Insuring the Future with Daniel Schreiber (Lemonade) | Disrupt SF (Day 1)
2:22AM Sep 6, 2018
So it's great. A great pleasure to have Daniel Schreiber from lemonade who is really trying to disrupt this. The insurance space, the insurance, tech space, when they're trying to challenge that industry, that very old rather sclerotic industry that we all know and semi love, rather rather resent, I think often
and it's, it's such an interesting space. So we're gonna have a great conversation now between him. And Danny Creighton, whose Editorial Manager for TechCrunch big round of applause, everybody, come
Thanks, Daniel, for joining us. So ensure tech, probably not an industry. Most people want to come to a panel, you know, at 10am in the morning in San Francisco. But I think today we have a really special treat. So Daniel here is the CEO and founder of lemonade, which I think is one of the most ambitious plays in this space. You go out your first licensed two years ago, in September 2016, raised 120 million from SoftBank just a couple months later, how do you do it? How did you get started? in an industry where you know, there's there's been hundreds of years of practice. It's very mature, as Mike pointed out. It's very sclerotic. How did you differentiate in the space? Well,
neither myself, nor my co founder come from insurance. We've been 20 years in the tech space, founding different companies. And when we encountered insurance, we encountered a sector that is vast, I mean, every household trillions of dollars, no matter by what currency measure. This is a huge industry, which is still dominated by the same companies that dominated 100 years ago. I mean, literally, the dominant companies today were found in the era of the horse drawn carriage,
and it's okay, well, if it's huge and unchanged, but it's loved, then that's fine. But of course, it's deeply unloved. And the Urban Dictionary defines insurance as a promise to pay later that is never fulfilled. So for an entrepreneur, looking at a sector that is that huge, that prevalent that unchanged and that disliked was too much to pass upon. So we really set out to create a different kind of insurance company and being newcomers to the space allowed us to go back to first principles. Rather than study insurance. We did the opposite. We locked ourselves in a white room with a whiteboard. And we said, if we were going to build an insurance company from scratch, how would we want it to be and the resultant company was different by just about every measure, the business model is different, the distribution model is different, obviously, the tech stack is different. And by rethinking those first principles, and having the benefit of no legacy, it allows you to do stuff that incumbents frankly, just unable to do. So let's walk through a little bit of that. So you're a certified B Corp. Let's talk a little bit about the business model and how that's differentiated. So in trying to figure out why people dislike insurance so much. And there's at least 700 years of record of people cursing in insurance companies. And we worked with one of our founding team members, Professor Dan Ariely, the famous behavioral economist who spent 10 years studying what it is that brings out the devil in us, why do we sometimes behave well, sometimes behave poorly. And he concluded that if you set out to create a system with a design intent, or bringing out the worst in humanity, would look a lot like an insurance company, just that everything that social sciences tell us, you must not do are manifested in insurance companies, there's an asymmetry of information, and as a win lose value proposition. And at its core, there's a perception at least, and oftentimes the reality of a conflict of interest, you claim $1,000 of me, if I don't pay you, there's $1,000, I'm 1000 richer, you're 1000 poorer. And that is a short recipe for a complicated relationship. And we decided that if, at the core of the business model, there's a conflict of interest, you cannot build a trusting, trustworthy, lovable brand. On top of that it is gonna manifest one way or another, sooner or later. So in that room where that whiteboard when we started rethinking about things, we decided that what we're going to try and build as an insurance company that takes a flat fee never takes more never takes less. And one of the ways to achieve that is to say, Okay, well, what happens if I charged you and a million people like you so much money, but the cleans weren't that much what I do with the leftover money? Well, I'm not going to throw it into the sea. And most insurance companies take it as an underwriting profit. And what we decided to do is to give it to charity, and to ask you the customer to say, hey, which nonprofit Do you care about, and then you achieve a few things. One is you neutralize I incentive to deny or delay or claim because we're not going to keep that money. Either way, we have set a limit on how much money we can make and making your life harder weren't profit us. And part of that is rooted in the notion that our moral fiber is no better than the next guy. If you put us in a conflicted relationship, that conflict will manifest with us as it does with others. So that was part of the system. And the other part of part of the thinking. The other part is, how do we bring out the best in you, 25% of Americans say it's okay, to embellish claims. I like to think that the other 75% were brought up not to admit that kind of thing.
