LMND - NYSE

Good morning, and thank you all for attending the Lemonade Q3 2025 Earnings Call. My name is Brika, and I'll be your moderator for today. [Operator Instructions] I will now hand over to the Lemonade team to begin.

Good morning, and welcome to Lemonade's Third Quarter 2025 Earnings Call. Joining us on our call today, we have Daniel Schreiber, CEO and Co-Founder; Shai Wininger, President and Co-Founder; and Tim Bixby, Chief Financial Officer. A letter to shareholders covering the company's third quarter 2025 financial results is available on our Investor Relations website at lemonade.com/investor. I would like to remind you that management's remarks made on this call may contain forward-looking statements. Actual results may differ materially from those indicated by these forward-looking statements as a result of various important factors, including those discussed in our letter to shareholders and the Risk Factors section of our Form 10-K filed with the SEC on February 26, 2025. Any forward-looking statements made on this call represent our views only as of today, and we undertake no obligation to update them. We will be referring to certain non-GAAP financial measures on today's call, including adjusted EBITDA, adjusted free cash flow and adjusted gross profit, which we believe may be important to investors to assess our operating performance. Reconciliations of our non-GAAP financial measures to the most directly comparable GAAP financial measures are included in our letter to shareholders. Our letter to shareholders also includes information about our certain performance metrics, a definition of each metric, why each is useful to investors and how we use each to monitor and manage our business. With that, I'll turn the call over to Daniel for some opening remarks.

Good morning, and thank you for joining us to review Lemonade's results for Q3 '25. I'm happy to report another very strong quarter. Our in-force premium grew to $1.16 billion, marking our eighth consecutive quarter of accelerating growth. Our revenue was up 42% year-on-year, while our in-force premium enjoyed 30% growth, growth rates we were not expecting before 2026. Happily, our strong top line metrics were matched by our profitability KPIs. Our gross margin climbed into the 40s, while our gross profit more than doubled to $18 million, propelling us steadily and predictably towards EBITDA profitability in Q4 of next year. All our products and regions contributed to this dynamic of accelerating top line and improving profitability, though it is worth spotlighting car, which saw 40% growth with more than half of that coming from existing Lemonade customers essentially CAC-less acquisition. That's transformative to car's unit economics as is the 16% year-on-year improvement in car's loss ratio, which came in at a lovely 76%. Staying with loss ratios, our company-wide gross loss ratio in Q3 was 62% and our trailing 12-month loss ratio was 67%, both our lowest ever. If nothing unexpected happens in the coming weeks, I anticipate that we will set a new record once more this quarter, Q4. Against this backdrop, it's worth remembering that while declining loss ratios and expanding gross margins are a thrill, they are not per se what we are solving for. As I explained at some length during our Investor Day 1 year ago, the metric we are looking to maximize is gross profit dollars. Loss ratios always affect gross profit but not always as a simple counter movement, whereby lower loss ratios yield higher gross profit. In reality, the relationship is non-onotomic, meaning that often a higher loss ratio will yield higher gross profit. The underlying mechanics are obvious when you think about it. Given the incredible price sensitivity in insurance, each percentage reduction in price can often yield outsized returns in terms of conversion and retention. Lower prices worsen gross margins and loss ratio, yes, but the attendant boost in revenue often more than makes up for that. This means that for some parts of our business, certain products, certain channels, certain segments, a higher loss ratio and slimmer gross margins will actually translate into higher gross profit. Given the choice, we will always privilege dollars over percentages, which is why when we see an opportunity to trade higher loss ratios and slimmer gross margins for higher absolute gross profit dollars, we will take that trade 10 times out of 10. And indeed, as noteworthy as our loss ratio progression has been in these past 2 years, during that time, our gross profit has surged by 261%. The full significance of this comes into sharp relief when paired with the fact that during the same time, our underlying expenses increased by single digits. This means that we've essentially transformed our variable expense into fixed costs. That's extraordinary. It's the hallmark of an AI-first company, and it is the reason why our gross profit trend line charts our path to profit and beyond. And with that, I'll hand it over to Shai.

