Thank you everybody. I'm Paul Newsome, I cover the Insurtechs, among other things, for Piper Sandler. Very happy to have the CEO of Lemonade, Daniel Schreiber here to chat about Lemonade and technology, and all the fun things that we have there. I apologize in advance, I'm getting over a cold, so I'm going to rely on you to do most of the speaking. Maybe we could talk, begin sort of the conversation with a fairly broad question. We're here at a Fintech X change Conference. How do you see insurance fitting into the Fintech ecosphere? How do you see the opportunities, just in a big picture perspective as it fits in broadly with Fintech?
Good morning. I wish you a speedy recovery.
Thank you very much. I'm on my way. I just sound awful.
A complete recovery. Good morning, everyone. Great to be with you. I think insurance is the most disruptable industry on the planet, but I'll use your question as a way to kind of highlight that. You've got kind of banking and lending and those kind of more traditional financial services, and you've got insurance as somewhat distinct. Actually they're pretty similar in size. If you look at the contribution to GDP, they both hover a few decimal points either side of 3%. If you look at the Fortune 100, you'll find that there are actually twice as many insurance companies as there are banks on the Fortune 100. They're sizable, huge kind of industries, and yet one has seen so much more innovation than the other. I'll give you a few measures of that, but last year, something like $115 billion was invested in Fintech.
About $5 billion in insurance. You're talking about an over 20-fold, 20x outspending by one sector over the other. If you look at the market caps of Fintech companies on public traded exchanges versus Insurtechs, you're talking again at an over 20x. If you look at penetration rates, Nubank and Stripe and Revolut and everybody, you're talking about hundreds of millions of customers, you're talking about single-digit million customers in insurance. All of which is to say you've got these two behemoth sectors, one of which has seen tremendous amount of innovation, one of which has seen it rounds down to zero.
You see this in the incumbent responses as well. Banks and major financial institutions outspend insurance companies about 3:1 on IT. There's a competition going on and they've had to invest, and you see major innovations, open banking. For traditional banks, over 50% of their interactions are now app-based.
You contrast that with insurance where it's still broker-based, going into your local State Farm agent and on the high street. It's really kind of remarkable how distinct they are. I think if you look at all the amazing value that's been created in banking and you contrast it with the value that's waiting to be made on insurance, that should give you cause for pause and ask yourself whether the alpha isn't really on the other side of the fence there.
If it continues to evolve in the way we think it's going to, does that imply a larger role for insurance then as technology becomes improvement, or as it becomes smaller? Well, you hear insurance folks talk about how different types of technology will sort of eliminate risks. I'm curious as to whether or not you think that what's happening from a technology perspective will make the industry bigger or smaller, would you say?
I hope it makes it smaller. The reason I say that is that insurance premiums are a direct correlate of risk and exposure. To the extent that technology can help us mitigate risk, we should be paying less premiums. We see that today with our car insurance, our auto insurance, where we use telemetry. All of our customers do that. We even have amazingly high connections to Tesla, and if you're driving with FSD, which is safer than any member of your household, we'll give you 50% discount per mile driven. As soon as you get to much safer technologies, and in the case of auto, it's lives are at stake, not just dollars, it is true that the cost of repair goes up a bit because these are computers on wheels rather than just mechanical machines, but the frequency drops pretty precipitously.
I would like to see it contract. It doesn't matter. We're talking about multi-trillion dollar sector. These are 11% of GDP is insurance today. You've really got such a huge sector, and at the same time you see I'm wishful thinking saying it would be great if it went down, but cyber exposure and other things like that keep going up. There's an offsetting going on there.
Yes. I've a long time said that the size of the insurance is equal to the size of the claims.
Yes.
As long as there are lawyers and inflation, we're all going to be in business for a long time.
That's pretty right, yeah.
One of the things that I've struggled with, as an outsider looking into the industry, is sort of the competitive moat of technology. I think part of that has to do with the fact that as an outsider, it's just very difficult to tell if company A has better technology than company B. Obviously, your company has been sort of a leader in trying to use artificial intelligence and other technologies.
Yep.
Could you talk a little bit about just sort of how as an outsider we can see that competition other than obviously the results over time?
Sure. It's a challenge. We founded a company in 2015, my first slide deck to my board was about artificial intelligence. Playfully said kind of artificial intelligence, not artificial delays. That was kind of the founding deck. We didn't discover AI in November 2023 or 2022. This is what we've been doing since the founding of the company, and suddenly everybody, of course, is talking about AI. It's like pixie dust that you sprinkle on your earnings, and there's loads of press releases coming out, and it makes it genuinely difficult. Is it AI or is it BS?