And so people
feel like they're entitled to embellish claims, which is interesting, fraud and insurance is not hardened criminals from the Far East hacking into your system. It's people like you and me who feel like they may embarrass often criminals except So for us, the B Corp and the charity is actually a core part of our business model. Because what it does is it changes the relationship insurance today is a bilateral relationship, it's me against you, as soon as we designate a nonprofit. And we say, Hey, this is going to be the beneficiary of any under underwriting profits and the leftover money, two things happen. One is I have no incentive to give you a hard time. And in fact, we pay a third of our claims within three seconds algorithmically instantaneously. And I averaged for paying claim is a day. So this is radically faster than what insurance companies typically do. But also you might bring your better angels rather than the devil to the table when you come to deal with us. Because you know that if you embellish your claim, you're not hurting some nameless, faceless behemoth with whom you have a conflicted relationship you're hiding that soup kitchen that you volunteer on and Sundays and that you've designated as a giveback cause. So for us, the whole nonprofit, the B Corp element and isn't just a cherry on top, it's really a core part of our business model had to change a bilateral relationship into a trilateral relationship in order to effect a radically different dynamic.
And let me ask you, so when you think about, you know, you're in the renter's insurance and broader sort of property insurance space. How do you think about competing with, you know, some of these incumbents who have been around a century in some cases, multiple centuries have collected data and compiled practices? How do you do that from scratch and de novo with no data in the database and sort of a blank, you know, my sequel file?
So yes, tell me percent of the Fortune 500 companies or insurance companies, and the average age is 100 years. So you're absolutely right, these these are companies that have established themselves and survive for centuries, oftentimes, and they have no doubt and massive data advantage. And us and data plays a huge role in insurance. predicting the future is entirely based on using data sets from the past.
And in our first couple of years and market, we've been at a data disadvantage. And so we started off with a wild ass guess we're kind of graduating the data comes into an educated guess. But we are still at a data disadvantage compared to incumbents. But we got about 100 times more data for every customer that we gather, because of the digital interaction. Everything with us is done through chat bots. And everything is done digitally through AI. And when you walk into your local State Farm agent on the high street, he or she will take maybe 20 data points from you will take about two orders of magnitude more data points, just because of the data sets available to us. So when take us 100 years to bridge that gap. And I guess to meet that we're probably 18 months, maybe 24 months away from parity, and there after those systems become massively advantageous. So I think that while we're still playing a game of catch up, and we're seeing the finishing line ahead of us, and we're hopeful that being built on a digital substrate with data and AI built into the fabric of our company will allow us to ultimately outperform not merely on user experience and instantaneous payment of claims, but in the hard work of predicting and pricing risk. Now, let me ask you, and,
you know, obviously, you've had incredible growth, I think the statistic was, is that in the first quarter of this year, in 2018, you had the same revenue that you had all of last year, and you had 1,000% growth, and this is first six months of this year, but it was reported in the last two weeks that you had a net underwriting loss of one and a half million and a net combined ratio of 138%, which meant that for every dollar and premium you took in, you paid out about $1, 40 out so that the bottle wasn't profitable, you know, how does that change in the next, you know, a couple of years? How do you sort of get to profitability, particularly on the coarser of insurance part of the business
for lemonade? I think that's exactly right. And it ties back to the question that we just discussed about data. So for us the first few years, we kind of wrote them off as a learning experience, until you have data, you can't have an expectation of pricing risk reliably, what you want to see as a project a project progression, rather, so that as the data accumulates, you become better and better at this underwriting and loss ratios. its own in the industry for the first half of 2018 was half of what it was in the first half of 2017. And we seeing that kind of progression gives us confidence. So it's a trend line that we're expecting. So yes, it's true that we cannot sustain the level of losses that we had in the first few months. But the trajectory is exactly a we would want it to be, in fact, ahead of where we expected to be. So it isn't an underlying concern. But let me develop this further. For a minute. I met with one of the heads of one of the largest insurance companies in the US and we'll having a chat and he says to me, hey, Daniel, our loss ratio is 54%, which passes for small talk amongst insurance people.
