Episode 172: Inside insurtech: The role of actuarial training from a CEO / Épisode 172 : Les coulisses de l’assurtech : un PDG se penche sur l’importance de la formation en actuariat

Welcome to Seeing Beyond Risk, a podcast series from the Canadian Institute of Actuaries.

I’m Chris Fievoli, Actuary, Communications and Public Affairs at the CIA.

The world of insurtech has been catching the attention of actuaries and insurance professionals, and this has led to new approaches when it

comes to developing and distributing insurance products.

One of the companies involved in this new era is PolicyMe, and today we are speaking to Andrew Ostro, their CEO and cofounder.

Thanks very much for joining us today.

Thanks, Chris. It’s great to be here.

So, to start, can we just get a bit of your background? And maybe tell us about your experience with the actuarial profession so far.

Absolutely. I’m Andrew Ostro. I’m the CEO and one of the cofounders here at PolicyMe. I have a very interesting background, probably quite unique for an actuary.

I started off my career as an actuary. I went to the University of Western Ontario, graduated first in my class there and started my career as an actuary at Towers Watson.

Working in the pension space, wrote my exams, became an FSA and a CRA, and then essentially left the profession directly after that.

I went into a more non-traditional role. I started working as a management consultant with Oliver Wyman. There, I was working exclusively in the insurance practice but not doing traditional actuarial

work.

It was more strategy work, working with executives at insurance companies on their finance and risk, distribution, strategy, operations – any of the different elements that were kind of at the top of the pyramid in terms of what they were dealing with at that given time.

So, it definitely helped that I had an actuarial background. I understood the companies. I understood how they operated. I understood how they made money. I understood the risk. But, I wasn’t directly using that skill set in that job.

Let’s talk about that a bit. Tell us about PolicyMe, and I want to focus on what differences would insurance customers see if they use your platform compared to what we’re traditionally used to.

Quite simply, PolicyMe is a direct-to-consumer life insurance platform, and what that really means is we provide a website and an application for customers to log in, learn about life insurance, understand their needs, learn about the different products available, get quotes for the products and ultimately get advised on what they need and submit their applications and get insured all in one customer journey – customer experience.

That was a very different path we took, and I think that’s really where I’d like to start in terms of describing the key differences in how our customers experience the platform and how our customers buy life insurance.

In our experience, what we saw was two different approaches to how the life insurance industry was approaching the direct-to-consumer space.

One was just to throw out simplified-issue products. And what that means is saying, look, we have a customer, the key problem here is typically that customer needs a broker to guide them through the process, sell that product and get them to buy. So, without that broker, we have a conversion problem on “How will we get the customer to convert?”

It’s been very clear that a lot of customers want to buy life insurance online. When you look at the statistics and the data, you see that the majority of customers actually start their search online.

Whether they’re Googling life insurance or looking for quotes online, there’s some indication that they want to buy that product online, but ultimately get stuck and go and work with a broker.

So the industry has kind of approached that problem in two ways. They said one was simplified products, which essentially means no underwriting, or very little underwriting, ask a very few finite set of questions, don’t go very deep into the medical history and essentially default to higher prices.

There’s always a trade-off between underwriting depth and price, so in a world where you want to keep the customer experience simple, ask less questions, cause less drops. The consequence is you have to charge more.

So, that’s been one approach: throw simplified products at this conversion problem.

The other approach we’ve seen is take the existing, fully underwritten products that the carriers already have built and offer through their broker channels – and their independent advisor channels – and adapt those for the direct-to-consumer space, which is essentially what we did as a broker when we sold other carriers’ products.

The issue there is that, quite frankly, the underwriting experiences, the operations, the customer journey are not designed for the direct-to-consumer space.

So, you end up with questions that are quite difficult for a customer to answer on their own. You end up with an underwriting process that involves a lot of handoffs and time in-between steps, which might be OK in the broker channel, where the customer’s not expecting real-time and automated underwriting, but ultimately you end up with an experience that isn’t directly designed for that customer, for that journey, and therefore creates a lot of drops.

So, our view was we needed to do something very different. And we didn’t want to go the simplified product route, as I mentioned. We don’t think a customer should ever sacrifice price for experience.

So, in other words, you should never have to pay more to get a better customer experience, which is essentially what the simplified product does.

So, our view was we needed to build new, fully underwritten products from scratch, ones that were designed explicitly for the direct-to-consumer channel.

