a16z Podcast - a16z Podcast: Reinventing Insurance

Episode Date: January 29, 2016

Your homeowner’s insurance didn’t anticipate Airbnb. Your car insurance certainly didn’t see Lyft and Uber coming. And when your car drives itself, it’s anyone’s guess how the insurance industry will wrap its collective head around that one. a16z’s Frank Chen and Mike Paulus talk insurance on this segment of the pod. Yes, insurance. Insurance may not be the sexiest part of your life (hopefully), but because of the changes in how we move through the world -- literally and figuratively -- insurance is due for a reinvention. What are the possibilities for new and better insurance, and which technologies and trends are driving it.

Welcome to the A16Z podcast. I'm Michael Copeland. We are sitting here in our somewhat cramped podcast room with Frank Chen, who runs our deal team and Mike Paulus, who is on the deal team. Welcome, guys. Thanks so much for having us. Great to be here. Frank, last year around this time, you wrote a piece for our 16 things package, and it was about insurance. Of all things. And it was about this trend for the future. And it turns out, actually, of all the things. and, or I won't say of all the things, but of some of those things, it got this incredible response. And all of a sudden, there were waves and waves of startups in insurance. I have to say I was the most surprised to see the reaction to this post. And in fact, Mike joined us after having read the post and saying, hey, look, it's an venture firm interested in insurance. So insurance, look, I personally pay for insurance. I don't get excited about insurance. But what is going on? Why are there entrepreneurs and people like yourselves focused on what seems to be a fairly hidebound and, you know, explainable, obvious

kind of thing. Well, let me go back to the post and the original motivation, which is, you're right. If you think about your relationship with an insurance company, I remember buying life insurance. I literally filled out forms in triplicate. After they wrote the policy, they sent me letters that I swear were generated by their mainframe application because it comes in one font and it's all caps. and that was the only app the font available

when they had the mainframe application and I really don't have any other interaction with them. Once a year I get this mainframe generated app that said a letter that says send me your premium and so if you think about

the opportunity to completely reinvent all of that we can reinvent all of that with technology with mobile first with data we can reinvent how they find me as a customer

we can invent the types of products that they sell me we can rejigger the price that I'm paying for those policies, we can reinvent how I change claims. So if you just think about it as a startup entrepreneur might think about it, which is let's change every single aspect of the relationship an insurance company has with their insured. Let's just do it.

And I think that's a lot of it. I mean, we're used to at the front end seeing the, you know, the guy go get-go. And so there's a marketing aspect to it. But I want to get deeper into what else can be changed and what the components of insurance are that in our software eats the world way. can be attacked by software. I think the really interesting thing about insurance is it's the first big data problem.

Really, what an insurance carrier is trying to do is evaluate you as a risk. And traditionally, it's been done based upon a lot of factors. It don't evaluate you, but someone kind of like you. So what is your age? What is your zip code? These sorts of things.

But with the explosion of data and with the explosion of this mobile phone that we always have with us, we can really say, who are you? What's your activity? What is your risk level? and allows you to price someone in a way that's a lot more fair to them. You can help them actually prevent claims.

And that's one really interesting thing about insurance companies is if they can use all this data to help my house not getting broken into or help me live longer, we're perfectly aligned there. So it's a really interesting opportunity both to better understand who you're insuring and better pricing and pricing in a more fair way and then also helping make people's lives better. Why is it that the only interaction you have with your life insurance company is basically their demand for your annual premium, right? Because if I'm in their

risk pool, we are perfectly lined. I don't want to die. They don't want me to die, right? Nobody wants you to die. Nobody wants me to die, right? But the insurance company of all the company relationships I have, they really don't want me to die. Right. Because then they don't have to pay out of policy. And so why aren't they at the forefront of encouraging me to live a healthy lifestyle? Why aren't they sending me my jawbone? Of course they should be, right? But it's just, it needs a change of thinking, and I think it's going to need startups to show the way of what insurance companies ought to do in their ongoing customer journey as opposed to the existing companies doing it.