And the loss ratio is for every dollar that I taken in premiums, what percentage do I pay out in claims? 35 percenters is a good number. And as is higher than that, and just said,
and I said, you know, the average American has one testicle
and in you don't hear that every day. No,
it doesn't pass with small talk amongst,
and he said, he looked at me, somewhat bemused as you might expect, and I sent him averages are kind of meaningless. What happens is that insurance companies because they onboard with these 20 data points and equals millions. And what I mean by that is, they look at millions of people as being more or less identical, because they don't have the resolution to tell them apart. So they look at these massive group of people, and they say, our loss ratio, and they take 10 million people, and they talk about them as if they were one monolith.
When you build insurance on a digital substrate, you get to about 100 times more
resolution. And just two years in market, we already finding that those groups that traditional insurance companies regard as monolithic are actually made up of subgroups. And we're already identifying three or four subgroups with three to 400% difference in their behavior and their loss ratios. But you need that degree of microscopic resolution to determine that. And we think that ultimately, this gets to perhaps not n equals one, but a much higher degree of precision. And therefore, ultimately, as you play this movie forward, you find that the digital infrastructure of an insurance company changes the very core of underwriting and pricing risks in a way that hasn't manifested yet the first act of lemonade, and companies like ours is about delighting consumers learning cost of acquisition, paying claims in seconds, act two is about using that those delightful apps, they generate phenomenal data, which ultimately allows you to price risk at a far higher resolution. So
give us an example just to make sure that I'm following this. So you know, in the old world, you might have zip code, is it a mount multifamily building? Or is it a single family building? And that was sort of the pool that you were in the New York, you know, large apartment building property pool? How do you think about categories? Like, what resolution Are you at today? And what would that looked like in two or three years? So
this is really goes to the very core of lemonade IP. So I'm not going to talk about the exact elements. But anytime you interact with an app, you're generating mountains of data that are otherwise lost. So for example,
when somebody goes through an app, do they bother reading the fine print or not? Do they play with default? So do they not? what time of day? Are they doing this from one devices? How many times what you know what channel did they come through you just getting mountains and mountains of data which turned out to be highly predictive of risk in that same conversation with the guy after kind of mentioned my comment that my distasteful comment
I said to him, he was asking me about our loss ratio. And I said to him that, you know, it's still a bit early. Yes, we have hundreds of thousands of customers. This is a month ago now, but we don't yet have enough claims claims always lag, and you need enough claims in order to find all the correlates. And he smiled at me. He says, Well, I can tell it, you guys are different, because no insurance company would ever say the words we haven't had enough claims.
So it really is that question. Oh, yes, allowing the system to play out. And I like to think I was never fly on the wall at Walmart all these years ago. But I like to think that when Walmart first encountered Amazon back in the 90s, they realize that this was something differently, took an interest. And they looked out the window. And so what was going on? And they said to each other, well, we'll see if this web thing catch catches on, we'll build a website as well.
And I think we're kind of at that level with insurance companies, they're all looking at us. They're curious, they passed the backhanded compliment of saying it's a neat app. And they think, well, if this thing works, we'll build an app as well. And what we've learned from Amazon 20 years later, is that building a company on a digital substrate, it's not about a neat website, that facade is the kind of outermost expression of a profoundly different company that every element from HR technology to processes to consumer interaction to every aspect is radically different. And I think in insurance, this will become even more the case where at the moment, it looks like, oh, they've got an app will create an app. But the level of differences so profound, that what happens is that with, as with Amazon, 20 years later, we're not seeing convergence, we're seeing continued divergence. And today, they passed a trillion dollar market in terms of market cap. And you just see that year after year after year, the differences become more pronounced rather than less. And with or without lemonade. I think that the insurance side he's going to face that kind of Amazon moment, over the course of the years to come.
