Tag Archives: model

A Simple Model to Assess Insurtechs

“The paradox of teaching entrepreneurship is that such a formula necessarily cannot exist; because every innovation is new and unique, no authority can prescribe in concrete terms how to be innovative.”

― Peter Thiel, Zero to One

Whether we’re talking about telematics, artificial intelligence (AI), digital distribution or peer-to-peer, investing in insurance-related technology (commonly termed “insuretech” or “insurtech”) is no longer considered boring. In fact, insurtech is one of the hottest investable segments in the market. As a 20-plus-year veteran in insurance, I find it surreal that insurance has become this hip. Twenty years ago, I gulped as I sent an email to the CFO of my company, where I proposed that there was a unique opportunity in renters insurance. That particular email was ignored. Today, that idea is worth millions of dollars.

What changed?

Insurance seems to be the latest in a string of industries caught in the crosshairs on venture capital. With the success of Uber and AirBnB, VCs are now looking for the next stale industry to disrupt, and the insurance industry carries the reputation of being about as stale as they come. The VCs view the needless paperwork, cumbersome purchasing processes, dramatic claims settlement and overall old-school look and feel of the industry and think they can siphon those trillions of dollars of premium over to Silicon Valley. It seems like a reasonable thesis.

The problem is, it’s not going to happen that way. Insurance will NOT be disrupted. While insurance looks old and antiquated on the exterior, it is actually quite modern and vibrant on the interior. The insurance industry is actually the Uncle Drew of businesses; it’s just getting warmed up!

The Model

Much of the reason I think VCs are unaware of their doomed quest for insurance disruption is that they are looking at the market from a premium standpoint and envisioning being able to capture large chunks of it. $5 trillion is a lot of money. Without an appropriate model, an outsider coming into insurance can naively think they can capture even a fraction of this. But premium is strongly tied to losses. Those premium dollars are accounted for in future claims.

I once had a VC ask me what the fastest way to $100 million in revenue was. The answer is easy, “slash the premium.” I had to quickly follow up with, “and be prepared to be go insolvent, as there is no digging yourself out of that hole.” He didn’t quite get it, until I walked him through what happens to a dollar of premium as it enters the system. And it was this that became the basis of the model I use to assess new product formation and insurtech startups.

There are four basic components to my model. Regardless of new entrants, new products or new sources of capital, these four components remain everpresent in any insurance business model. Even if a disruptive force was able to penetrate the industry veil, that force would still need to reflect its value proposition within my four components.

Component 1 – EXPOSURE

This is the component that deals with insurance claims: past, present and future. Companies or products looking to capture value here must be able to reduce, prevent, quantify or economically transfer current or new risks or losses. Subcomponents in this category include expenses arising from fraud and the adjustment of claims, both of which can add substantially to overall losses.

See also: Insurance Coverage Porn  

Startups such as Nest are building products that increase home security by decreasing the likelihood of burglary (or increasing the likelihood of capturing the criminals on video) and thus reduce claims associated with burglary or theft. Part of assessing the value proposition of Nest is to first understand the magnitude of the claims associated with burglary and theft and then quantify what relief this product could provide (along with how that relief should be shared among stakeholders).

Another company that is doing some interesting things in this model component is Livegenic (disclaimer: I have become friends with the team). Livegenic allows insurers to adjust claims and capture video and imagery using the mobile phone of the insured. This reduces the expenses associated with having to send an adjuster out to each and every claim. Loss adjustment expenses can be in excess of 10% of all claims, so technology that reduces that by a few basis points can be quite valuable to an insurer’s bottom line and ultimately its prices and competitiveness.

Component 2 – DISTRIBUTION

This component focuses on the expenses associated with getting insurance product into the hands of a customer. Insurtech companies in this space are typically focused on driving down commissions. This can be done by eliminating brokers and going directly to customers. Savings can also be achieved by creating efficient marketplace portals that allow customers to easily buy coverage.

Embroker is one of many companies trying to do just that in the small commercial space by creating a fully digital business insurance experience. Companies such as Denim Labs are providing social and mobile marketing services to companies in insurance. And then there is Lemonade, which is developing AI technology that it hopes will reduce the friction of digitally purchasing (its) insurance and making the buying process “delightful.”  Peer-to-peer (P2P) insurance is a fairly new insurtech distribution model that attempts to use the strength of close ties via social methods for friends and close associates to come together to make their own insurance pools.

Distribution expenses in insurance are some of the highest in any industry. As with the risk component, reducing expenses in this component by even a few basis points is incredibly valuable.

