Good day. Welcome to the Lemonade and first quarter 2021 earnings conference call. All participants will be in listen-only mode. Should you need assistance, please signal a conference specialist by pressing the star key followed by zero. After today's presentation, there will be an opportunity to ask questions. To ask a question, you may press star then one on your touchtone phone. To withdraw your question, please press star then two. Please note today's event is being recorded. I would now like to turn the conference over to Yael Wissner-Levy, Vice President of Communications. Please go ahead.
Good morning, welcome to Lemonade's first quarter 2021 earnings call. My name is Yael Wissner-Levy, and I am the VP of Communications at Lemonade. Joining me today to discuss our results are Daniel Schreiber, CEO and Co-Founder; Shai Wininger, President, COO, and Co-Founder; and Tim Bixby, Lemonade's Chief Financial Officer. A letter to shareholders covering the company's first quarter 2021 financial results is available on our investor relations website, investor.lemonade.com. Before we begin, I would like to remind you that management's remarks on this call may contain forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995. Actual results may differ materially from those indicated by these forward-looking statements as a result of various important factors, including those discussed in the Risk Factors section of our Form 10-K filed with the SEC on March 8th, 2021, and our other filings with the SEC.
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, such as adjusted EBITDA and adjusted gross profit, which we believe may be important to investors to assess our operating performance. Reconciliations of these 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 key operating metrics, including 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, who will begin with a few opening remarks. Daniel?
Good morning. I'm happy to be able to report on another quarter of strong advances along our key performance indicators. As compared to the first quarter of 2020, our top line, which is in-force premium, grew 89% to $252 million, representing an accelerated rate of growth compared to the prior quarter. Premium per customer also increased at an accelerated rate, 25% year-on-year, as recent product launches continued to bolster our economics. Tim will elaborate on all our numbers shortly. During our last call, I spoke perhaps cryptically about a new product launch we are highly focused on internally, and we've since unveiled that this mystery product is Lemonade Car.
Perhaps that wasn't a huge surprise. I do still get asked, "Why car insurance?" Well, when asked why he robs banks, the notorious bank robber Willie Sutton answered, "Because that's where the money is." I can say much the same thing about car insurance. The car insurance market is about $300 billion in the U.S. alone, that's about 70x larger than the renters insurance market and 80x larger than the pet insurance market. It's also 3x larger than all of the homeowners market. Given that Google Trends shows that searches for Lemonade Car insurance and Lemonade auto insurance outnumber searches for Lemonade home insurance, we believe we have a fighting chance of taking a sizable bite out of this enormous pie.
Setting aside the massive new market that Lemonade Car opens up, it will also hopefully be a huge unlock of value for our existing businesses. For one, we believe that our homeowners insurance customers today already spend about $1 billion on car insurance, but they've been unable to spend it at Lemonade, and our forthcoming launch will solve for that. For another, we've been selling homeowners insurance effectively with one hand tied behind our back since we can't bundle homeowners and car insurance in the way our competitors do. Lemonade Car not only opens up a huge new market, but I do expect it to be a boost for our existing homeowners' business as well.
The next question I get asked is something like this: with such forbidding incumbents like GEICO and Progressive who have truly achieved mastery over the direct-to-consumer auto insurance space, how can Lemonade conceivably compete? Well, those companies are indeed formidable; they've been doing their thing since 1936 and 1937, respectively. Now, they each have tens of billions of dollars of in-force premium, they spend billions of dollars a year on advertising and have done so for many, many years. That all adds up to real heft. We have tremendous respect for these competitors, as well we should. Strengths and weaknesses are two sides of the same coin; all that legacy and bulk comes at the expense of nimbleness. That may be a problem for them since the car industry is going through a once-in-a-century dislocation, that may favor the legacy- free.
As a rule, when innovations are continuous or incremental, the benefits of these innovations accrue to incumbents. When they are discontinuous or disruptive, they typically accrue to the benefit of disruptors or newcomers. I think that the transformations in the mobility space are very much of the latter kind. Cars are moving from being mechanical platforms to being digital platforms, morphing from being dumb appliances into smart robots, and from being isolated devices to being nodes on a network. Tesla is clearly showing the way. While the majority of cars will take some years to be as fully connected as a Tesla, their drivers already are.
The smartphone every driver brings to their excursion has exquisitely sensitive sensors, allowing us to derive gravitational, magnetic, location, and directional measurements that we can map onto driving metrics like how much a person drives, how aggressively, and whether on accident-prone roads or on relatively safe ones. Finally, and unlike data from connected cars, smartphone-based sensors also allow us to detect distracted driving, a highly predictive risk factor, and to track drivers across different cars they drive, rather than homing in just on a single car, regardless of the driver who is driving it. The upshot is that the data streams from cars and from their drivers allow us to graduate from pricing based on make and model, as has been done for generations, to pricing based on usage and behaviors. This could be transformative for the car insurance industry.