So for many of the startups that come through here and that we talk about and think about data is at the core of things, it would seem to me that one of the advantages that these incumbent insurance companies have is they have all the data, right? They had this historic, you know, if you're some large insurance company, you've been doing it for decades, you know, if not centuries, you know, go back to insurance. during, you know, Clipperships in England or whatever, Lloyds of London. Yes. But is that advantage no longer such an advantage, or is our startups and the way that they

collect data able to sort of quickly catch up and also, in some sense, surpass? What are you saying? I think the data that we're collecting today is very different. Things like your activity data from a mobile phone we've never had before, things like cyber risk that have a very different profile in terms of what makes someone a good cyber risk. You can't look at that in the same way that you're trying to model out a catastrophe like a hurricane. So as we look forward, the types of data that you can get around individuals are very different. I would say the way to think about it is the data that insurance companies had will become less and less important over time as these new data streams get unlocked.

Right. So the old insurance company asked me on paper in triplicate whether I was a – I had ever piloted a plane. Literally, this is one of the questions that you ask when you get at life. insurance. My phone knows exactly when I'm going to get on a plane, right? And so why isn't that data being reflected back into the policies in real time? In fact, why shouldn't it, like, when I book that private plane trip, why shouldn't the offer for a, hey, do you want to buy one-time life insurance for you piloting this plane or you getting on this private plane? Of course it should, right? But insurance companies just don't think that way because historically they haven't

been able to get that data. Therefore, you can't generate the right price for that policy. you can't get it in front of the customer. And now all of that has changed. Like you have the opportunity at the point where you get into a taxi cab or you start your lift driving experience or you get on a plane. All of those could be insurable and that's. So Mike and Frank, you just mentioned sort of a new kind of insurance. And Mike, you mentioned cyber risk. What are the new categories of insurance that you're seeing that are emerging because of our behavior, because of the services that we're all using and because of the way we're honestly just able to do more? I think the sharing economy has been massive. So Uber and Lyft and the whole on-demand world have really transformed and blurred this line between the commercial and personal driving worlds. And there's a huge opportunity there. Same with your house. You know, the typical homeowners policy, never thought that you'd be on Airbnb and be renting out your house. And then also be a consumer, be in another city and using Airbnb and hoping that you have property protection and things like that. So the sharing economy has.

has really been tremendous. And then also things like drones. More men bought drones than eyewatches for Christmas. And it's really unknown. What does it look like when the drone crashes into the neighbor? New things like Bitcoin. You know, it's fascinating. Something like one third of every Bitcoin ever mined is now lost.

But with some very common sense protective measures, it can be a very secure asset. So how does something like Bitcoin get insured? But there's a lot of folks out there that, have wallets that want to make sure that they have good insurance. Are you seeing any patterns? I mean, Airbnb is an interesting example because they're the ones who decided to sort of insure everyone, or at least the hosts, right? And I guess too, I guess, to a certain extent.

And so I, as an Airbnb customer, I don't have to worry. But are you seeing the kind of responsibility shifting because of any of these new trends? Yeah. So if somebody, if you're an Airbnb host and your house gets trashed, right, then Airbnb is ensuring that the losses associated with that, right? So that's not something that your traditional homeowner's insurance did. It's part of the price of doing a transaction with Airbnb, right? So existing products like home insurance are getting sort of modified for

these new sharing economies. And then we have entirely new insurance products, right? So cyber insurance is basically for retailers who are afraid of getting hacked because they're reading the stories of everybody else getting hacked. And they're looking at the costs associated with cleaning up, offering everybody identity insurance. Those are real costs to the business, not to mention the lost sales from people who no longer trust your online site. And so this is going to be one of the fastest growing categories of an entirely new type of insurance, which is I will insure you against cyber losses. Right. And I can imagine the pricing and the risk assessment just must be hard to fathom, honestly. Absolutely. And so we're seeing

startups trying to fill that void, which is let me bring tools and data to this process so that people can write these policies intelligently because you're right. How much will the policy cover? How much will the losses actually be? I like to joke that, you know, I get notifications all the time from the places that I've shop saying, oops, we lost your data. Here's your free year of identity insurance. And I think I'm going to get identity insurance for the rest of my life one year at a time, one hack at a time, right? So there's costs associated with that. There's costs associated with plugging the security holes. There's costs associated with hiring a brand-new insurance team, there's costs associated with lost customers. So how far do the policies go and