I have some theories.
Yeah. I think there are ways to pierce through and have a look, and I'll try and unwrap it for you a little bit. Insurance accounting is convoluted, and it makes it difficult, and they do not offer the kind of metrics that we use to track automation rates. There are two or three things that it's very hard to obfuscate. One is what we call the scaling quotient. What I mean by that is the following. Since ChatGPT came out, take that as a kind of good point, three and a half years, our business has grown considerably. We're talking about almost threefold of the revenue now than we had then. We've added not quite, but close to a 1.5 million customers.
Our gross profit has more than 10-fold increased over the course of three and a bit years, and yet our headcount today is smaller than it was then.
To be able to scale your business like that, see almost 3x the revenue and add millions of customers and shrink your headcount. That's an incredible telltale sign, and it has not been replicated by any of the incumbents. They're not growing at the rates that we're growing. We've now had, since GPT, 10 consecutive quarters of accelerating growth. Not just growth, accelerating growth. We're growing very rapidly, but no change to our operating expenses net of marketing, no change to our headcount over three years.
That is mind-blowing. It's very rare outside of the insurance space. It's nonexistent in insurance. That would be one measure that I think is helpful. Another one which is usually disclosed by incumbents and allows maybe the only true apples to apples metric is something known in the industry as LAE. Stands for loss adjustment expense. That is basically a measure of for every dollar premium I take in, how much do I spend or waste on the bureaucracy of managing claims? Not on paying your claim, but in the overhead.
Which is why it's such a helpful measure of efficiency. The more I have to spend on bureaucracy, the less efficient I am. Now, obviously, there's an advantage to scale. We're at $1.5 billion roughly, to round it there for a second. GEICO, Progressive, State Farm, you're talking about anywhere between $50 billion and closer to $100 billion. They loom over us, 50-fold bigger than we are. Yet, their LAE ratio stands, hovers at around 10%.
One point below, one point above. That seems to be kind of best in class. They're spending something like $0.11 on the dollar on the bureaucracy of handling claims. We're at $0.06.
That halved over the same three years. We have almost threefold more claims today with a smaller claims team than we had three years ago. We're just seeing this explosion of business with no correlating explosion of costs. We've already guided that we think as we double our business again, that might drop from six down to three or four. We've just got this new reality where our variable costs have become fixed costs.
As we continue to scale our business, the profitability just grows as a direct result of that. That gives you a clear snapshot of us versus incumbents who are 50 times bigger and yet twice as inefficient, if you like. There are other places that are harder to measure. For example, the precision with which you underwrite claims. Loss ratio is not a helpful measure. Although we've seen 10 consecutive quarters of improving loss ratios, but I don't actually point to that. You can get there just by raising prices. I don't think that tells you that you're using AI. Here's a nice snippet. We have a $500 million book of pet insurance.
In the last year, the sector, the industry took a lot of rate. They kept raising prices. You saw something like 27% increase in prices across the industry. Our rate increases were less than half of that. We were at 12%, so we took much less rate. We outgrew the industry 3:1 . They were growing at, like 17%. We grew at 50 something percent, 55% if memory serves. We took less rate, we grew 3x faster, and our loss ratios were better. That does tell you about the precision of the pricing. That's not just lazy raising rates.
That is modest raising rates in precise places so that the loss ratio doesn't move, and yet your growth means that you're being priced very, very competitively. I think there are, if you look for telltale signs, if you don't just look at the press release, you try to pierce through, there are numbers to be found all over the place. I'll tell you one other thing, that we are a young and fast-growing insurance company. We are outspent clearly on IT or technology by the incumbency. That is a very poor indicator of anything at all. State Farm, which is the largest P&C insurance company in the U.S., spends something like $3 billion a year on their IT.
Gartner estimates that this year the whole sector will spend something like a $250 billion . Just in the U.S., I think since we were founded in 2015, our competitors have spent something like $1 trillion on IT.
We've spent less than $500 million , so call it $500 million.
Yeah.
We're being outspent pretty dramatically at 2,000-fold or whatever it is. Yet our technology is better than any incumbent technology by a mile.
This isn't something that you can just throw dollars at. If your business model is one where you have human beings selling and being an agent and a broker on the high street, your data collection is appalling. Your ability to harness data is appalling. I saw an interesting study from 2019 that the number one cause of loss for some of the incumbents was other. Garbage in, garbage out. That kind of problem doesn't get solved by just throwing dollars at it.