I like to think of the kind of precision underwriting that technology is enabling as being akin to the revolution unleashed by the invention of the microscope. Before microscopes existed, everybody thought that a drop of blood was just a monolithic blob, whereas after we had microscopes, we could see red blood cells and white blood cells, and the fact that they are very unevenly distributed and of different sizes and perform different behaviors. I think the same could be true with these connected streams. Instead of pricing a large group of people as though they are monolithic, connected devices and connected drivers allow us to do precision underwriting. This could really be a game- changer. It's not that these technologies are unavailable to companies like GEICO.
It's that they might threaten their sizable book of business and undermine the competitive advantage, the old way of doing things that they built up over these many years. That may be why GEICO resisted telematics for a very long time and only reluctantly dipped their toe in the water not that long ago. Warren Buffett addressed this in the recent Berkshire Hathaway annual general meeting, and he said the following: "GEICO clearly missed the bus and were late in terms of appreciating the value of telematics." He added the following, "Hopefully, they will see the light of day before not too long." Why do many incumbents adopt these technologies half-heartedly? Often, when they do adopt them, they will underweight their signals in their rates.
I think it's because these new technologies have this ability to break up groups that they have been treating so far as monolithic and pricing them to their average. These new technologies will reveal that about half of those groups are actually overpaying. They are better risks than average. The adoption of these technologies will lead to lowering their rates, which will mean losing premiums, and they will also reveal that the other half of these groups are underpaying, that they are worse risks than the average. That will require raising their rates, which in turn will mean losing customers and again, losing premium. You can see why the always- connected car, the always- connected driver, amounts to a reset of how car insurance can be structured, underwritten, and priced.
This is advantageous to players without a legacy business to protect and who design their business from the get-go for these emerging realities. In a minute, I'll hand over to Shai. Just before I do, I'd like to switch gears and address the Texas freeze, also known as Winter Storm Uri. This was the fierce winter storm that hit Texas and neighboring states in February and impacted millions, causing power outages, icy roads, frozen pipes, and, sadly, a great deal of suffering. We received about an entire year's worth of claims in the first few days, providing an extreme stress test for both our operations and financials. The results, we believe, should be very reassuring to our team, our customers, and our investors. I'll start with the operations stress test. At the onset of Uri, our claims experience team activated our catastrophe or CAT operational process.
Our people and technology rose to the occasion. A majority of claims were fully resolved within one week of the storm's onset. As always, we put our customers first and are proud to have delivered best-in-class, delightful experience to them in their serious time of need. Net promoter scores for our claims interactions associated with the crisis were nearly 70, in line with our typical non-CAT experience and at a level that I believe is without parallel in our industry.
Turning to the financial stress test, all those claims from Uri and CAT in general in the quarter added about 50 percentage points to our gross loss ratio. Our EBITDA guidance for the year remains materially in line with analyst consensus prior to the storm. The explanation is pretty simple. We have extensive reinsurance programs in place for just such eventualities. They worked very much as promised.
All told, Winter Storm Uri was by far the most severe catastrophe Lemonade has had to deal with; it shows in the sudden spike of our gross loss ratio. That's pretty much the only major place in which it shows. You might have expected that a year's worth of claims packed into a single week would also crash our systems or overwhelm our teams, or lead to a degradation in customer satisfaction, or at least lead us to restate our EBITDA guidance. It has not. That is a strong testament to the financial and operational underpinnings of Lemonade and to the resilience of our tech, people, and partnerships. With that, let me hand over to Shai for some more updates. Shai, over to you. It shows in the sudden spike in our gross loss ratio. That is pretty much the only place it shows.
You might have expected that a year's worth of claims packed into a week would also crash our systems or overwhelm our teams, or lead, at the minimum, to a degradation in customer satisfaction, and probably make us restate our EBITDA guidance. It did none of those things. That is a strong testament to the financial and operational underpinnings of Lemonade and to the resilience of our tech people and partnerships. With that, let me hand over to Shai for some more updates. Shai, over to you.
Thank you, Daniel. Let me start with Lemonade Life. Last time we spoke, I mentioned that, unlike previous products, our term life insurance business will be launching gradually. We spent the first complete quarter post-launch ironing out the kinks by testing and improving the product to ensure it delivers on the Lemonade promise and provides a fantastic experience for our customers. We feel that we've made significant progress on this front and have moved to focus on growing the term life business.
On other fronts, we're happy with the rate in which our non-renters products are growing, with Homeowners, Pet, and Life representing roughly half of our new business in the last quarter, up from roughly a third a year ago. This diversification is strategically important to us. Our systems have become increasingly sophisticated at optimizing budget allocation for each product to ensure maximum ROI.
When we see a decline in profitability or volume for product A, we divert dollars to product B to sustain efficient growth. We're separately seeing attractive cross-selling trends across our business. This is one of the reasons we're excited to share a new metric with you today, annual dollar retention or ADR. Our ADR has improved considerably in recent quarters, up to 81% in Q1 2021 from 70% a year ago. Existing Lemonade customers who are purchasing additional products are a major driver of this improvement, contributing seven percentage points in the current period. As we look ahead to our Lemonade Car launch, we see potential for a meaningful acceleration in ADR as well. Lastly, I definitely share Daniel's excitement about Lemonade Car and wanted to share a sneak peek into how we are currently thinking about the product.