then how much should they cost? I think it's also interesting that we see really a convergence of things like cyber risk and traditional P&C. What's PNC? Property and Casualty. Traditional things like a fire, flood, something like that. So, you know, we had in Germany the first factory that was attacked in a cyber way and it caused physical damage. But as we see an explosion of connected devices where all of the machines in your factory are connected to the internet, all of a sudden, that fire or that flood or that shutdown or that interruption of your factory is as likely to be caused by a hacker as it is to be caused by weather or something like that. So it's not necessarily kind of new frontiers, but this convergence. If my car drives itself, then that cyber policy is going

to be the most important form of auto insurance that I have. Do you think in the same way that let's say I have a house that's on a hillside that's prone to landslide or it wasn't built to code, it's uninsurable because it doesn't meet certain standards. Do you feel like there's going to be standards that emerge around cyber risk? And then let's get to the question of driverless cars. If you don't have X, Y, Z in place, the pillars, the sort of underpinnings of good cyber security behavior, you won't get insurance, or will we see these things emerge as standards, do you think? I think standards will emerge over time, and this is sort of what a lot of startups are trying to work on, which is, let's say you were the insurance company and you wanted

to insure your favorite retailer against cyber losses, an intuitive thing to assess would be, well, how vulnerable are they, right? You'd use all the tools that hackers would use to try to figure out, like, how easy it is to break in. Right. And once I've broken in, like, what information can I get, right? So I think there will be emerging standards that let people assess what level of maturity these retailers have on their online presences. I think it won't be government. necessarily. I think it will be commercial standards, right, that sort of put you into a risk category, right? Are you risk one or risk six, right? Right. And you can either pay X or out the nose. Yeah, exactly. And then how do you get from six to one? Like if you wanted to be

in a less risky category, what things can you do to protect yourselves? And you'd expect insurance companies to take a much more active role, right, in the way that I was describing, instead of just the one year cough up your premium, here's that, hey, let's work together to get your security your risks down. And I think if you go back to the really the beginnings of commercial insurance, they've always had that role because the insurance company is really the risk expert. And when done right, they're not just someone you pay premium to you insure your risk, but they're actually your active partner. So, you know, Hartford Boiler came out of the Industrial Revolution, the start of steam boilers. And there used to be explosions everywhere. And there were a lot of

fatalities. And they invented the really kind of the first modern insurance policy. And they said, we're going to inspect every boiler. Only if it's a really good boiler, are we going to insure it? And then we're going to continue to see if it's well maintained. So not only did you have kind of the birth of a massive line of insurance, but you saw safety standards rise universally. And I think a lot of companies benefit from their insurance broker that comes in, whether it's practices of employment or your building or your car. They're often the drivers of safety and partners. So I think it's natural that as we go into the cyber world, the insurance companies are going to be the folks with the most knowledge across the broadest

said to sort of have those standards, but also make their clients a lot safer. And I think that the insurance companies that win in the next century will be those that insure leased, and that they pay the most claims, or they pay the fewest claims, rather, and they're the most preventive. Right. You've mentioned cars and driverless cars a few times. This is one of those things that's a head scratcher, because who's liable? If I'm not driving the car, cars driving itself, you know, who's the professional driver here? It's not me. That's exactly right. I think I was looking at this New Yorker style comic and I know if it was actually New Yorker or not, but it's a police car with another car that's been pulled over. And the caption reads, do you, did your car

know why? My car pulled it over? Right. Like, so right, who's liable? Or imagine this, right? It's only going to be a matter of time before somebody comes in hacks your garage, right? Your Bluetooth-enabled garage, lifts your garage door open, and then hacks your car. and drives it away, right? Because you have a self-driving car. Right. Okay. So who's responsible in that case? Did the homeowner have liability because their garage door was so easy to hack and the car inside it was so easy to hack? Right. Is it the car manufacturer who's liable? Because, wow,

like that was just an easy security hole that you didn't plug when you manufactured the car. So all these questions will get, you know, sort of actively decided as people write these policies, as people pay for these policies. And as the regulatory framework, the regulatory bodies get involved. So it's going to be fascinating to watch. And I would say the early trend has been towards a policy for the manufacturer of the car. So really Google, Volvo, and Mercedes have all said that for their cars, while they're in a self-driving mode, they have the liability. And then services like Lyft and Uber, which will likely be the first delivery mechanism for the mass market for self-driving cars, they're also covering that liability while they're in use. So I think that that is more likely going to be the service provider or the manufacturer.