Thanks, Daniel. I wanted to shed some light on something that captures one of the ways AI shows up in our results, the LAE ratio. For those less familiar with our industry, LAE or loss adjustment expense measures the cost of handling claims as a percentage of premiums. It's a simple but powerful indicator of operational efficiency, and it is one of the few metrics that truly allows apples-to-apples comparison of the underlying efficiency of different insurance companies. It should be noted, though, that this metric is influenced by economies of scale. And so the larger the insurer, the more they are expected to have a good LAE ratio. For reference, large carriers typically report around 9% LAE. In other words, they spend about 9% of their premium dollars to handle claims on top of the claim payment itself. I'm happy to report that our investment in automation has been paying off. And despite our relatively small size in comparison to the largest U.S. carriers, we reached a superior level of efficiency with an LAE of 7% on average across all of our products. In fact, in the past 3 years alone, we've cut our LAE ratio in half and the number of Lemonade claims adjusters actually declined, all this despite our claim volume growing 2.5-fold. Using blender, our AI-powered insurance operating system, claim adjusters are able to handle 3x the claim volume they could before, all while providing our customers with a more transparent and instant experience. But having the best-in-class LAE is not where we stop. We wanted to take this further and expect to cut the LAE ratio in half yet again in parallel with our next doubling of the business. With that, I hand it off to Tim, who will cover our financial performance and outlook. Tim?

Thanks, Shai. Let's start with our Q3 scorecard. In-force premium grew 30% year-on-year to $1.16 billion, driven by customer growth of 24% and premium per customer growth of about 5%. We added a record 176,000-plus net new customers in the quarter. Gross loss ratio was 62%, an improvement of 11 points year-on-year and 5 points sequentially, while trailing 12 months gross loss ratio improved 3 points sequentially to 67%. Prior period development was 5% favorable, driven by 2% unfavorable CAT prior period development and 7% favorable non-CAT prior period development. Total CAT in the quarter, excluding the cat prior period development was 4% -- favorable prior period development was driven primarily by home, car and EU operation, while the unfavorable CAT development was related primarily to the California wildfires in Q1. And on a net basis, prior period development was similar with non-CAT about 6% favorable and CAT 4% unfavorable for a net impact of about 2% favorable. Prior year development, which we report on a net basis, was $6.3 million favorable in Q3 and $18.9 million favorable year-to-date. Gross profit more than doubled to $80 million as did adjusted gross profit to $81 million for a gross margin of 41% and an adjusted gross margin of 42%. These metrics use revenue as their denominator. Adjusted gross profit as compared to gross earned premium was 29% in Q3, up 11 points from 18% in the prior year. Revenue grew 42% to $195 million, while our adjusted EBITDA loss improved by about 50% in the year to a loss of $26 million. And it's worth highlighting that revenue grew fully 12 percentage points faster than IFP, a dynamic we expect to continue through at least Q2 next year, primarily due to the recent increase in retained business through our quota share reinsurance structure renewed July 1. Our Q4 revenue guidance, in fact, implies a roughly 49% year-on-year growth rate at the high end of the guidance range. Importantly, adjusted free cash flow was positive for the second consecutive quarter at $18 million, while operating cash flow was positive $4 million. And we ended the quarter with just under $1.1 billion in cash and investments, of which $278 million is held as regulatory surplus. Annual dollar retention, or ADR, began to improve again as expected and was up 1 point to 85% versus the prior quarter. Operating expenses, excluding loss and loss adjustment expense, increased by $17 million or 13% to $141 million in Q3 as compared to the prior year. And let's break those expense lines down a bit. Other insurance expense grew by $4 million or 22% in Q3 versus the prior year versus a 30% growth rate of in-force premium. Total sales and marketing expense increased by $6 million or about 12% due to increased growth spend versus the prior year. In Q3, that growth spend was about $46 million, up 16% as compared to the prior year. We expect Q4 growth spend to be at a roughly similar level, which would put us at a total growth spend of about $180 million for the year. We continue to see both ROI strength and diversity across growth channels, where we've been able to maintain our LTV to CAC ratio above 3:1 across products, across channels and across geographies. Technology development expense was up 13% year-on-year to $25 million, primarily due to increases in personnel expense, while G&A expense increased 11% as compared to the prior year to $35 million, primarily due to an increase in interest expense. Headcount decreased sequentially from 1,274 in Q2 to 1,259 in Q3 and was up about 3.5% versus the prior year and essentially flat versus 24 months ago. Our net loss was $38 million in Q3 or a loss of $0.51 per share as compared to a net loss of $68 million or $0.95 per share in the prior year. Our adjusted EBITDA loss was $26 million in Q3, significantly improved versus $49 million EBITDA loss in the prior year. We're well positioned to continue to fund this growth to expand across geographies and continue to diversify our customer mix. With over $1 billion of cash investments, efficient capital surplus management and positive adjusted free cash flow, we're well positioned to fund our growth strategy without need for additional capital. Given strong year-to-date performance, we are raising our full year 2025 guidance across in force premium, gross earned premium, revenue and EBITDA loss. Our expectation for positive adjusted EBITDA for the full quarter of Q4 2026 remains unchanged. And with the recent change in our quota share ceding ratio, we expect our ceding rate to continue to decline in Q4 to roughly 40%. Our Q3 results show continued execution on and ahead of our targets, 30% premium growth, double-digit loss ratio improvement, a doubling of gross profit, revenue growth well outpacing premium growth, recurring positive cash flow and a strengthening balance sheet. We are delivering a unique combination of growth and profitability improvement and are doing both at scale with real discipline. Let's talk through our Q4 expectations, and then we'll take some questions. For the fourth quarter, we expect in force premium at December 31 of between $1.218 billion and $1.223 billion, gross earned premium between $283 million and $286 million, revenue between $217 million and $222 million and an adjusted EBITDA loss between $16 million and $13 million. We expect stock-based compensation expense of approximately $18 million and a weighted average share count of approximately 75 million shares for the quarter. And this implies for the full year, gross earned premium of between $1.044 billion and $1.047 billion, revenue between $727 million and $732 million, and adjusted EBITDA loss between $130 million and $127 million, stock-based compensation expense of approximately $61 million and a weighted average share count for the full year of approximately 74 million shares. And with that, I would like to pass over to Shai to answer some questions from our retail investors.