Component 3 – CAPITAL

This component focuses on the expenses associated with providing capital or the reinsurance backstop to a risk or portfolio. For many insurers, reinsurance is the largest expense component in the P&L. Capital is such an important component to the business model that the ramifications of it almost always leak into the other components. This was one of my criticisms of  Lemonade recently. Lemonade will have a lot of difficulty executing some of the aspects of its business model simply because it cedes 100% of its business to reinsurers. So, when it comes to pricing or its general underwriting guidelines, its reinsurance expenses will overwhelm other initiatives. Lemonade can’t be the low-cost provider AND a peer-to-peer distributor because its reinsurance expenses will force it to choose one or the other. This is a nuance that many VCs will miss in their evaluation of insurtechs!

For those seeking disruption in insurance, we have historical precedent of what that might look like based on the last 20 years of alternative capital flooding into the insurance space. I will devote space to this in future articles, but, in brief, this alternative capital has made reinsurance so inexpensive that smaller reinsurers are facing an existential crisis.

Companies such as Nephila Capital and Fermat Capital are the Ubers of insurance. Their ability to connect investors closer to the insurance customer along with their ability to package and securitize tranches of risk have shrunk capital expenses tremendously. Profit margins for reinsurers are collapsing, and new business models are shrinking the insurance stack. It is even possible today to bypass BOTH veritable insurers and reinsurers and put the capital markets in closer contact with customers. (If you are a fan of Michael Lewis and insurance, you will enjoy this article, which ties nicely into this section of the article).

In the insurtech space, VCs are actually behind the game. Alternative capital has already disrupted the space, and many of the investments that VCs are making are in the other components I have highlighted. Because of the size of this component, VCs may have already missed most of the huge returns.

Component 4 – OPERATIONS

The final component is often the one overlooked. Operations includes all of the other expenses not associated with the actual risk, backing the risk or transferring the risk from customer to capital. This component includes regulatory compliance, overhead, IT operations, real estate, product development and staff, just to name a few.

It is often overlooked because it is the least connected to actually insuring a risk, but it is vitally important to the health and viability of an insurer. Mistakes here can have major ramifications. Errors in compliance can lead to regulatory problems; errors in IT infrastructure can lead to legacy issues that become very expensive to resolve. I don’t know a single mainstream insurer that does not have a legacy infrastructure that is impinging on its ability to execute its business plan. Companies such as Majesco are building cloud-based insurance platforms seeking to solve that problem.

See also: Why AI Will Transform Insurance  

It is this component of the business model that allows an insurer to be nimble, to get products to market faster, to outpace its competitors. It’s not a component that necessarily drives financial statements in the short term, but in the long run it can be the friction that grinds everything down to a halt or not.


I have presented a simple model that I use when I assess not just new insurtech companies but also new insurance products coming into the market. By breaking the insurance chain into these immutable components, I can estimate what impact the solution proposed will provide. In general, the bigger the impact and the more components a solution touches the more valuable it will be.

In future articles, I will use this model to assess the insurtech landscape. I will also use this model to assess how VCs are investing their capital and whether they are scrutinizing the opportunities as well as they should, or just falling prey to the fear of missing out.

Originally published at www.insnerds.com,

Data Science: Methods Matter (Part 3)

Data science has grown in inevitability as it has grown in value. Many organizations are finding that the time they spend in carefully extracting the “truth” from their data is time that pays real dividends. Part of the credit goes to those data scientists who conceived of a data science methodology that would unify processes and standardize the science. Methods matter.

 In Part 1 and Part 2 of our series on data science methods, we set the stage. Data science is not very different from other applied sciences in that it uses the best building blocks and information it can to form a viable solution to an issue, whatever that issue may be. So, great care is taken to make sure that those building blocks are clean and free from debris. It would be wonderful if the next step were to simply plug the data into the solution and let it run. Unfortunately, there is no one solution. Most often, the solution must be iteratively built.

This can be surprising to those who are unfamiliar with data analytics. “Doesn’t a plug-and-play solution just exist?” The answer is both yes and no. For example, repeat analytics, and those with fairly simple parameters and simple data streams, reusable tools and models, do exist. However, when an organization is looking for unique answers to unique issues, a unique solution is the best and only safe approach. Let’s consider an example.

See also: Forget Big Data — Focus on Small Data

In insurance marketing, customer retention is a vital metric of success. Insurance marketers are continually keeping tabs on aspects of customer behavior that may lead to increasing retention. They may be searching for specific behaviors that will allow them to lower rates for certain groups, or they may look for triggers that will help the undesired kind of customer to leave. Data will answer many of their questions, but knowing how to employ that data will vary with every insurer.