Which begs the question, then, this is what's interesting about car ownership, et cetera. If, you know, it might become incredibly expensive to actually be your own driver. If the rest of the world is able to be insured because they're not driving their own cars, you know, I want to drive a car. It's probably going to, in the future at least, maybe it would cost me a fair bit. Yeah, and why shouldn't it if, you know, we're humans and we make mistakes and we're, you know, reckless drivers. Maybe as a social good, I should save money, which is the other way of looking at it if I let a much safer computer drive me around. So where are we going to see our insurance kind of experience change first? What's going to grab people? Is it going to be in the marketing

and sort of the front end that all of a sudden we're going to get new options? How do you see this playing out for your average consumer? Yeah. So I think what you'll see is you'll see insurance companies that deliver dramatically better customer experiences. So how you were able to compare prices, how you were to make sure it was apples to apples. So that will change. I think you'll see dramatic improvements in the way insurance companies do customer service, which is submitting claims with iPhones, taking pictures, right? We'll get much more mainstream. And then consumers will see new types of insurance products, right? So they'll get their existing products reinvented, hopefully improved. And then they'll see the new products that,

they hadn't thought about my self-driving car insurance policy, my protect me against hacking my garage door opening. And hopefully there's a lot of things that consumers experience but don't necessarily see. So, for example, if I can use big data to more easily identify who's trying to defraud me as an insurer, it means for everyone else there's a much better claims process. Or when I'm buying insurance on my phone, it doesn't ask me any questions that it should already know because it's smart. So just those, you know, like hopefully those experiences that you have that are technologically enabled that are delightful in their cleverness or their simplicity. Does the universe of insurance, if, let's think about the Internet of Things, if everything is smart and connected and therefore data is sort of ubiquitous, does the universe, can I insure anything at that point, you know, or in some ways bet on anything?

Like, do you see that happening? or are we still, for the near term, going to be focused on kind of the more fundamental pillars of insurance that we know? So in theory, if there were more data freely available, you ought to have new insurance products, right, to capitalize on that. And I think of the weather company, Climate Corps, did a series of experiments where you could buy insurance on weather events, right? So if I'm a farmer in Iowa and I want to say, look, if it's 100 degrees seven days in a row, that's very bad for my corn, I want to buy a policy on that particular event. It turned out not to be a great business for the Climate Court. So I think we need to run more of these experiments because, in theory, we ought to see a Cambrian explosion of insurance types because data is so easy to get. But we haven't seen that yet.

And I think that the other really interesting thing is there's a whole sector of folks that are cut off from the insurance markets today. For example, insurance is very expensive for teenage girls to get, for example, because their male counterparts can be very irresponsible. I know. I know I wasn't a great driver when I was 16. But through telematics, I can show, hey, I'm not the irresponsible 16-year-old. You're saying that teenage girls are treated as if they were teenage boys. Exactly. And there was a company in the UK called Drive Like a Girl. And it said, look, you know, we'll put a telematic device in your car. If you can show that you drive like a girl, i.e. are safer, less aggressive, will save you a lot on your insurance. So all of a sudden, a lot of people were able to save a lot of money. I think a more serious example is diabetics today. It's a huge part of our. a population and it's very difficult to get decent health or life insurance. But the vast majority of diabetics are managing. They're following their doctor's orders, you're exercising, and they're actually pretty good risks. And we can use wearables and devices now to say, hey, look, I might

be diabetic, but I'm actually a really good risk because I'm doing the things that I'm supposed to be doing. And all of a sudden, those sorts of folks have access to insurance policies and markets. So this idea of rewarding virtuous behavior, right, or less risky behavior is great. And one of the really interesting trends is this is going to extend to algorithms. In other words, there are going to be ethical programming that we have to do for our cars. So the prototypical example is, look, you're building the self-driving car and you're in this situation where you can make one of two decisions, right? Kill five kids on the sidewalk or kill 20 adults if you turn left, right? And so eventually we're going to need to program our cars to make these decisions.