Thanks, Tim. We now turn to our shareholders' questions submitted through the Say platform. Paper Bag asked, with the Local and L2 announcement, what tangible things will be accelerated as the number of car states we plan to launch in 2025 and beyond changed? Are there any new products planned to be coming out faster? And will we see further operating leverage in our engineering teams? Thanks, Paper Bag. The local platform represents a major leap forward in how we build and evolve our insurance products. And yes, it's already accelerating a lot of what we do. For those who aren't familiar, Local is what we call our next-generation LLM first no-code insurance product builder. And it effectively gives our teams a new way to configure, design, test and launch complete insurance products and experiences without needing to write or deploy code. Local is being built in a modular way. It is already deployed and delivering value in some parts of the business, even though much work remains before local is complete. And based on the rollout so far, processes that used to take weeks can now happen in hours. And yes, it accelerates our operating leverage by freeing our engineering teams to focus on higher impact initiatives since much of the product improvements and tests we're doing can be handled directly by our product and actuarial teams with no engineering involved. Paper Bag also asked, what is the reason or rationale for the recent board seat nominations of PayPal's CMO and Meta's VP of AI Product and are there potential partnerships with either company in the works? Paper Bag, the rationale for the additions of Jeff and Prashant to our Board is that both of their areas of expertise, AI and brand are central to Lemonade's strategy. Jeff and Prashant each bring exceptional experience that aligns directly with where we are headed as a company. Prashant is Meta's VP of AI Products, prior to which he was Meta's VP of Generative AI, giving him a unique insight into how cutting-edge AI can be deployed at scale. Jeff is CMO at PayPal and Venmo and was previously Global Head of Marketing at Airbnb. So he has shaped some of the world's most loved and enduring consumer brands. As we continue to leverage AI to deliver delightful customer experiences and ultimately transform the insurance industry, their experience and perspectives will be invaluable in helping guide our next phase of growth. There are no specific partnerships with either company to highlight at this time. Our rationale is strategic expertise and not corporate collaboration. There are several questions about the future of FSD and how we are positioning our car insurance product in that shifting landscape. So I'll share some thoughts responsive to that general theme. This is an area we pay close attention to. The time line for widespread autonomy is uncertain. It could take longer than the optimists predict or accelerate faster than most expect, and we're building with that range of scenarios in mind. Whenever autonomy reaches its tipping point, we believe we are well positioned to capture what many incumbents might see as a threat. The shift toward autonomy plays to our strengths. The future of car insurance is increasingly about pricing per mile driven and distinguishing between human and system-driven miles. Our system is built around usage-based pricing, real-time data and flexible coverage, precisely the infrastructure needed for that future. And we don't have any legacy systems or traditional business models holding us back. Lastly, there were a number of questions about our Tesla integration. We recently announced a direct integration with Tesla's API, which with proper customer consent allows us to pull driving data straight from the vehicle. This gives us access to a much richer and more precise array of data than what's possible through a phone app or plug-in device. Things like seatbelt usage and more accurate trip insights, for example. It's the kind of granular telemetry that becomes critical as cars get smarter and more autonomous, data that not only sharpens our pricing and underwriting precision today but also positions us to learn directly from the evolution of FSD systems over time. As for ensuring FSD miles at near 0 cost, we aren't able to share material updates on that at the moment but promise to do so when we can. What we can say is that integrations like these are early building blocks for the future where usage-based and system-driven pricing becomes the norm and where our platform is already designed to adopt. And with that, I'll pass it over to the moderator, and we will take some questions from the Street. We.

[Operator Instructions] We have the first question from Tommy McJoynt with Keefe, Bruyette, & Woods.

You noted about half of new car customers were existing Lemonade customers and thus were effectively CAC-less. How does that level compare to prior periods? And is the plan for the majority of new car customers for the foreseeable future to be CAC-less?