You can expect everything you already love about Lemonade now for your car. A delightful conversation with our chatbot, Maya, will help you get a customized, hassle-free policy in minutes. It will be a beautifully designed experience that is easy and fast. Claims will be paid quickly, and we'll continue to support charities on behalf of our customers. We'll also have attractive pricing for environment-friendly cars and EVs. This is in line with our company core values. As a B Corp and public benefit corporation, we have a commitment to the public good, and so our products will, of course, encompass that as well. With that, let me hand over to Tim for a bit more detail around our financial results and outlook. Tim?
Great. Thanks, Shai. I'll give a bit more color on our Q1 results, as well as expectations for the second quarter and the full year of 2021, and then we'll take your questions. We had another strong quarter of growth driven by additions of new customers, as well as a continued increase in premium per customer. In-force premium grew 89% in Q1 as compared to Q1 the prior year to $251.7 million. We believe that this metric captures the full scope of our top-line growth before the impact of reinsurance and regardless of the timing of customer acquisition during the quarter. Premium per customer increased to 25% versus the prior year to $229. This increase was driven by a combination of increased value of policies over time, as well as mix shift toward higher value homeowner and pet policies.
Again, as in prior quarters, roughly two-thirds of the growth in premium per customer in Q1 was driven by this product mix shift, including cross-sales, and the remaining one-third from increased coverage levels and pricing. Gross earned premium in Q1 increased 84% as compared to the prior year to $56.2 million, in line with the increase in in-force premium. Our gross loss ratio was 121% for Q1. This includes 50 percentage points associated with Winter Storm Uri and other CATs. In Q1 2020, our 72% gross loss ratio included two percentage points of this CAT impact. If we exclude the CAT impact from both of these periods, our Q1 2021 loss ratio was right in line with the prior year.
Operating expenses, excluding loss and loss adjustment expense, increased to 25% in Q1 as compared to the prior year, with sales and marketing expense up 52%, well less than the pace of our premium growth. The prior year G&A expense line, as a reminder, did have a one-time non-cash expense of $12.2 million related to the creation of the Lemonade Foundation. We also continued to add new Lemonade team members in all areas of the company in support of customer and premium growth in both current and future product launches, and thus saw increases in each of the other expense lines. Our global head count grew just over 100% versus the prior year to 661, with a greater growth rate in customer- facing departments and product development teams.
Our net loss was $49 million in Q1 as compared to the $36.5 million we reported in the first quarter of 2020. With a notably larger customer and in-force premium base. Adjusted EBITDA loss was $41.3 million in Q1 as compared to $22.4 million in the first quarter of 2020. Our total cash equivalents and investments ended the quarter at roughly $1.2 billion, reflecting primarily the net proceeds from our January follow-on offering of approximately $640 million, partially offset by the use of cash for operations of $40.3 million since year-end 2020. With these goals and metrics in mind, I'll outline our specific financial expectations for the second quarter and an updated full year of 2021. For the second quarter, we expect in-force premium at June 30 between $283 million and $288 million. Gross earned premium between $63.5 million and $65 million. Revenue between $26 million and $27 million.
Adjusted EBITDA loss of between $43 million and $40 million. Stock-based compensation expense of approximately $13 million, and capital expenditures of approximately $3 million. For the full year of 2021, we expect in-force premium at December 31 of between $376 million and $382 million. Gross earned premium between $279 million and $283 million. Revenue between $117 million and $120 million. An adjusted EBITDA loss between $173 million and $163 million. Stock-based compensation expense of approximately $50 million, and capital expenditures of approximately $11 million. As a reminder, please note that GAAP accounting rules are such that ceded premiums are excluded from GAAP revenue. As a result of the change in our reinsurance structure, effective last July 1st, to a significant proportional reinsurance structure, our year-over-year revenue and gross margin comparisons are not directly comparable in this period.
Accordingly, we publish in-force premium and gross earned premium as metrics that we believe are useful to analysts and investors because both capture the overall growth trajectory of the business before this impact of reinsurance. With that, I'd like to turn the call back over to Daniel, who will address some questions from our shareholders. Daniel?
Thanks, Tim. As we did last quarter, we invited our shareholders, regardless of the size of their holding, to post questions or to upvote questions so we can be sure to address the issues that are deemed most pressing by our community. Looking at the most upvoted questions this quarter, we see a few central themes. One was captured by Kayun about cars and specifically tie-ins with OEMs and particularly manufacturers of autonomous cars. Another from Neil is about growth and how we're planning to outpace our competitors. Our third most upvoted question was from Sanjay with an addendum from Arias about how AI is advancing and whether we'd consider offering it as a service to others. Finally, I'd like to address a question from Dean about blockchain and decentralized insurance and whether that poses a long-term threat to Lemonade. Let me address these in turn.
Regarding partnerships with car OEMs, I can definitely say that we see the power of that, and I expect it's a trend that we will see more of as OEMs like Tesla and others dip their toe into the world of insurance. Cars, as I said in my opening comments, are morphing into connected platforms, and that has tremendous implications for risk selection, risk pricing, and risk mitigation. At the same time, I think that much of the benefit of the connected car to insurers is already available today through connected drivers. The mobile phone not only does a great job of tracking risk factors, it itself is a risk factor, and a risk factor that cannot be tracked in any other way. Reportedly, one out of every four car accidents in the United States is caused by texting and driving.