When I say, “design a product,” it’s more than just: What are the features? What are the benefits? It’s: What is the underwriting? What is the application? What are the rules? How does the customer pay?

All the different elements around getting that customer insured more than just, “What does your contract look like?”

You mentioned underwriting a few times and I know anybody with experience in the insurance industry knows it’s a very long process. It can be very complicated.

Can you tell us a bit more about how your company is addressing that concern and perhaps making it easier and less time consuming?

I’d say there’s probably three different tenets to how we approach the underwriting problem and really try and clean that process up and make it more customer friendly.

And most of this was learned through our time as a broker. As I mentioned, the first two years that we were in operation, we were selling other carriers’ products.

We saw firsthand through thousands of applications what was working, what wasn’t working, what was causing drop-offs, what was getting customers to complete the process.

So the first one was minimizing handoffs. This might seem obvious, but handoffs are everywhere in the underwriting process today, where it comes to the application being submitted by the customer; that typically goes into a system where a new business rep, on behalf of the carrier, takes a look at the application, figures out what to do with it, sends it off to an underwriter, the underwriter takes a look through, they decide what requirements are needed.

Is there a blood test needed? Is there an APS needed? That goes back to the new business rep, and that goes back to the broker, then the broker communicates with the customer, and then up and down that chain – usually a couple of times – until ultimately a decision is made in terms of approval or not.

And even then, once the approval is made, the policy issuance requires a lot of handoffs with the new business rep and the broker.

So, first and foremost, it was centralizing all the information using automation, using workflows to make decisions to pass work between the different parties that need to look at it and really cleaning up the number of times that file is going between different people and different entities.

The second one is really around automating more rules. A lot of carriers have automation in their underwriting.

It’s usually first-pass automation and first-pass underwriting, where the data that flows in through the application goes through a set of rules that then assists the underwriter in really determining what to do next.

What we’ve found is that it’s very light. Typically, it’s only using the very high-level data.

There’s a lot of decisions that are still being made with the underwriter that could be automated and should be automated. And I’m not saying everything an underwriter does needs to be automated, but I’m saying there’s still a lot more that could be automated to support that underwriter with their decision making and really speed up that process.

So one thing we did is we looked at where can we add this? Where can we analyze underwriting decisions, look for patterns, look for ways to bring more of that decision-making into an automated workflow?

And then number three was just thinking about protective value in the same way that carriers think about expenses.

So, typically when a product is priced, there’s some component of risk – expected claims adjusted for time value, interest rates, all that stuff – but really the expected claims make up a good chunk of the premium.

And then there is the expenses. What is it going to cost the carrier to underwrite, to administer all the operations associated with the product, and then some margin on top of that. And it’s a little bit more complicated than that, but that’s largely how an insurance product is

priced.

So, our view was, when we have these extra steps, like blood and urine tests and APSs, can we measure the protective value? (Essentially what the reduction in risk that extra step is creating.) In other words, when we do the blood test, what is the outcome?

If we do a blood test for 100 people who specify that their BMI is over some threshold or in some range that we’re kind of on the fence on, do we need the blood test or not? When we do the blood test on 100 people who fit that criterion, let‘s analyze the impact of it.

And we might find that for 99 of those people, the underwriting decision was the same as it would have been if there was no test. In other words, the test didn’t tell us anything. For one of those people, it did.

And let’s measure, let’s say that for one of those people, we doubled the price. So, can we take that doubling, divide it over the hundred and think about, that’s the value, the protective value created by that blood test? That’s a cost.

What is the cost of the blood test? And think of it in the same way. You can gross up the risk portion, the claims portion of your premium, by adding that on and then reduce it from the expenses by not doing the blood test and all the money you save there.

Our view is if we do think of those in the same way and really bring that math, the science and the quantification back into that calculation, we can actually, not necessarily charge more, but calculate the risk, incorporate it into price and reduce the need for some of these requirements that really slow down the customer experience.

I know when you start up an operation, there’s a lot of challenges. I was hoping you could share with us some of the unexpected challenges that you faced as you were doing this.

There are so many things that come up that you just don’t think about when you make that decision to, hey, I’m going to go start a company, and hey, I’m going to go try and change the world and do all these different things and disrupt an industry.

It’s usually a very positive set of thinking and thoughts that occur at that moment when you make the decision, and then a whole set of just unexpected issues that come up in the following years that you’d never even think of.

But for us, I think the biggest thing is really some of the external factors.

I think when you really get down to what it means to run a business and what it means to innovate, you can start to think through like, “OK, what’s going to come up?”

