats and open up business opportunities.

The need for a multi-dimensional risk approach simply reflects this expanding diversity of climate risk drivers.

Even if we confine those to the current day, from one angle there are market factors, such as regulation and investors’ lengthening ESG agendas. From another angle, there is the societal pressure to consume less and reduce environmental impact. Then there is the role of science and advances in climate understanding and adaptation, together with mitigation technologies and what these tell us about the need to adapt collective behavior. Notably, many of the world’s central banks and supervisors, through the Network for Greening the Financial System (NGFS), have already upgraded their view on the financial risks from climate change. The risks from climate change are now increasingly seen as having "distinct characteristics," which means these risks need to be "considered and managed differently."

The potential impacts on operations are similarly diverse, not the least whether factors such as public policy and regulation may affect insurability of certain segments. Add in underwriting issues (risk assessment, pricing sufficiency/competitiveness), regulatory compliance (including solvency impact), capital considerations (risk accumulation for example) and emerging risks (and opportunities) – and you have a veritable cocktail of risk dimensions to consider.

ERM implications

In many ways, however, these risks are not new per se; they map onto existing categories of financial and non-financial risk such as credit, market, business, operation and legal risks that insurers have been managing for many years. But taking into account the vagaries of climate, the risks do present new challenges.

Specifically for ERM programs, they raise issues and questions that require explicit consideration:

• Governance, including the board’s role in providing oversight of climate risk responses and defining management responsibility for climate risk and ESG integration.
• Risk identification, identifying the key channels through which climate risks can affect the company and how these are articulated and monitored on a continuing basis.
• Risk appetite, including forming a view as to whether climate risk should be considered as a separate element or part of aggregate risk and how this will be implemented in practice.
• Risk measurement and reporting, including how to incorporate climate risk into financial risk models and reports and deciding on relevant metrics for decision making, a key element of Taskforce for Climate-related Financial Disclosure (TCFD) requirements, for example.
• Investment – how does the investment approach meet ESG objectives and respond to investor pressure to reduce or eliminate funding of high-carbon industries, for example?
• Reputation risk, including identifying public communications needs and a strategy for communicating a firm’s climate and ESG response.

And because all in turn feed through to strategic business considerations such as earnings, product development, long-term direction and acquisitions and divestments, having a solid understanding within the business of the connections between physical, transition and liability risks is increasingly essential. This also means that the risk and governance frameworks need to be holistic and that each aspect cannot be treated in isolation.

See also: An Early Taste of Climate Change Disrupting Insurance

Devil is in the details

Conceptually, this all probably makes sense. Where it starts to get trickier is getting into the long weeds of risk impact and mitigation. For that, quantification is key.

This requires proven analytics tools and methods that are constantly being refreshed to reflect the latest science and predictive climate change scenario datasets and the expertise to provide the context of how business decisions can affect potential futures. Typically, quantification will also entail a collective, systematic and open data collection initiative to capture appropriate data to represent the key risk-related attributes of assets and, equally importantly, to include the valuations needed to feed through into balance sheet and other decision-making views.

Examples of the types of outputs needed will include hazard and climate risk scoring and mapping, determination of hazard- and climate-adjusted financial losses and advanced modeling of current and future climate risks. And beyond the numbers, transparency of models, scenarios and parameters is also key to the credibility and flexibility of the approach.

Our view is that there are some key analytical building blocks in helping build understanding of climate risk. Even if these may represent a kind of analytical nirvana at the moment, principally due to lack of data, there are options. Drawing parallels with emerging cyber risk, many insurers relied on scenario analysis and a sort of risk disclosure statement to not only quantify risks but also to set risk appetite metrics:

1. Identify hazards – review of the existing portfolio for exposure to climate and natural catastrophe perils to establish the hazard levels.
2. Quantify current climate risk for key perils – modeling of the current portfolio of risks, taking into account the vulnerability of assets and the level of hazard with reference to past events.
3. Quantify future climate risk for key perils – modeling of future portfolios of risks for key perils at different times (e.g. 2030, 2050) and climate development scenarios. This should also consider the connections between perils – compounding and cascading risks are difficult to model, but they are the real world.
4. Identify opportunities to mitigate climate risk – identification and assessment of loss drivers and mitigation opportunities to help reduce the financial loss potential of climate change.
5. Determine transition risk and opportunities – evaluation of potential transition routes in line with modeling and taking steps to embed them within the risk framework.
6. Quantify transition risks – through breakdown of the top transition risks by region/climate scenarios.