By some estimates, driving while distracted by a mobile phone is six times more likely to cause an accident than even driving while drunk. While we're looking forward to leveraging all the tech that's embedded into next-generation cars, and perhaps to partnering with OEMs in the fullness of time, we don't think we need to wait for these connected cars, which remain a tiny fraction of the cars on the road, to become prevalent in order to drag car insurance into the digital age.
Already today, every car, even that 1973 hand-me-down from grandma, is in effect equipped with a slew of precision sensors capable of generating predictive insights and streaming them in real time to an appropriately trained AI. All that technology is packed into the driver's pocket, and we plan to leverage that power from day one without waiting for connected cars to become mainstream.
To Neil's question about growth, as Tim just recapped, we're actually seeing not only strong growth but accelerated growth. Given how young our company is and how extraordinarily large our market is, and just made that much larger by entering into the car insurance space, it's my expectation that we will see many more years of very rapid growth, something which our competitors really don't these days.
So long as our unit economics continue to be healthy and strong, we'll continue to lean into growth investments. That's what happened in Q1. It yielded, as I say, accelerating growth in in-force premium and in net customer adds. Also, as we continue to launch new products, we've been reporting accelerated growth in our premium per customer, too. Here, too, I think that we have many years of significant growth ahead of us. Hope that addresses your question.
Sanjay asked about our AI. Given the nature of our closed- loop feedback system, our AI continually gets stronger with each customer interaction. As we expand our product offering and broaden the scope of our relationships with our customers, our data becomes richer and we train our AI engine to increasingly complex problem-solving and improving capabilities. There are no concrete plans for an AI day to your question. Let me share some of the specific examples that may help illustrate the far-reaching impacts of AI technology on our business. One is something that we call Watchtower or the Eye in the Sky. We use machine learning to analyze signals coming from orbiting satellites in space to detect catastrophic events around the globe in real time.
This technology allows us to enter into our CAT response protocols faster than probably any other insurance company and to deliver a best-in-class customer experience, but it also allows us to reduce the prevalence of bad risk in our book by blocking campaigns and sometimes customers and customer acquisition in impacted regions much quicker. Another example that comes to mind is something that we call The Sixth Sense. For each customer or prospective customer, we're able to collect data that allows us to produce something like a digital footprint or stuff like device ID and IP address and face detection, and all different other things that I won't elaborate on. When a customer whose claim has been denied or is suspected as fraudulent will create a new account and submit a new video task or otherwise make claims that are dishonest.
We are pretty good at instantaneously detecting those and shutting them down, something that incumbents using email or call centers would really not be able to do at all. You just can't connect the dots in the more traditional ways of working. The upshot of that is that we can catch fraud networks that previously, to the best of our knowledge, simply went undetected and those claims would have been paid. We also use satellite imagery and computer vision to extract information about properties we insure, whether it's roof types or whether there's a swimming pool or other important signals. We do that during the onboarding of our customers. This allows us to deliver what I still believe to be the fastest and most delightful, and most straightforward onboarding experience while increasing the precision of our pricing and underwriting.
We see this in other parts of our business as well. Our customer support and claims management engines are very substantially driven by machine learning. We are able to use technology to automate a significant portion of our customer interactions. More recently, with certain subgroupings of interactions, for example, in p et insurance, which we launched just a few months ago, for Pets' preventative claims, we developed technology that effectively removes the need for humans entirely. Not only does that enable us to pay out a meaningful percentage of our claims instantly at a lower cost, it also delivers a great customer experience. Excuse me.
It does that while allowing us really across the board to achieve very levels of operational efficiency and scalability, and depending on the incumbent, we've achieved a ratio about 5x , perhaps more than that, the number of customers served by every Lemonade employee relative to what the industry knows. Five, sometimes 10x greater level of efficiency in terms of at least headcount. It is pretty impactful. For example, just last month, we gave back to our customer support team something like 3,200 hours because our AI technology solved and executed on a growing share of incoming customer requests than it had previously. All different things, whether a customer wants to change their address or resend documents or asking about their payment history or adding a spouse to their policy, all different things like that.
We really do see this impacting our business across the board. With some of the top-notch talent in the world in the field of machine learning and AI at Lemonade, the development of these tools is now part of our DNA, really has been from the get-go. Tools like this clearly don't just impact one area of a business. Rather, the impact is very broad and touches every aspect of Lemonade, from customer onboarding to claims management, really to the fundamentals of our business model. Data and machine learning, and automation, because it's built into the fabric of the company, that's why to Arias's question, we don't see ourselves renting it out or licensing it to others. Our company is vertically integrated, and the AI is deeply integrated.