For example, each insurer’s data contains the secrets to its customer persistency (or lack thereof), and no two insurers are alike. Applying one set of analytically derived business rules may work well for one insurer — while it would be big mistake to use the same criteria for another insurer. To arrive at the correct business conclusions, insurers need to build a custom-created solution that accounts for their uniqueness.

Building the Solution

In data science, building the solution is also a matter of testing a variety of different techniques. Multiple models will very likely be produced in the course of finding the solution that produces the best results.

Once the data set is prepared and extensive exploratory analysis has been performed, it is time to begin to build the models. The data set will be broken into at least two parts. The first part will be used for “training” the solution. The second portion of the data will be saved for testing the solution’s validity. If the solution can be used to “predict” historical trends correctly, it will likely be viable for predicting the near future as well.

What is involved in training the solution? 

 A multitude of statistical and machine-learning techniques can be applied to the training set to see which method generates the most accurate predictions on the test data. The methods chosen are largely determined by the distribution of the target variable. The target variable is what you are trying to predict.

A host of techniques and criteria are used to determine which technique will work best on the test data. There is a bucketful of acronyms from which a data scientist will choose (e.g. AUC, MAPE and MSE). Sometimes business metrics are more important than statistical metrics for determining the best model. Simplicity and understandability are two other factors the data scientist takes into consideration when choosing a technique.

Modeling is more complex than simply picking a technique. It is an iterative process where successive rounds of testing may cause the data scientist to add or drop features based upon their predictive strengths. Not unlike underwriting and actuarial science, the final result of data modeling is often a combination of art and science.

See also: Competing in an Age of Data Symmetry

What are data scientists looking for when they are testing the solution?

Accuracy is just one of the traits desired in an effective method. If the predictive strength of the model holds up on the test data, then it is a viable solution. If the predictive strength is drastically reduced on the test data set, then the model may be overfitted. In that case, it is time to reexamine the solution and finalize an approach that generates consistently accurate results between the training data and the test data. It is at this stage that a data scientist will often open up their findings to evaluation and scrutiny.

To validate the solution, the data scientist will show multiple models and their results to business analysts and other data scientists, explaining the different techniques that were used to come to the data’s “conclusions.”  The greater team will take many things into consideration and often has great value in making sure that some unintentional issues haven’t crept into the analysis. Are there factors that may have tainted the model? Are the results that the model seems to be generating still relevant to the business objectives they were designed to achieve? After a thorough review, the solution is approved for real testing and future use.

In our final installment, we’ll look at what it means to test and “go live” with a data project, letting the real data flow through the solution to provide real conclusions. We will also discuss how the solution can maintain its value to the organization through monitoring and updating as needed based on changing business dynamics. As a part of our last thoughts, we will also give some examples of how data projects can have a deep impact on the insurers that use them — choosing to operate from a position of analysis and understanding instead of thoughtful conjecture.

The image used with this article first appeared here.

The 10 Top Trends From a Pivotal 2015

Many will pinpoint 2015 as a pivotal year – the turning point in the transformation of the business of insurance. External influencers and rapid technology advancements are resulting in major shifts in strategy, areas of focus and investment. Many insurers are thinking big – beyond the typical incremental change and toward bold moves that will establish them as leaders in the digital age.

Here are the top 10 trends that laid the foundation for this pivotal year and positioned the insurance industry for an amazing 2016 and the years beyond. The trends are dominated and enabled by technology developments, which continue to be interwoven into the fabric of insurance. The trends are:

  1. Digital transformation is taking hold, even in insurance.
  2. Innovation and innovative thinking have no boundaries.
  3. Huge $dollars$ being are being poured into start-ups.
  4. New ecosystems are emerging.
  5. Distribution channels are under strain, leading to shifts in investments.
  6. Core modernization is required and continues to consume insurers.
  7. Positive shifts are occurring in customer focus and priority.
  8. New tools, data and models are being embraced but are still a struggle to adopt.
  9. Many technologies are maturing and being adopted – cloud, analytics…
  10. Tech advancement is still outpacing the ability to consume.

Insurance executives can no longer ignore or play down these trends. Although the terms “disruption” and “transformation” are popping up everywhere, they are no longer buzz words but reality.

It would be a mistake to dismiss the magnitude of the shifts. As one senior industry executive put it, “Our industry will be substantially different five years from now. Companies that do not aggressively transform will be at risk of failing.”

This view is shared by many industry leaders, who sense that the tide is shifting in the new digital era. Unfortunately, many others are hoping to ride out the rest of their career without driving change, an approach that is risky.

The full research brief describing the trends is located here.

A Wedding’s Lessons on Customer Insight

Enjoying the emotions of my second son’s wedding was the highlight of our family’s year. After the drama of the ceremony (including the comedy of a fire alarm that wouldn’t give up), we enjoyed great food at the reception, moving speeches and then the joys of drink and dancing into the wee small hours. Everyone agreed it was a great day with brilliant weather.