As they become armed with this sort of information, insurers should be able to identify the regions and perils that are driving climate risk now and how this distribution could change. Critically, this capability will help to quantify and reduce the cost of climate risk and enable insurers to feed the results into reviewing and updating the risk appetite and management frameworks on a regular basis.

Given the evolving investment focus on the "social contract" and sustainable returns, the capability will also be increasingly important for being able to inform potential investors of both the impact of climate change on an organization and steps being taken by the business to reduce its climate impact.

This need has been accelerated by recent regulatory moves focused around reporting and disclosure, including proposals and consultations in some countries to make TCFD reporting mandatory sooner rather than later. Add to this the idea that COVID-19 may accelerate the broader appetite for ESG as financial markets look to build resilience to systemic risks, and there is an even stronger case for enhancing understanding and response.

The upside is that the positive reputational impacts of disclosure, enforced or otherwise, are likely to be more far-reaching than just compliance – working through this process provides a holistic stress test of strategic decision making and company direction.

Eye to the future

So where might the gaps lie? To be truly strategic, thinking about climate risk needs to properly address current climate risks and project five, 10 and 20 years into the future, at least. That means developing the climate trajectory scenarios and metrics (the areas incidentally where insurers say they expect to need most help, according to our TCFD survey) that are increasingly being demanded by various stakeholders to assess a company’s climate transition plans and contribution.

See also: COVID-19 Is No Black Swan

Not all companies will be equally affected, but it’s apparent that, in relatively quick time, climate will have to be a central component of ERM and strategic direction. Those running ERM programs at insurers are uniquely placed to ensure their companies are prepared to meet those rising and multi-faceted expectations of investors, regulators, employees, customers and other stakeholders.

Embedding climate risk into existing frameworks and ensuring boards are taking a strategic approach to the changes that are already happening, and those to come, will put companies in a position to deal more effectively with the threats and embrace the opportunities of a future low-carbon economy .

The unexpected, unprecedented events of 2020 have turned the world upside down. Like every business sector, commercial lines insurance has had to adapt and adjust throughout the year. Many commercial lines insurers have experienced significant financial hits from the pandemic due to increased claims, lower business volumes and the decreasing payrolls of their customers. Large reserves have been set aside, but the continuing uncertainty means that there may be long-lasting negative impacts to financials far into 2021. How much has the economic environment affected insurers' strategies and plans for 2021?

SMA's recent research report, 2021 Strategic Initiatives: P&C Commercial Lines, provides insight into how strategies are shifting. All indications are that the transformation that began several years ago will continue in 2021. However, some significant changes are occurring in strategies as insurers consider the new realities of the business environment, the risk landscape and shifts in the workforce. Prior to 2020, insurers focused primarily on level one transformation, aimed at business optimization and innovation in the context of existing business models. As 2020 approached, leading insurers were moving to level two transformation, emphasizing true innovation via new business models, new products and bolder strategies. Then the pandemic hit.

As commercial lines insurers plan for 2021, there is a movement back to level one transformation. Our research shows that business optimization continues to be the top driver of strategies, as it has been for the last several years. Innovation had risen to become a major reason for tech investment, but it has fallen significantly, and fewer executives cite it as a strategic driver. In terms of major projects, core systems and business intelligence initiatives continue unabated. These foundational systems are too important and have too much momentum to slow down. In addition, all types of digital projects are moving forward, and some are accelerating. And the increased emphasis on improving the agent/customer experience remains critical.

One of the most important things to explore in analyzing 2021 commercial line strategies is how the priorities differ for companies focused on the small commercial market and the mid/large markets. For example, those focusing on both small and mid/large markets still seek growth through their existing products, channels and markets. But insurers focusing on small commercial are more apt to seek growth through new lines and markets than those in the mid/large sector. Another example lies in distribution strategies, where small commercial is further along in expansion plans, while insurers serving the mid/large segment are still in earlier strategy stages.