This isn't some external bolt-on that we could license out or that we would want to license out. It's the very essence of our company, and we expect it to remain that way. Finally, I want to address Dean's question about blockchain and whether it poses a threat to Lemonade. Let me say that both Shai and I, and others on our team, have taken an interest in blockchain. We actually are both moderately invested there too, with a bent towards dApps and particularly DeFi, decentralized finance. These are powerful technologies and paradigms. I do think there are some applications for these technologies in certain areas of insurance. For example, Nexus Mutual is doing some very interesting work on ensuring smart contracts.
I do think decentralized finance and blockchains like Ethereum present opportunities for companies like Lemonade to build new products that ensure new realities and new classes of assets. I don't, for the moment, see how they can provide an alternative or become a threat to today. Homeowners, Life, Pet, Car insurance, these cannot be profoundly bettered by crypto, and here's why. Decentralized finance, DeFi as it's known, its power comes from the trustless nature of the blockchain. When determining whether a major global event has occurred, oracles, as they are known, provide a decent way to pump real-world events into smart contracts in a trustless way.
If I want to place a bet on a blockchain prediction market on the outcome of an election, for example, there are platforms and protocols that will allow a smart contract to know if the event occurred, and it does that in a trustless way, and then it can settle the bet accordingly. The same would apply to insuring against certain highly visible outcomes, say, whether some big concert took place on a given day or whether a landmark building was standing on a given day. If the question at hand is whether my laptop was stolen at my local coffee shop, there really is no API for that information, and asking the community to stake coin and vote would be terribly inefficient, and I think probably produce very unfair outcomes anyway.
Personal lines insurance is fundamentally about trust, and that's why at Lemonade, we've worked so hard to build a trusting and trustworthy company. It's really built into the fabric of our company, into the business model, into the game theory and behavioral economics upon which Lemonade is predicated. Trustless personal lines insurance, at least using the current state of the art of blockchain, would not be an improvement on what we've already built. While the world of DeFi opens up new opportunities, new products, new markets, I don't see yet how it can be leveraged to do a better job at what Lemonade does today. I will say, though, that if future developments change that assessment, you can be sure that we will be at the cutting edge of developing and deploying them.
With that, let me turn the call back over to the operator so we can now take some questions from our friends on the streets. Thank you.
Thank you. We will now begin the question and answer session. Today's first question comes from Matt Carletti with JMP. Please go ahead.
Hey, thanks. Good morning. Daniel, I wanted to ask the first question on auto. You've talked quite a bit in the past about the flywheel at Lemonade and kind of how your computers get smarter with time, kind of the more iterate to go through. Elsewhere in auto, particularly telematics, we've seen other companies kind of go through that process and take three, four years to kind of hit their stride and get their things zeroed in. Can you help us square those two and how for Lemonade, kind of you'll hit the ground running with your auto product? Will you do the telematics in-house, or might we expect to see you take advantage of a outsourced company like CMT or TrueMotion that is far down that path already?
Hey, Matt. Good morning. Thanks for that. There's a number of factors that I think will interplay here, and we're not giving too much information about exactly what we're doing in car until we're ready to launch, but I'll try and give you enough to give you a sense of how we're thinking about these things. In other words, good customers in homeowners products tend to be good customers in car insurance as well. Responsible human beings tend to express their responsible nature across different things when at home and when out of home as well.
With 1 million customers that we'll be targeting who are on our books, and car insurance will at least have a fair bit of data before we even promote to whom to promote our new products. That's one thing that I'd say. The second is that some aspects of the rating of telematics is a matter of public record. We don't have to start entirely from a standstill. The way the insurance space works in the U.S. is that we can take a look at how different behaviors, different telematics signals have been rated by competitors. Much as we do with all of our products, when we start with and we don't have our own data, we do tend to use the best available data out there as a starting point, and then we refine it over time.
Rather than just guessing, we're actually in a pretty strong position to get the best practices that are out there today as a starting point and then build on them. Frankly, whether the signals are collected through one technology or another, and you mentioned a couple of vendors, is a secondary question. In other words, there's two things. Which technology is collecting the sensory data and kind of sending it home, and then secondly, how are those rated as an insurance rating factor? As I say, regardless of how the technology is developed, the rating factors that are being applied by competitors today is a matter of public record and will serve as our starting point.
Great. That's very helpful. Thank you. Then one other, if I could just on Uri and the Texas, the NPS score of 70 on the claims, which is pretty impressive. Really, the question is what do you plan to do with that? That's a real asset. Should we expect to see some sort of marketing campaign in Texas or elsewhere based upon that? Have you seen any just kind of word of mouth, new customer trends following the storm?
We have had, and we continuously have, satisfied customers spreading word of mouth. They tend to tweet about their experiences. If you look at the Lemonade Twitter handle, you'll see quite a lot of people talking about their good experience. We do see referrals from customers on a pretty regular basis. It's not just about yearly, it is about the continuous day in and day out obsession with high levels of customer satisfaction throughout the life cycle. I don't know that there'll be a yearly specific event. I'm not sure how tasteful that would be. The idea that treating customers well and delighting them and getting to high NPS does yield long-term value to the company is a formula that we're very committed to.
Yeah. Fair enough. Thank you very much for the answers.
Thank you.
Our next question today comes from Michael Phillips with Morgan Stanley. Please go ahead.