“I’m going to have to work with insurance carriers and reinsurance carriers. We’re not going to have the capital to work our way through this industry and find partners.” And there’s going to be issues with customers who aren’t used to buying like this – we’ve got to solve for the customer experience.

Those are certainly challenges. I think those ones are a little bit more expected when you think about starting a business in this space. The ones you don’t think about are the external factors.

So obviously COVID. Obviously, it caused a huge change in the way we work, in the way the industry operates, things that you don’t really think about or plan for at the time of starting the business.

And even from that, just this whole shift to working from home, even post-COVID on how to think about your employees and the culture, and everyone works together and keeping people motivated and effective and efficient.

Now then interest rates causing huge impacts on the way we think about funding and our capital, inflation obviously driving up our costs.

Even more recently, the Silicon Valley Bank that collapsed there, we weren’t directly exposed to them, but that obviously had huge impacts on the broader tech ecosystem and VC landscape.

I think it’s just really – I mean there’s no way to really predict these things. I think no one can predict them and I think of you, as a startup founder, if you try and start worrying about this stuff, it’s just going to distract you, but I think it’s just being aware that things aren’t going to go to plan.

There’s going to be things completely out of your control that are going to come in. Really focusing on a sound business model, having a good business that can weather a lot of these storms, good unit economics, good underlying financials and really just some robustness into your business model that can adapt and survive these external factors.

Well, as you know, actuaries tend to manage risk, mitigate risk; we tend to be risk averse. But starting up this company, that represented taking on a huge risk.

How do you balance those competing perspectives in your own mind?

I think for me when I think about what an actuary does – yes, that’s right, they mitigate risk – but even one step earlier, it’s quantifying the risk.

I think that was the big take away from my actuarial education and limited experience as an actuary, but going through that program and that process was: It’s not always about mitigating; it’s fine to take risks.

And I’m of the mindset that you need to take risks, and the only way to innovate is really by not playing it safe and going after some big wins and recognizing that comes with big losses.

But, I think being able to quantify what that risk is and understand what it is you’re taking is just so critical in anything you do, especially in starting a company. So, I think that kind of mindset of saying, “Look, our goal here isn’t to create this single scenario where no matter what happens, we end up in the same outcome.” We definitely want this range of outcomes.

Obviously, you’d prefer not to have any downside, but that’s just impossible when you’re trying to incorporate some upside, and really, that’s what you’re doing when you’re innovating. So, it was less about “How do we turn this range of outcomes into a single outcome?”

It was more about “How do we understand what that range of outcomes are?” and “How can we work towards making the upside scenario or some of these wins more probabilistic?” And so that’s where I think the big difference is.

And, for us it was – it’s never easy to take on a lot of risk. It’s always scary. All else being equal, you’d rather know the outcome and not have to worry. But, again, as you put it, you can’t start a company without taking on huge risk.

Again, really thinking through “How do we understand what it is we’re up against?” and “How do we understand the different scenarios that can play out?” and “How do we do everything we can to work towards the ultimate scenario on the one that we really want?”

Now I assume that you have to deal with investors in the course of operating your business, and I imagine they would have a hard time understanding some of the oddities with the insurance business. We have a very long-term perspective. There may be a long payback period.

How do you get them comfortable with this new world?

This was really difficult. A lot of technology investors, a lot of VCs, do not understand insurance at all.

It is a very complicated space – especially life insurance, when you start to get into some of these permanent products, investment-linked products, long-duration products. They’re tricky.

The pricing isn’t intuitive. You’re burning cash. A typical life insurance carrier takes four and a half years to get paid back when they sell a policy.

So, the more they sell, the lower their income, the bigger their loss for the year. It’s not intuitive, in a lot of ways, to have investors typically think about software businesses, which is usually the recurring revenue SaaS model.

That’s the one a lot of investors really know quite well, which is sign a bunch of customers, usually in the B2B space. You sell them software as a service. They become a customer, paying a subscription, and that is a recurring number.

You look at churn. What is going to be the drop off on a yearly basis? What’s my new customer rate? And then the math kind of flows from there. Investors really understand how to evaluate those businesses, how to model them. And then you start to look at a business like ours, which is

very different.

These are one-time buyers. Sometimes they buy multiple products, but they’re not going to be recurring buyers.

Commissions are front loaded. The economics are very tricky. There’s investment returns and tons of different assumptions that go into pricing models that impact profitability of these products.