See also: AI Investment in Commercial Lines

While everyone hopes for a more predictable, stable year in 2021, the prospects look uncertain, at least for the first part of the year. The scaling back on some of the more ambitious plans and concentration on operationalizing strategies from existing projects will serve the industry well in any scenario. At the same time, it is evident that leading insurers still plan to pursue the level two transformation as they tackle bolder strategies in distribution, product, underwriting, claims and operations. There are also likely to be new partnerships and new business models launched into the marketplace in 2021. All in all, the next year will be interesting and require regular monitoring of the external forces and the ability to adapt strategies and plans for success during these turbulent times.

You may have seen articles last week about a breakthrough for artificial intelligence in medicine that managed to be both arcane and exciting at the same time. Google's DeepMind research arm solved a 50-year-old problem related to predicting how proteins fold themselves -- news only for geeks, right? Think again. Understanding how these chains of amino acids fold themselves into 3D shapes, providing the structural components for the tissues in our bodies, opens up all sorts of possibilities for exploring our inner workings and for rapid development of drugs.

What I haven't yet seen explained -- amid all the speculation about just how many Nobel Prizes in Medicine will spring from the work -- is that the type of AI that DeepMind developed to solve the protein-folding conundrum should also provide breakthroughs in insurance. This type of AI can take dead aim at some core issues in insurance, especially in underwriting and claims.

AI is funny. It tends to be talked about as a single thing, but it's really a whole bunch of things, pushing against limits in a wide range of directions. And some of the progress is flashy without being all that important.

For instance, when IBM's Watson defeated the greatest Jeopardy champions in 2011, IBM talked about sending Watson to medical school. After all, if it could beat Ken Jennings, what couldn't it do? But Watson's breakthrough was in natural language processing, a great advance if you want to be able to talk to a computer but little help if you're trying to cure cancer. Similarly, when DeepMind beat the world champion at Go in 2017, the event made for fun headlines but not much more. The AI is terrific for any setting where there are a small number of rules and where the computer can play games against itself ad infinitum to optimize its approach, but how many real-world situations fit that description?

By contrast, what DeepMind accomplished in solving the protein-folding problem is of deep significance because the approach the scientists used -- known as supervised deep learning -- can be applied to so many business situations, including in insurance.

Without getting too deep into the details (which you can find in this excellent piece in Fortune, if you want to geek out like I did), the scientists faced a problem far more complex than businesses face: trying to figure out how a protein folds itself, in the milliseconds after it is created, based on a host of forces. While we've been able to sequence the human genome for more than 15 years now, you also have to know how the string of amino acids folds, because the shape determines so much of how the protein behaves.

Although a famous conjecture in 1972 said it should be possible to predict a protein's shape just from the sequence of amino acids in it, the computation had proved to be too complex. Instead, the shape of a protein had to be determined through a complex chemical process and, often, through the use of a special type of X-ray produced by a synchrotron the size of a football stadium. The process could take a year and cost $120,000, for a single protein.

(I realize I may be giving you flashbacks to high school biology and chemistry and perhaps some unpleasant memories, but I'm just about done with the science and am getting to the implications for insurance.)

What the scientists had going in their favor were two things: a sort of answer key, because of some 170,000 proteins whose shape had already been determined experimentally, and some coaching tips that could help the AI focus on the key variables.

That starts to sound like a business situation, especially, in terms of insurance, in claims and underwriting. If you want to train an AI to take over tasks, you have underwriters and adjusters who can tell you what the right answer is and who can guide the AI's self-training by steering it toward certain variables. Over time, that AI can become as good or better than a human at, say, looking at photos of the damage in a car accident and estimating the damage.

At least, that's how it worked for DeepMind on a much harder problem. On a scale where 100 is perfect accuracy, the previous best AIs scored about 50, well below empirical methods, which scored about 90. But in a recent competition in which AIs predicted the shape of proteins whose forms had been determined experimentally but had yet to be published, DeepMind's median score was 92 -- a computer prediction outscored that year-long, expensive, physical process. Importantly, DeepMind's AI can tell scientists how confident it is about each prediction, so they know how heavily to rely on it.

The immediate application for the DeepMind AI will, of course, be in medicine. There are some 200 million proteins whose shapes haven't yet been determined, and the AI can quickly go to work on those. (The required computing power is only perhaps 200 of the graphics chips used in a PlayStation.) Understanding the shapes will help researchers see what drugs might interact with which proteins, potentially reducing drug development time by years and lowering costs by hundreds of millions of dollars.