And so this idea of sort of ethical decision-making, right, doing something that's sort of better for a specific audience is actually going to come back and influence insurance rates too. What kind of ethical programming does your car have? And therefore, what kind of insurance is appropriate for it. What you need in that situation is the Speed Racer Mach 5 button where it then jumps over the five kids and the 20 adults and like that's the problem. That's the solution. So give me some examples of some entrepreneurs and companies that are at this pushing. the envelope on insurance. Well, we got to start with Xenafits, I think. Okay, so let's talk about our portfolio and then we can talk about other companies too.

One of the fastest growing companies in our portfolio, who would have thought? It's an insurance company. Right. And they figured out a magical way to deliver value to customers in the form of HR software and collect insurance money, basically, as a way to monetize it. And fantastic, right? So they figured out the magical business model. And we love innovation like that. Doesn't have to be a breakthrough algorithm all the time. It was a breakthrough business model. Right. You know, we have a couple others in the portfolio that you might not initially think of as insurance companies, but the applications have been numerous. So smart car is doing some really interesting things around an API to access your car and the insurance applications

there are really interesting. Human API, which allows you to have access to all of your health data in a more electronic way, has had some really interesting early insurance opportunities, making it a lot easier to get underwritten, for example. So these are both, like, in the first example, it's how my driving habits. And the second example, it's my health habits. Exactly. And your health care record. Right.

And then there's some really interesting things. You know, we've seen in the U.S. at least, I think Clover and Oscar are two really interesting approaches to try to solve a health care experience that is expensive and probably suboptimal from a patient perspective and taking really interesting approaches. I think another really interesting company that's flown on. under the radar a little bit is zong on. And that's an insurance company backed by Alibaba and Tencent that's raised almost a billion dollars, an $8 billion valuation, which would make it one of the most valuable private companies in the world,

really bringing digital insurance to the Chinese market using messaging apps and a lot of social distribution, fascinating and really disruptive company. And they're insuring what? I mean, all kinds of things? Yeah, so they started out insuring the shipping cost to return a digital good. So you get something shipped from Tencent or Alibaba and they insure the cost of you shipping it back to them. And now they're doing home, auto, all the different lines you might want. And of that billion dollars raised, as much as that just as sort of the pool of capital that they need to follow through on their insurance promises? I would say in this particular

case, there definitely is a regulatory component. Most of that's going to go out and conquer the Chinese market. Right. And I think it's one of the best examples. People are people are always wondering when the Amazon or the Google or the really big tech players in the U.S. are going to make their real splash. And I think this is a really intriguing example in China where it's happening. In the insurance game. Yeah, because, you know, what you need is money and then you need to be able to market to people. And then, you know, you need to be able to underwrite risk the right way.

But there's nothing, I guess, proprietary about those things. Exactly. And so, you know, what it really sort of foreshadows for us here in the States is where do I buy my insurance product, right? And I think we're going to see a lot of disruption. Today, traditionally, I'd buy it from a broker who's highly trained and scented and, in fact, makes a ton of commissions when they sell me insurance products. Why aren't I buying more insurance products at Costco or Amazon or through Facebook Messenger? And so, like, that will come. Right. Do I buy it from Google or do I sort of rent it every time I take a ride or have an experience? Yeah, that's exactly right. So maybe we

unbundle the insurance products, right? They become more event-driven rather than I cover your entire life. Or the other way to do it is basically the insurance product flexes as my life changes, right? Which is the way life insurance is sold today is basically you figure out, look, how much money will you need? Like, if you die, your kids need to go to college and so on, and they're like, let's add up all those numbers and multiply by some safety factor, and you buy that policy. But the reality is your needs change over time as you make money and save it and your kids get older and maybe you get your kids through college. And so how about a life insurance product that automatically adjusts based on your life situations because

they're already talking to your bank and they know your life situation. Right. That would be awesome, I have to say. Well, this is fascinating. I have to say, I've never been so interested in insurance in my life. Mission accomplished. Frank, Mike, thank you guys so much. And as this continues, and it will, we'll be talking about it more. Fantastic. Great. Thanks so much.