It relates to that, one thing everyone notices is that the industry's large and there's an enormous range of performance, particularly in personal lines. Could you talk a little bit about how broadly you might compare to some of these companies that are spending a ton of money that are doing much better, right? It's one thing to compare yourself to State Farm or a generic regional insurer, which clearly has green screens and programming that I probably did when I was a young guy. We can tell other stories of that. Versus like a Progressive or GEICO, which have managed to have very low expense ratios overall. As an outsider, why should we think, hey, Lemonade's really gotten ahead of even some of the better ones as well?
Is there something we could point to, or maybe that's just going to be the subset will be those handful of folks who it won't be just Lemonade at the end of the conversation?
Progressive, from an outsider's perspective as well, but my sense is that they are leading the pack. I'm not sure I'd put GEICO in the same breath. At last year's AGM, Berkshire Hathaway, who owns GEICO, spoke about this. Ajit Jain, the Vice Chairman who runs insurance-
said that GEICO is doing very poorly on technology. They've got this wonderful honesty policy at Berkshire, so it's easy to get a visibility there. He said GEICO has 500, and then he paused and corrected himself and he said, "Actually, over 600 disparate systems that don't talk to one another.
Lemonade has one.
It wasn't bought, it was built. We're a culture engineering team. We built it in-house. We control everything. It's the same system that will allocate the marketing dollars and acquire the customers, and we'll use 50 different machine learning models to make predictions about every single person hitting our website. What is the likelihood to claim, to convert, to churn, to cross-sell? We'll amalgamate all of that and produce a lifetime value prediction on every single person hitting our website in real time, the likes of which doesn't exist by any of these systems. We'll allocate dollars based on a kind of an algo trading system of which campaign is generating the best return on marginal dollar spent.
An AI will sell you the insurance once you click on that link and you come through, and we've got our AI, Maya, and she will use all of that information in real time to make the best offering to you to give you the right defaults, to offer you the right add-ons.
When you have a support question before or post-purchase, that will be handled by AI. Finally, when you make a claim, the majority of our claims are handled without a single human being in the loop at any point. Millions of claims. Some people bemoan the lack of the human contact, not our customers. If I can pay your claim in two or three seconds, which we do all day long, you're not bemoaning the lack of the human, "Please wait. Your call is important to us, and it will be answered," and all of that. I think these things speak for themselves. You look at NPS as one metric and you'll see that notwithstanding the fact that we are a cost leader, we are a service leader as well.
Technology lets you do that. With a three-second claim, you're happy and my costs just crush, which is why the LAE is where it is.
Honestly, when you start thinking about the incumbency in general, and I'm not picking on names, GEICO happens to speak about these things.
Please do.
There are structural reasons why it's very difficult to get rid of those green screens or to transform your business. They start with, I know quite a lot of the CEOs of the largest insurance companies, it starts with them having the wrong investor base.
Their investor base want their 5% dividend and total stability, and what's needed is massive transformation.
They have the wrong management team. They were groomed for business preservation, not for business transformation. Their systems date back to the '80s, and instead of having a black box, they have a black hole, and they just throw billions of dollars into it. They have a distribution network. That means they'd love to move to an app-based distribution, but it would sacrifice all of their current channel conflict, which would be unmanageable. They are deeply encumbered. The reason we founded Lemonade is because we don't know of a solution to their innovators dilemma.
If I thought I could just sell them shovels and pickaxes and tell them, "Hey, just bolt this on top of what you have," we would have done that. I think their job is much harder than ours. Starting from scratch as a tech company. Is very helpful. Remember that insurance is the most disruptable, I said earlier, for financial reasons, but think about it from an AI perspective. It is an ephemeral product that is all about statistics. That's what insurance is. I am monetizing probability theory. Everything else is distribution.
At its core, it's about ingesting data, having high-quality data, being a data leader, using machine learning to find the multivariate correlates, and being able to price accordingly, and then serve customers at a lowest cost to serve. One of those maps onto loss ratio, one maps onto expense ratio. Of course.
Yeah.
I think that the fact that everybody else is so encumbered with these old systems and these other issues that we discussed makes it very difficult to drag themselves into the 21st century.
Sort of probably look at some right to script to look at. Your thoughts on distribution. Obviously hot topic with the brokers because.
You're welcome to share my water if you like.
Yeah. Maybe you could talk about what you see as the technology changes, especially recently with distribution. Obviously you have direct channel, but I'll let you talk to it.
Sure. You okay there?