However, how this AI moves into the mainstream remains to be determined. DeepMind functions as a research arm of Google, not as a business, and has promised to ensure that the software will “make the maximal positive societal impact,” but you could hardly blame Google if it tried to recoup the development costs through charges to Big Pharma. Only once this AI filters through medicine will it, I imagine, spread to other business problems, such as those that insurance faces.

For me, it's enough to know at the moment that this sort of AI is possible, because that means that a lot of smart people will accelerate their efforts to bring supervised deep learning to insurance. While the wins at Jeopardy and Go were startling, the AI that solved the protein-folding problem will prove to be far more consequential.

Stay safe.

Paul

P.S. Here are the six articles I'll highlight from the past week:

Smart Contracts in Insurance

Smart contracts will likely be used first for simpler insurance processes like underwriting and payouts, then scale as technology and the law allow.

Time to Try Being an Entrepreneur?

With businesses cutting back, many are asking that question. But there are huge misconceptions about how to think about the issue.

Surging Costs of Cyber Claims

With home-working widespread because of COVID-19, security around access and authentication points is critical.

4 Stages of Dominance in Performance

Chances are, you have natural gifts. However, many of the skills you need must be developed, nurtured and maintained intentionally.

Vintage Wine? Sure. But Vintage Tech?

Legacy systems that have evolved over long periods can be bloated and far less efficient and cost-effective than more modern technologies.

Do Health Plans Have the Right Data?

Health plans strive to deliver efficiency and great customer experiences and improve care outcomes. But what data are they missing?

The phrase "WWJD," or "What would Jesus do?", became popular in the late 1800s after the widely read book by Charles Sheldon titled, “In His Steps: What Would Jesus Do?” The phrase had a resurgence in the U.S. and elsewhere in the 1990s as a reminder to Christians of their belief in a moral imperative to act in a manner that would demonstrate the love of Jesus.

Today, as we continue to struggle with coronavirus, economic crisis and social and political dysfunction, I ask, WWYD – What will you do? 

This is the whole you – the individual, the family woman/man, the parent, the breadwinner or the bread eater, the independent or dependent person, the believer or the non-believer, the investor, boss, business owner, employee, consultant, etc. 

The lucky ones and risk takers have options – others must merely play the cards life deals them or commit to a life of dependency. The choice isn’t always ours, yet the consequences too often are our future!

Look at the work-from-home movement. For many, this now is a necessity, for some a luxury and for some a little of both. Most of us are creatures of habit – now that many have gotten into the habit of working from home, will we be willing and able to go back to the office?

Are you prepared for the future as it will be – with transformational change (read, chaos)? Or are you still assuming a neat and clean and comfortable life with incremental change? Tomorrow includes pandemic, social and economic chaos, generational and VALUES transitions, “a house divided,” cultural differences, shifting demographics, the have-nots and the haves with different world visions and personal expectations, etc. 

See also: 5 Transformations for a Post-Pandemic World

The story was told of three GIs under attack in a foxhole. Things looked bad. All three were non-believers. One said, “I think we should pray, but I’ve never done that before.” The others were equally spiritually challenged. Finally, one said, “I never went to church, but I lived next door to a Catholic Church, and on Wednesday nights I could hear them praying in their church hall. Let me try to remember what I heard. Here goes: Under the B number 7, under the G number 48, under the I number 25… BINGO!” 

SMILE!

Then decide: What will you do?

As customers press insurers for the best possible rates -- especially in the wake of COVID-19 -- companies must map their position against competitors to be able to offer optimal prices. The market offers several online tools and platforms to compare general insurance quotes based upon the details that prospective customers provide. However, for insurers that wish to compare their quotes with competing products in the market, not just in terms of price but in terms of attributes such as coverage and benefits, these platforms and tools are insufficient.

A 2015 survey by Earnix of North American insurers engaged in personal lines, almost every company surveyed (93%) performs rate structure and rate competitive analysis. Other common types of competitive analysis include product features (79%), coverages and contracts (70%), financial metrics (58%) and underwriting guidelines (56%).  