Thanks, Ed. Good morning, everybody, and congrats on the news with the auto. Looking forward to see how that develops. One question there, too, from me on the auto side, I guess, can you talk about, you mentioned no impact in guidance for the year on in-force premium since you're not even rolling it out yet, but I guess you're doing stuff there. Maybe talk on impact on both sides of the underwriting margin, both the expense side, I assume you're hiring people, and maybe doing some filings and things like that. I would imagine there's going to be maybe a little bit of pressure on the expense side.
Maybe a little bit longer term once you are live, thoughts on the loss ratio side, even if you are good and been around for a while, new business does have a bit of a higher loss ratio. Any pressure you may expect there on the loss ratio side on that? Those two sides, please.
Yeah. Hey, Mike. It's Tim. A couple thoughts there. As we have with prior launches, we have taken into account probably the vast majority of the expected expenses and are quite conservative on the top-line impact. We've taken that approach also with Car. Car's a little more amplified. It's a bigger product. There's a larger team focused on it. In fact, the largest team in the company from a development perspective is focused on Car, as we noted. It's a longer lead time. We expect to launch it, but it's not in the next couple of months. For this full year, I would expect we've got the vast majority of our expenses. We anticipate some of these unknowns, and we have done that in the past.
As we get closer to a more concrete launch date, obviously, we'll update how we see the top line developing from the car launch. This is consistent with how we've done the prior launches with Life, with Pet, and our shift to homeowners. From a loss ratio perspective, we kind of face this already today. We've talked about what we think of as a new business penalty, which is when you launch new products, you enter new markets, you bring in a new type of customer, things are a little bit tougher. That's in the model today. Daniel noted that something like half of our business at this point coming in is something other than renters. New things are a significant part of the book today. Homeowners, Pet, Life now starting up. This is something we've dealt with previously.
We do realize that a car is a different animal. It's a bigger market. There's more challenges. Again, I think I'd point you back to our initial comments that some of these uncertainties, some of these disruptions, we think can favor the more agile company. This is something we've, I think, performed quite well with in three consecutive product launches just over the past year or so. We're gearing up, and we're excited for the launch when it comes.
Okay. Tim, does that mean, obviously, you don't have any policies yet, and you haven't even rolled it out or launched it, and so no impact on the guidance in-force premium? Does that mean the guidance for the year on EBITDA contemplates any incremental expenses from the rollout?
It contemplates all the expenses we expect to hit this year.
Yeah.
Yeah. Yes.
Okay. No, cool. Thanks. Second question, you mentioned a couple times, I just want maybe a little more details on new entry points for customers and kind of new channels that you're seeing or using. Can you talk a little bit more about that?
Yeah, I think it's a theme, right? It's not a sudden new change that we're seeing, but part of the reason we're able to increase our marketing efficiency, in fact, this quarter, I think we brought in something like 40% more dollars for every dollar invested. That's after quite significant improvement last year. We're continuing to see improvement in marketing efficiency, which, at the scale right now, is a great thing to see continuing. In terms of the channels, we're always looking and testing new channels. The real indicator, I think, is the proportion of the customers that come through the largest channels, and that has declined, which is good news. A couple years ago, something like 90% of our business was coming through the largest channels. That has improved now to 80% last year and something like 70% now.
Large channels by definition, there's a bit of an 80/20 rule, but we're continuing to see diversification. We're testing new channels. Again, I wouldn't hold your breath for Super Bowl ads, but we are active in all the places where we think our customers are spending time. Whether that's online, whether that's social media, whether that's Google, Facebook, YouTube, and sort of the usual suspects, but also complemented by newer channels that we're testing. I would expect that concentration to continue to diversify over the coming couple years.
Okay. Thank you, Tim. Appreciate it.
Our next question today comes from Jason Helfstein with Oppenheimer. Please go ahead.
Thanks. First, just want to say our thoughts are with the team in Israel, with everything that's going on. Two questions. One, Tim, you just acknowledged in the last question: you know you are leaning more into marketing. Is there a way to kind of quantify, like you're leaning more into marketing, but you're also seeing higher lifetime value? How you would compare that versus a year ago at the IPO? Do you want us to kind of use that retention metric that you're announcing as a proxy? Just some more color on, again, how those marketing investments are actually, how that connects to higher LTV. Secondly, is there a way to think about what the adjusted gross profit would've been without the 50 points of CAT impact? Just the actual dollar adjustment. Did you have to take any of that on your gross profit line? Thanks.
Sure. Start off with the marketing comment or question. Leaning in is really two things. It's spending more dollars in absolute terms, but also seeking continued efficiency gains. Now, the counterforce to that is we're doing new things. We're launching new products. Pet continues to grow really nicely. Life is very new and nascent. While we're always focused on acquiring profitable customers, we'll be a little bit more aggressive on developing our newer products and our more mature products. Renters, for example, I would think of as the most mature. The returns continue to be very strong, and so while we'll not proactively go after unprofitable business, those are the two balances. I would maybe think of historically there's been a ratio of our advertising spend, which is our direct growth spend as a percent of total sales and marketing.