Yeah, I'll be fine.
From our point of view, actually, distribution preoccupies me very little in terms of the AI ramifications. I would just say that the two core metrics in insurance are loss ratio and expense ratio. Loss ratio is about precision of pricing. Expense ratio is about cost to serve. You combine the two, and it's called, rather unimaginatively, the combined ratio.
That's what that's about. LLMs and agentic AI maps onto one, and machine learning and deep learning maps onto the other, which is why it's so disruptable. The distribution piece, as consumers move more and more to LLMs in order to recommend their insurance or even send their agents to buy the policy for them, we're just fine with that. If you now ask your LLM of choice about pet insurance or renters insurance or car insurance, Lemonade will be over-indexed quite significantly.
The reason for that is that consumer set going back to things we said earlier, we tend to be a cost leader because we use technology to get to the best cost. We tend to have the highest NPS. Because of that same reason. Technology is the common denominator to both of those. LLMs scour all the sources out there, ingest that, and then will play it back to you when you ask them who's good at insurance. We're actually finding that we are punching considerably above our weight on LLMs.
To the extent that distribution becomes increasingly headless, where it's your agent talking to my agent, we're fine with that as well. We're not relying on human agents or what we call ogents. We're very comfortable with the machine to machine. Everything that we do is MCP or API based. In some places, like in the U.K., which tends to be price comparison website based distribution rather than anything else, we have the equivalent of algo trading going on there, where we can bid for every lead that comes in very effectively. Being a cost leader has got to be an advantage in the current scheme, but even more so when all of the GEICOs and Cockney accents get displaced by just agents talking to each other and looking at the core facts.
Right.
That's an advantage to us rather than a disadvantage.
The key issue is to be essentially distribution neutral as a carrier set?
I think the key issue is to be a cost leader. To offer the best product or the best service not because you're living on thinner margins, but because your underlying cost structure is automated through AI, and that costs less than humans, and that will then flow through in all the distribution methods.
Excellent. Getting towards the end. I should ask at least a few numbers questions. Fourth quarter, positive EBITDA very focused. Any thoughts about mechanically how that's going to emerge over the course of the year and the sustainability of that over time?
Sure.
Is it just more of the same, or is there something else that we should be thinking about in the next year or two?
I think the simplest way to think about our business model is to think about something that's Y-shaped.
Where the top is tracking gross profit growth. Gross profit is much more helpful than revenue or premium because it incorporates the quality of the revenue. If you've got a bad loss ratio, then you don't get much gross profit. Track gross profit dollars, not margins, because oftentimes higher loss ratios will yield more dollars given the price elasticity of demand. Track gross profit dollars. This last quarter, we announced results which reflected 159% growth year-on-year of our gross profit dollars. This is just a rocket ship.
Track our underlying expenses. If that continues, where we have underlying expenses, net of marketing spend, but all of our OpEx basically stands still.
Gross profit keeps surging, you just know that that translates into profitability. If you look at our EBITDA margin over the last several years, it's a straight up and to the right line. We announced about four years ago that we're going to be profitable, EBITDA profitable in Q4 of this year. That's still what we're saying. The machine is operating in a highly predictable way. We just can calculate the rate of growth and how many gross profit dollars are needed to drop through to the bottom line, and this is our third year of cash flow positive. It's not like we've been burning cash along the way, but the GAAP accounting follows for whatever reason. We'll get to EBITDA profitable, and that will keep, as far as we can tell, that will continue for ever and a day. We're not expecting any near-term reversals on that.
Is there a terminal value to what point you get to true scale and looking at?
This is one of the exciting things. Insurance, the prize at the end of the rainbow is stunning, right? The dominant insurance companies around the world today date back, the young ones, to the 18th century.
You've got the Lloyd's and the AXA. AXA is over 200 years old. Lloyd's is 300 years old. Aviva in the U.K., 330 years old. They grow to be $100 billion, $150 billion a year. That is where our sights are set. We're not planning to slow down anytime soon. We're going to, as best we can, continue to compound. We are in all 50 states. Rather unusually, we also operate in Europe and in the U.K. For some reason, the Atlantic seems to be a barrier for most insurance companies, but we're operating all over the place, and we see tremendous opportunity, which we hope will compound for many years to come. We could 10x our business, and we would still barely be noticeable to our competitors. We'd have to 10x again before they really start paying attention. That's a nice position to be in.
It is. It's big business.
Yes.
Well, I want to thank you guys for.
Thank you.
For being here. Appreciate it very much.
Thanks, Paul.