To perform such multifold comparisons, insurers require access to a business intelligence solution that is interactive and allows users to analyze across different consumer segments and product/policy features. Such a business intelligence (BI) tool should not only be capable of extracting the quote information from various raters/agents/online-platforms/competitors but also consolidate this information so insurers can compare other metrics, by:

• Providing a clear understanding of the standing of client vs. competitor quote, by dollar amount. 
• Generating insights to develop predictive models that can assist in increasing the conversion rates of policies.
• Focusing on the binding and issue rates, to measure success.
• Assisting in cross-selling, marketing and targeting of customers from other lines of business.

This article provides a framework for developing a quote comparison BI solution and contains metrics for assessing competitiveness across multiple lines of business.

DISCLAIMER: Because insurance consumers often purchase home and auto insurance at the same time, while building a quote comparison BI solution it is essential to include the competitiveness of both lines together by combining multiple line of business reports into one. This entails setting up specific data warehousing processes in the background, which are needed to power up the visualizations in the dashboard/report output.  

Key Metrics to Use in This Rating Methodology

An insurer trying to compare the pricing of its competitors needs to establish a set of metrics to rank different quotes for a given consumer profile, created by analyzing past policy purchases and consumer details data collected by brokers.  

One of the several challenges is that the competitiveness information received for quotes belongs to a single business line. This information is sometimes categorized and not in absolute numbers. In most cases, a carrier gets the following values directly: its premium, low/median/high carrier premiums and rank for a quote. Thus, there is a problem with joining auto and the home quotes to form a combined quote.

While a carrier can determine its own combined premium (auto premium + home premium), it cannot do the same for the low/median/high carrier premium info. For instance, one does not know whether the carrier that quoted the lowest auto premium is the same carrier that quoted the lowest home premium. In fact, the combined low premium (set to low auto + low home) is, in most cases, going to be lower than the actual combined low premium. The opposite is true for combined high premium info; the estimate will uniformly be higher than the true high combined premium. The combined median should be higher than the actual combined median just as often as it is lower, so the overall median combined premium should be roughly in line with the actual combined medians.

The carrier performing the rate comparison can use different approaches to determine the best estimate of all the competitiveness measures used in the views of a single line of business. Some of the metrics used in this approach are:

• Ranks based on the premium dollar value quoted for both auto/home.
• Categories based on the ranks where the insurer is No. 1 in a category, and rank sorting among the top few ranks.
• Number of carriers quoted for that policy. 
• Difference in premium dollar value quoted for both auto/home and the low/median/average/high of all the premiums quoted.
• The average rank based on premium dollar value quoted for each line of business.
• Percentage of quotes issued. Issuing is the act of completing a quote and binding a policy (generating a policy number). The issue percentage (issue rate/close rate) is the percentage of quotes in which a policy is issued at the end.

See also: Best AI Tech for P&C Personal Lines

Ranking Insurance Quotes

The ranks given to competitors do not have the information about the number of competitors quoting for a given consumer persona (or a consumer). P&C carriers can adapt a new ranking methodology that identifies the most robust metric that combines rank with the number of carriers quoting for a given consumer persona/profile. For example, consider a scenario where a customer gets three auto insurance quotes from his independent agent. Here, to be ranked second isn’t as impressive as being ranked second out of 10 quotes. The following three rules can be used by carriers when determining ranks in their dashboards:

Expected win rate: Expected proportion of Rank 1 quotes.

Rank scale metric: This is essentially distance from “expected win rate.”

LOB filters constraint rule: Consider an insurance carrier offering home and auto and wanting to be able to offer them alone or in combination. The BI solution for price and rank comparison should provide filters for visualizing such combinations.

Analysis and comparisons

Geography-Level Analysis 

One can showcase metrics like numbers of quotes and conversion rates, etc. Analysts have access to line of business and geography-level filters like state and county, urban vs rural, etc.

Rank Level Analysis

Here one can show product performance based on ranks for different lines of business for metrics like quote vs. pre-quote.

Line of Business Analysis

Each line of business has specific attributes -- for example, driver age and vehicle age for auto insurance -- so one needs line-of-business-level views for auto and home and a combined graph.  

Competitive Market Analysis 

This involves relative quote ranking to understand the insurer position concerning competitors, analysis around quote distance from average market price and segmentation analysis to understand which insurers are incredibly expensive.