I would say that, that 50% rate has been consistent, plus or minus for a few quarters now. So it's a good number. It's a stable number. The CAC-less approach is without question, part of our focused, driving customers to multiple policies. We've seen growth in the multiply policy rate above 5%. It has increased sequentially every quarter for quite some time. But I would think of that 50% plus or minus number is a good stable number that we expect can continue.

Just to add to that, in addition to these customers being CAC-less, they are remarkable in other ways. They tend to have much better loss behaviors, loss patterns. So they are less costly not only to acquire but to service. They tend to have higher retention rates. There's a lot to love about these cross-sold customers. We -- in the letter we refer to them or in my comments refer to them as CAC-less, it would be more accurate to almost think about it as negative CAC. These are customers that tend to be profitable in whatever line of business we acquire them through and then they add a car policy on to that. So it's really an important part of the business. I will just add that while 50% or as we said, over half of our customers coming this way is a big deal. It's a core plank of our strategy, always has been. That part grows more organically, whereas the one that we target is less organic, and we have dials that we can dial that up or down. So the more you find us spending on acquisition, the more that will affect those ratios over time.

Got it. And then switching over, looking at the ceding commission revenue line, was there a contingent or profit share tailwind in that ceding commission in the third quarter? It looks like it was a higher percentage of ceded premium than it had been running at.

The bulk of that ceding commission is driven by loss ratio and because the loss ratios came in quite nicely, a record low in the quarter. As you know, there's a sliding scale of commissions. So the commission varies somewhat up and down based on the loss ratio. There's a cap and a floor, a high and a low. So at some point, you cap out when your loss ratios get really, really good. So that was the main driver in the quarter. And probably worth a reminder, the ceding commission that you see on the face of the P&L is about 4 points different than the actual ceding commission, and that's an accounting nuance. I think you'll see an effective ceding commission rate of about 28% in the quarter. But on a P&L basis, because of the accounting nuance, you see about 24%. So you're exactly right, a couple of points better both year-on-year and sequentially.

Your next question comes from Jason Helfstein with Oppenheimer.

I'm going to try to sneak in like 2 and then a quick housekeeper. So obviously, we're seeing like impressive improvements in kind of the contribution ratio efficiency. No doubt you are finding ways to use AI to make the business more efficient. That being said, where would you rate yourself on like at a 10, this would be us using all of the AI tools out there that we could and where you are? That's question number one. Question number two, again, you've got the business dialed in now between kind of growth and marginal contribution improvements. Is there anything philosophically to think that you're going to lean more into growth because of the way the business is and the metrics are playing out? And then lastly, Tim, just expenses were up on a year-over-year basis and sequentially in the third quarter, like OpEx, i.e., technology and G&A more than we've seen in a while. Just is there just anything to call out from an expense standpoint in the quarter?

Jason, -- so the AI is now -- the impact of our AI deployments, I think, is now reflected on pretty much every line in our P&L. So you're quite right. We see it almost anywhere you look. We spotlighted the LAE as a way to really provide apples-to-apples comparison and that way you can see how dramatically different it is from the incumbency. You could look at the fact that we've OpEx -- sorry, our gross profit has gone up tenfold in the last 3 years, whereas our headcount hasn't moved, has actually moderately declined. So there are a lot of indications of something pretty dramatic happening in terms of the AI, and we see that in terms of all the efficiencies. And if you look at the life cycle within kind of the customer engagement with Lemonade, you'll see AI everywhere. It starts with how and where we deploy the marketing dollars that attract you as a customer. So as you know, about 90% of those dollars that Tim referenced earlier that we deploy to acquire customers, about 90% of them are guided by AI, some 50 different machine learning models that optimize how we spend, where we spend based on LTV to CAC predictions of every customer, every segment, every advertising campaign. Then when you come to us, our recommendation of products and also some other settings and cross-sells during the purchase process is AI-driven. And then later, when you engage with us and ask customer support or claims, and you'll see again the majority of our claims being settled without human intervention by AI. So by one measure, I would say that we score very high on your 1 to 10 scale. We really do use AI across the board. The majority of our code, software engineering is now written by AI. So we're really seeing this everywhere. At the same time, I think if you take a zoomed out perspective and you kind of judge today by where we will be a year, 2 years, 3 years from today, I think you'd rate us as a. I think we're just getting started. And there is so much more that we see that we can do. We're scrambling to do it all. As we are doing that, the ground beneath us is shifting because models are becoming so much smarter, so much faster. So I think both we have done a lot and we have done very little, one measured against what the industry knows and the other one measured against the potential that we see coming over the course of the next few years. Another word about your second question. Can you -- Tim, do you have that?