We talked a little bit about IP, and you're, you're sort of building up this sort of data model the summer you had a little bit of a legal kerfuffle with a German sure one insurances we Fox sort of product, you ended up having to do a cease and desist. But they were copying the app, they were doing a bunch of claims to try to figure out exactly how your model work and how do you think about, you know, the IP side about it and, and protecting it as well as, you know, staying open to the values of like a B Corp, you know, how do you sort of balance those two sides.
So I wrote a blog post a few months ago about giving out awards for the number one copycat of
and actually Liberty Mutual took the prize that created a website at the time, which I've since closed down called Ludo, which I figured would be called apple juice. Yeah,
there you go.
And we're torn between. On the one hand, I think any company any company that uses elements of what we're doing to be more customer centric, to add charitable components to the business model to deal with consumers fairly and in an expeditious way to reduce the hassle of making claims, because it tarnishes the reputation of our industry. All of that is fabulous. The only place where we draw a line is when people break the law, and they don't relate to the particular case they asked me about, but align passes very family where the law laws get broken. And everything within the bounds of the law is fair game.
And let me ask you, I mean, talking about the law, you obviously work a lot with regulators, and how do you how do you build a relationship? So for those who don't know, in the insurance space, muster the regulations at the state level, not at the federal level? So you're not just actually working with one regulator, you work with 50 state regulators all across the world, the country, how do you sort of engaged that men and and sort of sell them on these sorts of new models? I mean, has it been harder in some states and others have everyone sort of understood, oh, this is great. You know, we want everyone to sort of move this model, what's the divide there?
And we do have have a wonderful relationship with regulators. And I'm not just saying that, because there's an audience, we do, but it is a challenge. I don't underplay it. Yeah, and I think actually, we are unusual in the insurance space in the technology insurance space for being a regulated insurance carrier. And so we believe that to affected profound change, you can't just create a pretty website, a nice facade in front of 100 year old edifice, the amount of change that you can affect that way is very, very limited in in our telling the core problems of insurance for reasons that we've already discussed, like the business model level, and you have to change that you have to neutralize conflicts of interest, you have to change the policy language, there's just so many things have to be changed in the plumbing of the business rather than just that website that you put up in front. So we did want to create our own carrier, get our own licenses and not be beholden to the incumbents and and for regulated That was tough. I mean, the idea of having a novel business model, we give underwriting profits to nonprofits. It took them a while to get their head around that but I think if you come with clean hands with experience people, you know, we had on Dan Ariely on the behavioral economics and we had the President of Product from AIG and the chief underwriting Officer of Liberty Mutual, we really built a whole team of people with deep expertise, each in their own domain, and you come well funded, so Sequoia and Google and SoftBank that carry some weight as well. And you bring that all together to regulators, and then they run water through the pipes. So they, for example, took all the data from Sandy the disaster that hit back in 20,012, and ran it through our model to make sure that would be able to withstand that. And ultimately between that and having strong reinsurance partners, and getting a rated by rating agencies, all of that comes together, and then they monitor you. And now two years in, I think they feel really good that they give us the license, but I will tell you this, it was touch and go. Yeah, it could have ended very differently. I think he flipped that coin 10 times. Maybe it comes up heads once.
I'm not sure that I would. Now knowing what I know, I'm not a smart thing to do.
And how many states are you giving your product today? So today,
we have about 60 something percent of the US population,
and that's keeps growing. Got it. And what are the new initiatives you just announced? I think last week was this thing called a policy 2.0. Could you talk a little bit about that? Yes.
This is a great example of why you want to be a carrier and why just creating a facade doesn't work. So we have renters insurance, we do homeowners, we do renters, we do contests, all different kind of product, we were licensed to sell any property and casualty insurance. So you take the renter's insurance policy. Now I'm a newcomer to insurance. I've been in it for three years,
and I'm a recovering attorney. I've been clean for 20 years. But
I'm a recovering attorney. And my English is possible. And I could not understand our own insurance policy. Honestly, we started off with an industry standard policy. That's what you do. And that gives comfort to regulators and others that using an industry standard policy. This thing was 20,000 words long, many of them from Middle English. I'm not talking Shakespeare. I'm talking Chaucer.