Price Elasticity Analysis 

This focuses on price sensitivity and customer profile. Segmentation analysis helps understand the pace at which the rates can increase or decrease to accept the change. Similarly, different studies can be carried out for the sales and marketing team to increase the retention of profitable customer segments.

Price Optimization 

Price optimization is more of a prescriptive analytics approach. A scenario-based analysis can help check price sensitivity by market segments and build a reusable platform for carrying out state and countrywide optimization analysis.

Quote & Coverage Optimization and Customer Segmentation 

An insurer may not offer the best price but may still bind the policy because of better coverages, or the insurer may have the best price but not win the business. These are the scenarios that a carrier needs explicitly to look into as they will help the carrier while quoting new customers.

Location-Based Analysis

Location has a significant effect on premiums. Differences in competition, state and local rules and cost of living account for this. Insurers take into consideration this information when they decide on the premium dollar value. For example, in a region with higher car theft rates, one may pay a higher premium than a place where car theft rates are low.

See also: Foundational Tech for Personal Lines

Conclusion:

With COVID 19, pricing has become more crucial than ever as it can become a great source of competitive advantage. A useful quote comparison tool that encompasses various metrics of comparisons and different data elements can lead to better customer acquisition and retention. The visualizations provide an effective way to stay on top of customers' expectations and price right to keep your organization ahead of competitors. Deploying data analytics to integrate home and auto and customer personas can help deliver the right products and the correct prices to existing and prospective customers.

However, to kick off this transition, it is crucial for insurers to carefully collate a list of metrics to compare, rank the insurance quotes to take into account the number of competitors and then perform the needed analysis. The best way to determine this is through a comprehensive, systematic audit of the current quote comparison process. A carefully planned audit will help insurers to: 

• Identify gaps in the current quote comparison process. 
• Identify the right personas and the right metrics.
• Prioritize the customer population who are more likely to expand the wallet share and purchase a policy.
• Build a single source of truth of quote information for effective pricing decisions

Leadership is a necessity in times of comfort and crisis. But now is the time for the insurance industry to lead by subsidizing the cost and distribution of a specific necessity: disposable face masks. Now is the time for insurers to be true to their respective brands by respecting the urgency of the present, so they may respond to the challenges of COVID-19 by presenting the public with ample supplies of personal protective equipment (PPE).

Leadership of this kind is a lifesaving measure, which also offers immeasurable savings. The alternative is wrong as a matter of morals and money, because when an act does not make sense, when an act violates common sense, the cost comes in the form of many dollars and cents; billions of dollars in medical bills for millions of patients nationwide.

Were insurers to subsidize one of the least expensive but most effective ways to stop the spread of COVID-19, were insurers to advance the issue of public health by promoting this issue on the outside of every face mask, were insurers to show their faces to the public, the reaction by the public would be huge.

An insurer’s logo would be a mobile advertisement. Truth in advertising would no longer be a contradiction in terms. Not when it would be impossible to deny what people could see: PPE in action.

According to Vitali Servutas and Brent Dillie of AmeriShield:

“Compliance governs the insurance industry as much as it inspires the industriousness of our commitment to public health and personal safety. By complying with the rules of the CDC and the Berry Amendment, we give hospitals, businesses and consumers a safe, affordable and convenient means of protection against COVID-19. Disposable face masks are essential to winning this fight, which is why insurers should support or subsidize the use of these masks for everyone.”

I agree with this statement because the words speak to a third “P,” patriotism. 

Helping Americans by increasing jobs for Americans is good for all Americans. This policy is wise, too, because it highlights the value of oversight and quality control. Put another way, what works well for consumers is a policy that works to expand wellness.

Insurers have every reason to support this policy, given the nature of the pandemic and the pandemic’s toll on the nature of how we live now; of how we live to survive, for now.

We need protection, yes, but we also need to know we share the same goal: that we are in this fight together, that we will hang together, that we are and will be stronger together.

See also: Insurance CEOs Spec Out a Post-COVID World

Insurers have the resources to achieve great things. If they summon the will to do this one thing, protect the public by providing the public with disposable face masks, then insurers will achieve greatness. 

History will record these things, just as people now living will tell future generations about the good works that make insurers institutions of greatness.

The public welcomes this moment.

Today, health plans (also referred to as payers) are busier than ever. They strive to deliver efficiency, great customer and stakeholder experiences and improve care outcomes. To do this, they need to use more data, and they have much data at their disposal. But what are they missing?