That's a number we've disclosed, and it's in the Q. That ratio's been kind of 70%-75%. That's continued in that range. That tells you as the absolute dollars increase, we're able to maintain that ratio. By leaning in, that's really what we're looking at. From a gross profit standpoint, rough order of magnitude, this is part of the reason we're able to, I think, have a good result with the Texas freeze. Part of it is the benefit of timing, right? We had a little bit of a preview on our last quarterly call because we didn't have all the information. We had a good amount of information to help guide us. The overall losses were in the roughly mid-$20 million range. The impact on us, because of our proportional reinsurance, is roughly 25% of that.
I would think of that, and it is a bit embedded in the gross profit number, but I would think of 25% of that total number as the rough impact. Absent that, you would've seen a nice year-on-year growth in gross profit as we've seen in prior quarters.
Thank you.
Our next question today comes from Ross Sandler at Barclays. Please go ahead.
Hey, Daniel, one technical question and then one for Tim. Daniel, how might the iOS 14.5 update with IDFA impact the company's ability to collect data from the phone? I know that's been part of the strategy in the past. How important is phone data versus the information that customers provide you in the onboarding flow? Tim, thanks for the new metrics around retention. I think Shai said that seven points of the recent 11- point increase was from bundling products. Is that mostly the story, or is there something else that's driving up that retention rate? I think the definition's a little different than what was in the S-1. Could you just bridge the difference between the recent disclosed retention and the S-1 retention? Thanks a lot.
Sure. Hey Ross, good to talk to you. Thanks for your questions. The Apple changes don't materially affect us, as best we can tell. For one, we tend to be mobile- heavy, sorry, web- heavy in terms of our onboarding. A lot of our transactions happen on websites. We don't do anything other than when customers are logged in. Indeed, it's a good point and time to reiterate, we don't sell the data, we don't use the data, we don't advertise, we don't do anything with the data other than use the data just for the purposes of the app itself. We are in pretty safe waters, and we don't see anything meaningfully changing in how we operate or our ability to collect or use data changing really at all as a result of 14.5.
Ross, in terms of the retention metrics, just as a refresh, since a year ago, since the S-1, we've talked about our customer retention, unit retention, the number of customers that we keep over time. That has continued to be stable. Obviously, much larger absolute numbers, much larger market, much greater penetration into the U.S. and that retention has continued to be stable. Remember, at these growth rates, 80%+ annual growth rates, the bulk of our business are first year customers. Sustaining and maintaining and seeing some improvement in those retention rates is a real positive that we see. The new metric, annual dollar retention, is okay, in those cohorts that you're retaining, what are they spending? They consistently spend more over time. A couple of things are happening. One is we now have multiple products to sell them.
Yes, cross-selling is a key component. Today we've got customers who can have two policies, in fact, three policies with the addition of Lemonade Life and hopefully we'll add to that with Lemonade Car before too long. We're also seeing increased coverage amounts though. People get a little more knowledgeable. We have outreach programs. They become a little wealthier. All that can drive a greater level of coverage and greater average premium. It's really a combination of those. The S-1 metric is very specific to a 12-month period. The dollar retention now really captures all the different ways that we can generate more value from customers as we retain them.
Thank you. Our next question today comes from Ralph Schackart with William Blair. Please go ahead.
Good morning. Thanks for taking the question. Two, if I could. First, go back to the first question. Just curious, how much of the auto policy pricing will be initially based on third-party data sources? Maybe more importantly, maybe speak to your ability and opportunity to leverage your own connected auto and driver data over time, and how you think that's going to be an advantage or a significant advantage potentially compared to incumbents. Second, your claims process historically has been known to be very hassle-free, frictionless. I'm just curious with auto, will it be a similar experience or process that will require a simple picture or video to be submitted to the bot? Just curious how that submission process looks versus your current products. Thank you.
Sure, Schackart. Shai hinted at some of this in his earlier comments. I'll take your second question first, which is that a lot of the experiences that our customers have come to associate with Lemonade will absolutely be carrying over into the car product. Again, without being too specific about how that will work, the idea of getting a hassle-free experience, instant claims, a great deal of automation and seamlessness is something that we're very committed to, and I think everybody who's been waiting for this product will have felt that the wait was worth it. This is going to be a really exciting and I think compelling product, which will be differentiated in terms of the experience quite significantly from what's out there today.
I think that'll be pretty exciting, and yes, the kinds of things that you're asking me about are the kinds of things that you can expect to find seamlessly integrated into the product. In terms of the data sets, it's worth distinguishing between data sets and rating rules. You, Ralph, today don't have the data that a GEICO or Progressive or State Farm has, but nevertheless, you can go in and see how they price based on the rates and forms that they have filed with regulators on a state-by-state basis. The lookup table that says make and model is rated this much, and driving experience is rated this much, and credit worthiness is rated this much, et cetera. All these factors are available even if you don't have that data to hand; you can know how the data is deployed.