Yes, it was about just philosophically now that everything seems to be kind of working would you consider leaning into more growth and pushing out like kind of profitability targets and then there would be housekeeping expense.

Okay. So yes and no, and we tried to touch on this in our earlier comments. We see ourselves turning EBITDA profitable in Q4 of next year. That's not moving. I don't anticipate any change in that. That's been our expectation for some 3 years, and it's becoming increasingly obvious, I think, to people outside the company and why we're so confident of that. So that particular profitability metric is unlikely to change. But there are other metrics that talk to profitability, such as gross margins, which we see as more pliable. So what we are optimizing for is gross profit dollars. And in pursuit of maximizing gross profit dollars, there will be segments where we will let loss ratio rise because the elasticity of demand is such that, that will spike demand and retention in a way that offsets the margin becoming a little bit more constrained. So depending on which -- pick your metric, and I'll give you a better answer, the gross profit dollars, we expect to maximize, and we don't expect to take our foot off the pedal there at all. And the ultimate EBITDA breakeven is locked in for Q4 of next year, and we're not anticipating that changing.

Great. And then a couple of notes on the expense side. You're right that the tick up in this quarter was a little higher than is typical. We do see it vary quarter-to-quarter. I don't see that as a step change or an ongoing change. But particularly in the quarter, growth spend, obviously, is a notable year-on-year increase, and we break that out. We're spending a bit more for tech personnel, and you see the offsets from that in efficiencies elsewhere but that's a dynamic where if you isolate that line, over time, you can see some increase there year-on-year. Some of it is just purely inflation. The team size doesn't grow dramatically but the cost goes up modestly. In G&A, our interest expense growth, and that grows with our growth spend more or less because we're -- as you know, we're financing about 80% of that growth spend. So from an expense standpoint, it jumps out. But from an overall cash flow benefit standpoint, obviously, that's a terrific benefit to our -- the IRR measures of the company as a whole. A little bit of noise in our merchant fees, which can be seasonal, meaning they can move a little bit more or less than the premium in the quarter. So a number of little things, but the big picture is unchanged, single-digit expense growth and 30% plus top line growth, and you see that in the chart that we published, and that's what we expect going forward.

Your next question comes from Katie Sakys with Autonomous Research.

A couple from me. I guess, first, it sounds like there's a bit more growth scheduled for 4Q than previously messaged the last time you hosted a call. So I guess I'm just trying to reconcile the change in the IFP guide for the full year '25 given the magnitude of 3Q results relative to previous guidance. It doesn't sound like you're messaging necessarily a pull forward in growth into 3Q from 4Q, but it kind of does seem like the full year guide implies a bit of a sequential deceleration next quarter back down below the 30% growth rate. So I'm just looking for some additional color there on the change in the full year guide when 3Q IFP netted out relative to the previous guide.

Sure. Katie, your math is right. So when we have a big beat on a key metric in a quarter, then obviously, we evaluate how much of that we expect to continue forward and how much we want to make certain adjustments on the top line, that IFP number captures the entire business, not just the additional sales or the new sales or the growth rate. So while our growth spend has increased and our new sales, we expect to increase as well, we're cautious about retention. Our Q3 results were actually quite good, and we're able to overperform but we're somewhat thoughtful about that top line going into Q4 because that captures the entire business. The opposite is true on the other line items. So in gross earned premium and revenue, we captured not only the beat in Q3 but additional increase in Q4. So there's a little nuance there between the metrics, that's what's going on.

Okay. Yes. No, that makes total sense. It's just -- I mean, ADR, like to your credit, improved versus last quarter, showing upward progress there. I understand, obviously, some of that is coming from the lapping of nonrenewals on home from last year. But I mean, it looks like you guys are doing well in terms of retention versus maybe we were at the start of this year. So I'm just curious about the conservatism, like you were able to exceed the 30% IFP growth rate this quarter. So what in the financial plan is potentially looking a little bit less positive as we end up the year, especially as retention continues to improve?

I would think of it as all quite positive if you're looking for our view and how we see things rolling out, particularly in the fourth quarter where we're a month plus in. We have pretty good visibility. I'd remind that we continue to be really thoughtful about our home book of business. The underlying numbers actually look quite good, the loss ratio and the other metrics. But we continue to work through what we've called our clean the book exercise. That continues unchanged. Actually, it will have a level of impact in the second half that's similar to the first half. But that continues, and that's part of our plan. So we're growing at a 30% rate despite that sort of pruning of our customer base. So all your questions are fair, but I think we're quite optimistic I just want to be thoughtful about the parts we know about and the parts we don't yet know about, which is the remainder of the quarter.

And maybe, Katie, sorry, just for the benefit of people listening on who haven't done the math as you have, our guide does anticipate a 30% next quarter at the high end of the guide. We've guided at something between 29% and 30% growth for Q4. So we're certainly not anticipating or guiding to any considerable reversal or slowdown as guided.