They can be missing key patient information; that is, they don’t see the whole picture. The insurance industry is no stranger to data gathering, coding, tagging, analyzing. In addition to a long history with data, in the recent decade or two the industry has seriously upped its game in terms of converting legacy data to newer usable forms and has upgraded systems to establish true master data management infrastructure (whether in-sourced or in/out-sourced). We have even seen a flurry of third-party data use and the occasional implementation of new ways to digitize unstructured data. All of this is here to stay, which is good for data purveyors, health plan analysts and application developers and for the business of cost management and reimbursement. Providers and patients alike will benefit.

But the challenge remains: Health plans can’t keep up with everything all the time. And they cannot use all of the data as thoroughly as they want to. 

I advise health plans to take stock of their needs and assess whether current data sources will get you where you need to go. If not, additional patient level data — identified or de-identified — from a new outside source could very well be in order. 

Let’s review some of the cases where a health plan could tap into some of the available ecosystems to solve key challenges.

Data Efforts Are Getting Budget Dollars

In 2016/17, it was estimated that the life and health insurance industry spent over $3.5 billion on marketing and advertising activities. (Estimate is compiled from more than one source and may include some commission payments). 

On top of this, according to Novarica, Gartner and other watchers, the industry plans to spend hundreds of millions of dollars a year through the 2020s on data and analytics talent, technical infrastructure that supports AI and machine learning, advancements in digital capabilities and modeling, as well as improving content and communication management systems. Throughout, every functional department of a health plan will seek data-driven understanding and confidence. 

Sample Payer buckets illustrated below: 

Large Payers (Top 10): In-house data management leaders and large-scale analytics teams at large payers often can be funded to focus on select hot topics of the day, and sometimes they build things themselves. The opportunity for larger payers is to acquire an assortment of sample datasets in the size and with necessary permissions so they can focus on what they are good at: the analytics, evaluation of new product designs, improvement to sales plans, sales enablement and sales effectiveness, negotiating network contracts (think value-based care) and delivery of care.

Next 100 Payers: While these payers have smaller in-house tech and analytics teams, they can still have resources in terms of staff and dollars to spend on services or data. They might find it very useful for a fixed amount of time to engage third-party data sources so they don’t have to commit to hire permanent staff. I have seen many engage analytics expertise to go with it, or buy/license just the data.

The rest of the market: As we move to smaller, regional or independent health plans, often I see that they have small in-house teams dedicated to data management and small teams dedicated to analytics. (A team might be as small as two or three persons.) It is not uncommon that they outsource part of their data management and data analytics capabilities. They tend to have more updated technology platforms and can easily acquire third-party data on demand, plugging it in quickly, to complement their curated internal data sets.

See also: Overcoming Human Biases via Data

Use Cases Abound

In general, richer patient-level data can help health plans address needs in two big categories: market segmentation and risk management. Below are market trends that drive needs in these categories, with an opinion as to where the opportunity lies for helping plans address them. Please share your thoughts with me in the comments section below.

Good people, friends and former colleagues, are losing their jobs as big insurance companies lay off staff. The sliced tether to the mothership has some considering making the jump to insurtech. As a co-founder in the insurtech space with a corporate background, I’ve been getting a lot of calls, and answering the same set of questions: What is the startup scene like, who is hiring, how to get started?

These are logical questions. I even have some decent answers.

These are also the wrong questions. They’ll help you find a job, but they won’t help you understand if you’re going to be excited to get out of bed in the morning or whether an entrepreneurial job slowly crushes you into desiccated powder.

The right question isn’t about logistics – it’s about the internal transition you’ll need to make, and whether you want to live that change.

The right question is: Who do I need to become to thrive in insurtech? (Or really, any corporate to entrepreneurial transition.) 

By thrive, I don’t mean start a unicorn. If we knew the steps to do that, 75%-plus of venture capital funds would not fail to make a profit for their investors. I also don’t mean wantrapreneuring – turning not doing into a career. Wantrapreneuring is skating from meetup to meetup, asking for lots of advice about what to do (then pushing back with a strong opinion of how it should be done, all the while not… doing).

I mean, doing the work. Finding an idea. Talking to customers. Convincing a co-founder or two and a team to join you. Or joining the team. Designing the product. Checking the font on every piece of customer communication. Figuring out why your freaking payroll vendor’s system doesn’t just WORK. 