And despite my training and everything else, honestly, I misread a lot of stuff. And I'm thinking how can you build a trusting and trustworthy product, when the core product is impenetrable to people who even moderately trained, let alone your consumers? And how do you present that in an app and a five inch screen with a few icons, when there are so many exceptions. The policy language in a standard policy has exceptions running six levels, deep exceptions to exceptions to exceptions to exceptional success. And I am not exaggerating, I counted them
and the several hundred exceptions, and you just think, Jesus, how you do this. So what we did was, we went back to the drawing board, we wrote a policy from scratch, I actually spent several days on this myself, I thought, Okay, I'm going to write this in English, like a blog post. And we wrote it, we took down out 90% of the words, we only use contemporary language, it reads seamlessly, it introduces some legal risk, because usually normal terms rather than terms of art, and then there could be some litigation about what those means. But we think that if you are looking to help consumers, and put it in plain English, ultimately, that's, that rewards you in the end of the day. And let me stack it up on GitHub. So we took this policy, we did our best of making it work. And we said, we're going to do two things. One is we're gonna make it an open policy. So any of our competitors may use it, that's unheard of, we currently pay a vendor for using their policy is we said, this is open source, anybody can use it. And secondly, anybody can contribute. And people are going on to GitHub and adding comments, and making suggestions and doing pull requests. And that's how we're writing the next generation of policy, you guys can write their next generation policy from your laptops right now.
And let's go to talk a little bit about a moral hazard. So I think one of the challenges, you know, you're simplifying these policies, obviously, making a product that connects with more and more consumers, you have this issue of moral hazard, which is sort of where people's behavior changes as they buy the product. So they buy insurance, and that they've therefore take larger risks, they set fires in their house, and then demand for claims, how do you how do you sort of handle that part of the business around around moral hazard,
we don't see a lot of that. And I think actually, that what happens with insurance typically, is that people are reticent to make claims because they perceive it as being a horrible process.
And if it's a claim that you can afford to handle yourself, say, to hell with it, I just don't want the headache of weeks of paperwork, fax machines, which I don't even have any more and all that kind of stuff. So we've radically simplified that. And now you make a claim with your app, you do it in seconds, and about a third of our claims get approved instantaneously by the but and the money is back on in your account seconds after you press Submit. And that that creates not what I regarded as a moral hazard. But that creates an incentive to make more claims. And our answer to that is, we're fine with that. When you buy insurance, you buy insurance, and ready to get paid when a genuine event happens. What we're not fine with is fraudulent claims. But if you are legitimate customer, you've paid us money, this is what you paid it for. And I always tell our team, we're in the business of paying claims. That's what we are here for. That's what our customers want of us. And once we neutralize our own moral hazard by saying we're not going to make money by denying claims we don't get to keep that money. You set that aside, then you're all about delighting the customer and use technology to do that use behavioral economics. We have this giveback component, but it's really about if it's a legitimate claim, pay it and paid as quickly as you can and delight customers. There's a site called clear assurance that does reviews of websites. It's a consumer feedback site. We have nothing to do with them. They ranked 270 of the insurance providers in the US two years since our launch, we now ranked number one in the US in terms of customer satisfaction. And I think it's largely to do with that we have about a minute here left on stage. Give us a short kind of picture of the next couple of years for the company. You already raised. SoftBank round,
two years out of out of licensing. What's next for the company? How do you how do you even spend 120 million bucks
our ambitions are expensive. And so when we look back over two years, we feel deeply grateful for had the stars aligned have how we've been backed by such venerable companies by the team that has joined us by the customers are put trust in to us. So we feel like we've achieved in two years more than would set out and we feel very, very grateful for that. But frankly, as measured against our ambitions, we've barely made a dent, we would like lemonade to be one of the dominant insurance brands of the 21st century. And by that measure, when our competitors are hundred years old, and 100 billion dollars in business. We've got decades of work ahead of us. So it will be about expanding into new products. And we're talking about in the coming months expanding globally and in an industry that measures itself in trillions of dollars, really, there's endless growth opportunities ahead of us. And that's what we'll be focused on.
Daniel, thank you so much for joining us on stage and thank you.