Okay. And then if I could just sneak in one more. I can appreciate that the trailing 12-month gross loss ratio is trending well below the 73% target you guys have previously messaged. Just kind of thinking about that in the context of the changes to the quota share structure and ongoing maximization of gross profit dollars. Is 73% gross loss ratio still the right target for the business at this point? Or do you eventually see a pathway to taking that target down lower?

It's a great question, Katie. And to be honest, we've tried to be responsive to questions such as this one and provide a target loss ratio. But I do want to give you an insight into how we think about this, which is that there isn't a target per se. Loss ratio is an input, not an output. It's a lever which we use to optimize the business. It's not necessarily evident that the business is optimal. So because we are as efficient as we are and our other cost structures are declining as they are, we are in a position to be price leaders, a thesis that we developed at some length almost a year ago during our Analyst Day on November 20 of last year, which is to say, we think that there is a structural advantage that Lemonade enjoys where in a price-sensitive market like ours, but oftentimes a 1 percentage move on price will yield a fivefold increase in conversion or other metrics. It may make sense for us to continuously refine within certain markets and certain segments get to very competitive price points. And that will put pressure on gross loss ratio. But just to give you a hypothetical, I spoke earlier about the CAC-less acquisition of great customers in the car business. Why do we need to optimize to a 73% or any other particular number for these customers where there is no cost to acquire the customer and almost no cost to service the customers. You can envisage a situation where we could lower prices so dramatically where we would be profitable with a 90% loss ratio. The math here and the degrees of freedom that we have is pretty dramatic and something that will be very, very hard for the incumbency to replicate. So we are using the data to guide us in terms of what is optimizing gross profit. At times, that will mean selling a lot more with thinner margins, at times not. Some of our products are more price elastic, some are less, some campaigns are more elastic, some are less. So it will aggregate into a loss ratio that we will report on a quarterly basis. But we're thinking about loss ratio less and less as one big aggregate number with a target more and more as fine-tuning of optimization of by product, by campaign, by region, and that will result in different loss ratios, different product lines but always in the service of maximizing gross profit. I hope that gives you -- I hope that helps give you an insight into how we are approaching the question that you're asking.

We now have

So I also just wanted to follow up on the gross loss ratio, so down 5 points overall, up 13 points in Europe. So can you just talk a little bit about what drove that decrease in Europe? Is it benefits of scale? Was it product mix? And then just I'll get my follow-up on that topic as well. I think previously, you've talked about U.K. being really strong in Europe from a growth perspective, maybe Germany being a little bit weaker. Any updates there within the European market of what you're seeing?

Sure,

Yes. Just general good news across the board, I think, and particularly from a loss ratio perspective, we're starting to see some mix benefit. So as the U.K. grows and in particular, the renters book in the U.K., that sports a nice effectively low loss ratio, that starts to show up in the total. So that's part of the driver. Some of it is prior period impact, also favorable in the quarter, and that's good news. That just means when you have a younger book of business and you're more thoughtful in your reserving, you can at times have a favorable release of prior period. We saw a bit of that in our French book, which is a smaller book of business. With the U.K. heading in aggregate above the 50% level, that bodes well. But we're also seeing nice improvements in other territories as well. We've gone from having really no home business in Europe to having a really nice and effective home product now in 3 of the 4 territories. So Europe is really hitting on all cylinders. It's still a relatively small portion of the book but it's become material, and you'll likely hear more about it from us as we go forward.

Your next question comes from the line of Andrew Andersen with Jefferies.

Just looking at pet, it's been growing pretty well and the loss ratios seem pretty stable there. I was wondering if you could just touch on kind of the competitive environment you're seeing with pet, maybe how you feel your pricing is relative to some of the industry? And if you could maybe touch on what you're seeing in terms of loss trends there.

Yes. Again, I feel like a broken record. The things I said about the EU are also true in pet, super stable and predictable loss ratios at this point, a little bit of seasonality. The partnership with Chewy continues to hum along. about almost 5% of the business now has been driven through that partnership. From a competitive standpoint, we've done in 4 or 5 years. I think what it took pet-only providers that are really, really strong players in the market, 10 or 12 years to do. So we really like what we're seeing from a pet perspective. From a pricing perspective, I would think of it as similar to our other products where while we don't aim to be -- to underprice the product or to be price anything as a loss leader, we will often be, if not the most -- the least price, a super competitive price. So we do lose business if it doesn't satisfy our LTV model requirements but we're typically quite competitive with the strongest players.