And enjoying it. Coming into yourself in this space. Feeling like every challenge stretches you in a new direction. All the while handling the emotional extremes (which I guarantee are rawer and realer than corporate).

So, having had a corporate career before becoming an entrepreneur, here’s my read on the person you’ll need to become:

A shipper, not a soother

You know all those meetings to get opinions on a project before you actually start it? Aimed a little at understanding what your colleagues know, and a lot at tamping down later aggressive politics from people who feel left out? 

Just stop. 

Draft something, share it with your teammates, tear it up and make it better with their feedback and SHIP IT! 

The scales tip the other way here – the issue isn’t that you might offend by putting something on paper, it’s that you’ll never get to your destination if you don’t complete anything. (See wantrapreneur, above).

See also: COVID-19: Technology, Investment, Innovation

I promise you it’s leftover corporate-induced anxiety that’s preventing you from shipping. And you 100% need to find a way to force through it in the entrepreneurial environment. So ship the pitch deck, the blog post, the story, the code. Relentlessly focus on your own output. 

(Also, if you don’t write it down, or type it, or draw it, or record it, it doesn’t count. In your head is not done. So do it.)

A no-seeker, not a yes-orchestrator

You know the pre-meetings? The ones you do with your boss’s seven peers to get their input and objections before the big leadership meeting? Your goal is to avoid a no from the big boss, so at least you can keep moving. 

That’s not a model for a startup. 

Of course you should get lots of feedback (mostly from customers), and of course you should take your partners’ politics into account. 

However, the biggest gift in startup life is a quick no. 

The biggest gift in startup life is a quick no. (I said it again – this one took me too long to learn.)

And it’s amazing how many people won’t have the decency or understanding to give it to you. A maybe is not a yes. A maybe does nothing but eat up runway. When you’re small, you’re surviving on a shoestring and updrafts of hope. You need to find all-in, strong-yes partners. 

And if you’re working hard for the yes that’s not coming, these aren’t your people. Sorry. 

A lightning rod, not a moderator

In corporate life, being someone with a “strong personality” will show up in your performance review. “Tone it down,” they say. “Watch the humor,” they say, until you realize you’ve risen in the ranks by sanding down every corner that makes you, well, you. Your personality, your opinions and your willingness to argue something from the heart are the cost of fitting in. 

In the startup world, nobody funds boring. Nobody joins boring. Nobody takes a chance on boring. Average gets you nowhere. Inoffensive is a lack of conviction. 

Be prepared to own your ideas, your journey, your very self and argue them strongly. Don’t play to the crowd. Better to irritate a few people if it means pulling the ones who can help you into your slipstream.

A doer, not a delegator

Early stage, there’s just too much work and nobody to do it. You can’t set up half your payroll system, design just the principles of a user experience or draft an outline of a letter to a customer – these aren’t partially done – they are an absolute waste of time precisely because they are incomplete, and therefore unusable. 

You can’t delegate completion when there’s nobody to delegate to. Do your work all the way to the end. Let go of the perfection of corporate life and the 87 rounds of reviews, and content yourself with a customer letter you think you’d understand and with a quick proofreading. 

Oh, and delegate complete tasks, not fragments. You need a team that can also finish their work.

See also: Step 1 to Your After-COVID Future

Transition means change

I don’t buy the arguments that people are either successful in corporate environments or in entrepreneurial environments. That’s accepting a world in which none of us can learn and grow, and in which we’ll never succeed at anything we didn’t try in our 20s. It’s nothing but a package of hubris and negativity all mixed up together. 

It is true, though, that corporate and entrepreneurial environments test us in different ways. If you have a corporate job and you want to work at or start a startup, can you find a job or can you start a company? Of course you can, given time and resources. 

But can you thrive? You need to be willing to change.

Becoming an entrepreneur is just that, a becoming.

Entrepreneurship strips away our masks, for founders and for team members, both. The fate of the business is in your hands. Your work stands for itself. You stand up for yourself. 

So, don’t overweight your thinking to whether you can find a job in the entrepreneurial world. Think about who you’ll need to become to thrive in that space. Does your heart sing with delight when you think about becoming that person? 

You, and the people you will work with, deserve that.