It takes 10 years to really understand whether a product really was as profitable as you thought it was. Just a lot of things that are just non-intuitive to a typical tech investor.

And then when I think, even the ones that do spend a little bit more time in the insurance base, it tends to be on the P&C side, where certainly it is more complicated than a typical software business, but not nearly as complicated as the life side.

And what I mean by that is when we look at the P&C space, I mean, it’s one-year policies. The name of the game is generally loss ratios and taking on better underwriting to reduce claims.

You’re dealing with generally high frequency in events. So, for an auto policy, you might expect a claim – I don’t know – one every three years, the customer might have a claim.

Therefore you sell a few thousand policies, you might expect 3,000 policies, you might expect 1,000 claims in the first year, and in a couple months you can very easily evaluate whether your underwriting is effective, whether the claims ratio is going to be where you thought it was. It’s just more intuitive to think about that.

But then on the life side, I mean, you’re dealing with a one in two-, three-thousand event, especially term life insurance to a 35-year-old, 40-year-old.

You change underwriting, you’re looking at years to really evaluate, to have enough data to really understand whether your claims experience and your loss experience is worse than expected and not just random.

You see one or two claims in the first year. You don’t know whether that’s a sign of poor underwriting or just random noise. It’s a combination of both the investors not really understanding it, but then also it leads to difficulty to innovate in life insurance for that very reason, right?

I think what we’re talking about here is underwriting process rules. That involves a lot of risk, right?

You’re actually changing the way you evaluate whether you sell a policy or not and doing it in a way where you really can’t measure effectiveness for a long period of time.

And also with life insurance, you can’t change the price, right? With property and casualty insurance you sell someone an auto policy, you sell a block of business, you find your claims ratio was way off in a year, you can increase the price in year two.

You can’t do that in life insurance. You can increase the price for new customers, but once you sell that 20-year policy to a group of people, their price is locked in for those 20 years.

So there’s just a lot of fundamental aspects of the life insurance base largely relating to the low-frequency events, the long duration of the policies – just the binary event of a claim-no claim that you’re either $0.00 or $1,000,000 in a loss, right? There’s nothing in between for an individual policy.

All those kinds of features just make this space very difficult to (1) explain to an investor, and then (2) very difficult to innovate in because of the reliance on capital and the reliance on partners with big balance sheets to really support and take some chances here.

So, it’s been tricky. I mean, it’s not an easy space to win in. And then you have the headwinds of a lot of the public companies or a lot of the startups that have recently gone public in the US in the P&C space just getting hammered on valuation over the last year or two, where you’ve seen these companies’ IPO and then just have the public markets completely devalue them within a year or two.

And it’s easy for an investor who doesn’t really understand our business or really understand the products to say, “Oh yeah, I saw this P&C insurance company in the US – looks kind of similar to you; distribute P&C products, trying to innovate on the underwriting process. The public markets have just devalued them. Why shouldn’t I be drawing that analogy to you?”

So I think for us it was really about finding the right investors and finding the right partners. It’s just so critical in this space.

When we started, I’ll admit I was very naive at the beginning, thinking, “OK, I have a great business idea. I have a great team. We have a great opportunity, a huge space. Every investor is going to be all over this.

I’m going to go after the big tech investors that everyone knows and everyone’s heard of, and they’re going to love this idea.” And you just find out very quickly that’s not how it works.

They don’t understand what you’re doing. And when they don’t understand what you’re doing, they generally try to stay away from it, so it was very important as a founder to not just say – just to rephrase it a little bit – it wasn’t just about, "We’ll take any investor, and it doesn’t really matter – they’re all going to want it.” It’s really about finding the right investor.

And money is not money. Capital from one partner is not the same as capital from another partner, and I learned that very quickly.

I think for us we found very good investors, very good partners I’d say about a year or two into building up the product and the company. That’s really allowed us to scale, to innovate and to make a big difference in what we’re doing.

That sounds very interesting. Sounds stressful too, but you seem to be doing OK, so that was really a fascinating look at this side of the business.

Thanks very much for coming on the podcast today to talk about it.

Yeah, thanks. Any time.

So if you enjoyed today’s conversation, we invite you to subscribe to our podcast series and catch up on prior episodes.

And if you happen to have an idea for a future episode or you would like to contribute to our Seeing Beyond Risk blog, we would like to hear from you. Contact information can be found in the show description.

Until next time, I’m Chris Fievoli, and thank you for tuning in to Seeing Beyond Risk.