And I just want to go back to some of the LAE comments and the potential for improvement in that ratio over time. I'm just trying to think about how you are managing kind of maintaining a similar customer service level but also taking into consideration, I imagine at some point over time, there will be a pivot back towards some more homeowners and auto will be a different or a higher mix of the book. So how do you kind of manage through the different customer service levels and the changing needs there, but also using automation efforts?

Andrew, you'll note in the letter, we break down the LAE by product. And you'll see that there's a uniformly down to the right shape to all of the curves, all of the products, including the more complex ones that you're asking about. So we are seeing that we're able to use AI across the board, across the product line to great effect and to achieve dramatic improvements in terms of automation. The very nice thing about using AI to do this work is that it's never at the cost of customer service. It is to the delight of customers. The overwhelming majority of complaints that we get, I think well over 90% for that things that humans do rather than AI does. So when we deploy AI to do these things, it's not the old thing that you used to get when you dialed United Airlines and you have to repeat yourself 7 times to be understood and press 1 and press 3 and press 5 and you knew you were interacting with a machine. These are very high level -- we only deploy the technology once it reaches very high levels of customer satisfaction. And once it does that, it usually exceeds or in the areas that we agree to let it go live, it exceeds what humans do because it's much faster. The error rate is often lower. So we're seeing it able to handle ever more complex things. Jason asked me earlier about kind of our scale of 1 to 10, and I think that would apply here as well, which is you can see how much we've done. And at the same time, we just think that there is a whole lot more that we can do. We really do see a blue ocean in front of us of areas that we can improve. So we're fairly bullish on kind of if you zoom out on the prospects of AGI within the next few years, which really means that machines will be able to do every activity, every intellectual activity that humans do today. And therefore, the idea that some of these products are more complex and require humans today is both true and transient. I think in the coming years, you will find that we'll be able to deploy systems to take care of all of our customers' needs, lowering our costs and raising the level of customer delight.

And I think add a thought, sorry to interrupt. I think there's a note or 2 in the letter that's kind of elegant around this concept of shifting variable cost to fixed cost. And so if you think from a customer satisfaction standpoint or a customer experience standpoint, very specifically, in the older world, even if you automated responses or interactions with customers, you had to have a human evaluating and improving those responses. So they weren't -- so they were constantly improving and getting better. And that human evaluating those responses became a variable expense. They had to review and think and make judgments even though they weren't actually responding to every request with the tools -- with the AI tools we now have at hand, even that review process with a human intervention can be automated such that an improvement in the response can filter out to our entire customer base in real time. And so this concept of constantly looking for variable expenses that we can convert to fixed expenses is really -- it sounds simple but I think in the world of AI, it really helps to kind of sharpen the focus on how these things actually turn into things you can see on the P&L.

[Operator Instructions] We have Jack Matten with BMO Capital Markets.

This is Charlie on for Jack. I'm sorry, we joined late, so apologies if you addressed this. But we saw Shai tweeted this morning that Lemonade plans to start lowering rates. Can you elaborate more on the timing and magnitude of when you may plan to file for these rate cuts and which lines of business are you talking about specifically?

Charlie, yes, we've addressed this both in my opening comments and in answer to previous questions. So I'll keep my comments brief. I didn't see Shai tweet what was alleged. So I think what Shai said is that -- or certainly what he meant -- the point that we're trying to get across is that there is a sense in which we -- and there was a question about this, we've achieved everything that we said we were going to achieve in terms of loss ratios and they're at record lows, and we're anticipating them potentially going even lower this coming quarter. And yet, we don't always see -- this is a moment to kind of take a victory lap and we're thrilled with it and it's excellent but we don't always see lower as better. That's what we were saying. And there are times when you can optimize gross profit with higher loss ratios as well. And it can be counterintuitive because you think lower means more profit but it also means taking a hit in terms of conversion and retention and therefore, growth. And the smart thing as far as we're concerned is to optimize not for a particular loss ratio number but to optimize for gross profit. It's what we do. And all that means is that different loss ratios for different products, different campaigns, different regions over time. There's nothing dramatic. We're not signaling any findings that are imminent or we're not guiding to a new target loss ratio or anything like that. We think the loss ratio, in fact, will continue to improve in the near term, just saying that it's important for our investors to be aligned with us about what metrics are important ultimately. And we think gross profit is the one that we're solving for and loss ratio is an input to it. I hope that clarifies that.

Yes. Sorry about that. And I guess for my second question, I know you've already adjusted your main quota share program to retain 80% of top line. Are there any other changes regarding your broader reinsurance program that you're thinking about heading into the new year given the expectation for reinsurance costs to continue to moderate?