Reflecting on this time afterward reminded me of the importance of such events in all our lives, including the lives of our customers. Whether it is getting married, the birth of your children, moving home or even (as I’ve had the joy of experiencing) the arrival of your first grandchild, such milestones affect us all.

Does our marketing or customer insight work always reflect this? Do you target your marketing on the basis of important and appropriate trigger events in your customers’ lives?

Common practice can be to assume that the gold standard of targeted direct or digital marketing is to use logistic regression propensity models and perhaps optimization across multiple models to determine next best action. However, think for a moment about your own life. Do you feel that you walk around with a more or less permanent level of propensity to buy something? Apart from perhaps coffee, chocolate or alcohol, I suspect not.

In our lives, is it really more about “events, dear boy, events,” as former British Prime Minister Harold Macmillan quipped? Different experiences and special occasions help us to mark out the progress of our lives and trigger us to reflect on other needs and aspirations. These can be as mundane as the annual renewal cycle for home insurance or as momentous as the birth of our second child, for considering life insurance or the need for a new home. Those who an predict the timing of a trigger event will outperform those who target using any propensity model.

How long ago did you reflect on the right timing to talk with your customers? Does your customer segmentation capture the key life events that shape their thinking about new needs, where your products and services could help them?

It is perhaps also time to acknowledge that the whole concept of an apparently permanent “needs-based segmentation” is looking dated. Customers rarely have such semi-permanent needs, at least not ones that they are aware of or will consider at every point in time.

Perhaps it’s more helpful to think about the “jobs they want to get done” when the right triggers arise in their lives. Segmenting based on the jobs your products and services can help your customers to “get done easily” can be very powerful. Even more so if you can combine that behavioral analysis to predict the trigger events or actions that prompt such a job requirement.

Have you experienced this shift to thinking more about timing in your analysis and marketing targeting?

What Socrates Says on Customer Insight

Are you and your Customer Insight team too often frustrated that you’re not making a difference in your business? Do your internal customers ever criticize what they receive from your team, asking, “Where’s the insight?” Sometimes this is because of technical skills or barriers that need to be addressed, but very often it’s because of poor communication. Do you need to get a better brief?

What I mean is this: Marketers or other stakeholders within your business can come to Customer Insight and ask for a piece of data/analysis/research. If the analyst just gives them what they asked for (or a version of that based on their understanding of what they heard), it’s often a recipe for disappointment. Analysts can feel limited by work that’s not creative or using their technical skills. Your internal customers can be disappointed, to receive something other than what they meant, and that doesn’t meet their real need.

This communication challenge is of, course, well-known in the field of project management. This tree swing example normally helps to illustrate this dilemma.

But there is more, beyond the challenge of documenting requirements clearly, in a good brief. Have you also found that what your internal customers doesn’t ask for is what they really need? Is what they want not what they need? That’s my experience, too. So, to help analysts improve their questioning skills in this area, I’ve been borrowing a technique from the world of leadership coaching.

Trained coaches will likely have come across Socratic questioning. It is a style of inquiry, aimed at helping the one being questioned to critique his own thinking, assumptions and viewpoint. Working with both experienced and junior analysts, I’ve found that the principles of Socratic questioning can help them in questioning what they are asked to provide, to get to the real need.

Here’s a very brief intro to this style of questioning, as proposed by the great Socrates himself:

Conceptual clarification questions: “What exactly does this mean?”; “Can you rephrase that, please?”; “Can you give me an example?”

Probing assumptions: “You seem to be assuming…?”; “Please explain why/how…?”; “How can you verify or disprove that assumption?”

Probing rationale, reasons and evidence: “Why is this happening?”; “Would it stand up in court?”; “How can I be sure?”

Questioning viewpoints and perspectives: “Another way of looking at this is…, does this seem reasonable?”; “What would… say about it?”;

Probing implications and consequences: “Then what would happen?”; “Why is… important?”; “How does… fit with what we learned before?”

Given previous advice on being action-oriented throughout any customer insight work, I find it helps to add another line of questioning to this model. That is to explicitly ask what action is going to be taken as a result of this request. This is important, to avoid precious analyst time being taken up with questions that are just out of curiosity. You need to know what action is planned.

None of the above is intended to be used word for word, or imposed without intelligent interpretation, in the language and culture of your organization. However, applied sensibly, I’ve seen that it can help empower analysts to question more and to improve their skills in eliciting real business needs.

When the real need is understood and captured in a clear brief, then you stand a much better chance of getting real insight.

What have you found works? How do you get a better brief?