Yes. I would say we're right on track with our typical approach to reinsurance, which is we're constantly thinking thoughtfully about what we might change or improve. But structurally, that renewal comes in July. We have the opportunity to add or subtract things during the course of the year, which we do almost never but we certainly have that opportunity. So we're constantly looking at those ways that we might help manage both the benefits of reinsurance from a volatility standpoint as well as managing capital surplus, and that really is the driver there. But we are in a great position with the renewal that came through July 1. We're heading towards a point by midyear next year where we'll be ceding just about 20% of our premiums and losses to our quota share partners. As you know, it takes a while to flow through the book once you get to a renewal as the business renews over the course of the year. The impact of that will be such that in Q4, our effective overall seed rate might look more like around 40%. So you're seeing, as expected, that decline as we move closer and closer to the next renewal. In the early part of the year next year, we'll start to get more serious with our partners as we have in the past and think through what that next renewal might look like.

If I could just sneak in one more. Any color on the competitive environment in pet? It feels like we've been hearing more public insurance carriers talking about that business more and more.

Nothing notable. I think we're still finding -- it's funny when you kind of look at the market from a competitive standpoint, you hear about either Google algorithms changing or competitors getting more aggressive and these things definitely happen from time to time. But if we look at our Q2 results, our Q3 results, our view into our guidance for Q4, it's really steady as she goes. Even frequency and severity of claims in the quarter was not notable, and that's good news for that book of business because it is -- continues to grow in terms of its share of our overall business. So while we kind of track the competitors, it's not top of the list of the things we think about. The things we are doing are working. They're working well. And pet as it has been for quite some time, is a key pillar that enables us to grow at 30% plus.

Our final question from the phone lines comes from [ Luke Nelson ] with Cantor Fitzgerald.

I just have a couple of brief questions this morning. My first question being with card, it's roughly around 15% of IFP today. Where do you guys kind of see that mix trending long term? So is 25% the right ceiling? And are there limits on auto exposure we should be thinking about?

So best indicator, I think, is to kind of think back a bit to our recent Investor Day, which is about a year ago now. So it's not quite so recent but we sketch out a plan and a vision to track and drive growth at the company from $1 billion of premium to $10 billion. And what we sketched out at that time was a CAR component of that of around 40%. I would think of that as sort of a thematic share but a pretty good one. It could be more, could be less. The TAM for car is in just the U.S., not to mention Europe, which we don't have a car product in yet. But in just the U.S., it's just an enormous potential market and even just our own customer base is an enormous market for us. So there's really no restriction from a TAM perspective. It's about us optimizing the LTV to CAC, really driving that cross-sell dynamic because that's what helps us with retention. The gross loss ratio improvement was terrific. So if you think about a mid-teens ratio today and a 40% CAR share at $10 billion, your number is not far off. 20%, low 20s is certainly within reason in the coming couple of years. We -- the nice thing about Lemonade is the mix of business is quite diverse. And so that number can ebb higher or lower, and we'll still be well able to track to our growth rate targets overall but I think CAR will end up in that range that you're thinking about.

Got you. That makes sense. And then just my last question is a 2-parter, and you might have touched on it previously, but I noticed retention increased to 83% but ceding commission increased as well despite the reduction in reinsurance. So can you kind of walk us through that dynamic? And where do you expect retention to trend over the next few quarters?

Sorry, if you could -- I think I misheard your question. Was it around ceding rate? Or was it around retention?

Right. So yes, my question was, I noticed retention increased to 83%, but at the same time, ceding commission income also increased. So can you just kind of walk through the dynamic between the 2? And where do you expect retention to trend over the next few quarters?

Yes. So a couple of metrics just to pull apart there. So we disclosed a retention metric, which is a customer metric. So ADR is annual dollar retention. And just as a reminder, that's the dollars from any given cohort of business, 1 year later, how much have you retained. And that number has tracked upward nicely from the 70s to the high 80s over many, many quarters consistently. It dialed back a couple of points over the past few quarters because of our home effort to clean the book, and we had some nonrenewals there that camped that number down. We've now seen that reverse as we expected. It went from 80 -- up 1 point this quarter sequentially. So it feels like we might be back on track to have that number increase. That's customer retention, stable and improving. From a ceding commission standpoint, that's a bit of a -- that's a different part of the business, and that's really related to the premium we share and the losses we share with our quota share partners. And so that, I'd kind of send you back to our earlier comments about the quota share renewal. So at July 1 this past year, we were ceding -- or June 30, we're ceding about 55% of our book of business. That has shifted such that it will move from 55% to about 20% over the 12 months from Q3 to Q2 that we're in right now. The commission we earn on that is a variable rate commission, and that's -- you'll see that pretty clearly outlined in our 10-Q disclosures that we'll file today, so you can kind of dig into the nuances there but we continue to get a mid-20% roughly ceding commission on all the premium that we see to that partner. And so we'll see fewer dollars. That's a good thing but we'll continue to earn a healthy commission rate on all those dollars that we see for our partners.