My earlier comment was really saying that the cold start problem of where we've got a great stream of data coming in because from day one we will have very rich data coming from our customers, but you haven't had the cycles yet to know how to map those data onto risk factors and therefore onto rates. That problem gets solved by using the existing rates that have been approved by regulators for others, and you can see how different risk factors have been rated, and that provides a very strong starting point. What we've done with other products and I expect we will do with car as well is start there and then have our feedback loop make us smarter and hopefully we achieve parity pretty quickly and then we end up with an advantage play.
That is how we start and how we solve for the problem. It's not so much a lack of data or using anybody else's data; it's about using rating factors and weights that regulators have approved and mapping them onto data that we ourselves will generate. I'm simplifying, but I think in broad strokes that gives you a sense of how we approach this.
Great. That's helpful. Thanks, Daniel.
Our next question today comes from Tracy Benguigui with Barclays. Please go ahead.
Thank you. Good morning. You mentioned that as you grow in auto, that would also have a growth impact in homeowner. I'm just wondering if you could provide some context to us on how we should think about a normal CAT load in the future, as homeowners would become a more meaningful part of your business mix?
I would think of a couple things. I think there's maybe two questions there. I think there's just perhaps optimistically pent-up demand perhaps in our customer base for those customers who may prefer to have multiple policies or are bundling benefits, whether it's for discount or cost or ease of use or whatever. We've found that more policy types brings us better results, higher value per customer, greater retention, greater customer satisfaction. We don't see cars any different. In fact, car can be more of a driver for many customers.
From a CAT perspective, our transition to a less concentrated book of business in the renter's product and more in homeowners can have an increased impact on the CAT load, but greater diversification across the country, which has been significant. We started out in three or four states. We're now in close to 50 states, depending on the product.
We're starting to see the opposite effect, where we're more diversified across the country. Our largest states used to be a larger proportion of the business. Now that's declining. There's kind of opposing forces, I think, that will move the loss ratio impact around. It is traditional that car providers tend to have a somewhat higher loss ratio. That fact is not lost on us, and it's something we'll think about as we get closer to things like the discounts and pricing. I think it's a little premature for us to say much about that.
Okay. Looking forward to hearing more about that in the future. I don't have any further questions.
Thank you.
Our next question comes from Arvind Ramnani with Piper Sandler. Please go ahead.
Hi. Congrats on a good quarter. I also wanted to ask about auto insurance. Given that this auto market is very large, how are you thinking about cadence of marketing and signing up customers? Are you initially going to offer this just to your existing customers, and then you'll open it up to new customers? I am just trying to get an understanding of your go-to-market approach.
Yeah, it's a little early. We're not quite to the point where we're ready to give real specifics about the launch. We'll do that as we get closer. I would look to our recent launches of other products as a guide. We've found there to be a pretty significant pent-up demand for all of our products. In fact, you'll see as many searches for Lemonade Car insurance as you might for products that we actually currently sell. That's a good indication. We've opened up the funnel to collect email addresses and interest from potential customers as of a few weeks ago with the announcement. That's another way that we kind of get a good feel for the market acceptance. Our existing customer base is absolutely the most interesting place to start. 1 million+ customers, 1.1 million customers.
What we found with the Pet launch, and we're beginning to see sort of the green shoots in the life launches. There are folks for whom the Lemonade product is a decision point. When we launched Pet, there were dozens of different ways and vendors where you could get pet insurance, yet we had scores and scores of customers on the first day buying pet insurance from us. That tells you that there's something about the Lemonade brand, the Lemonade experience that's distinct.
I would hope that we'll see that same dynamic for car. Something like half of our new life customers and our new pet customers were existing customers, but the other half were people who are new to Lemonade. Those, I think, are good guideposts, and we'll share more before too long as we get smarter about our specific launch.
Great. I may have missed this, but two really quick ones. Just from a regulatory perspective, how many states will you start with?
We've not announced that specifically, but again, I would point to our other launches. We like to have a broad launch, but we're also speed to market is important, so it'll be a balance of those two. We'll be launching in a good proportion of our customer base. Again, as we get closer, some of the regulatory hurdles can be a little unpredictable in terms of timing. We do know the outcome, but we don't always know the exact timing. Again, that's part of what the real focus is this year, is both internally on the product and the customer experience, but also the external regulatory environment. We're parallel pathing on all of those efforts.
Great. I know you clarified kind of margin impact from investments you have to make in the auto product. Does revenue guidance include auto, or is it just without it? You included auto just on the margin impact.
We've not given specific dollar figures, but I can guide you in a couple ways. If you look at our employee base, 661 employees, about half of those are customer-facing, and about 25% are product and engineering. What you see today, most of the focus and the effort for our new product launch is in the product and technology area. While we're continuing to hire, and those numbers will grow over the course of the year, a significant proportion is already reflected in that number and is definitely reflected in the guidance for the rest of the year. The customer-facing investment comes closer to launch. Again, we've layered into the guidance all the expenses we expect to incur this year based on our current launch plans, and essentially no top-line impact. It's a pretty conservative approach.
It is the largest team within the tech group focused on Car. Everything's layered into the guidance.
Thank you.
Our next question today comes from Josh Shanker with Bank of America. Go ahead.
Yeah, thank you for taking my question. Is there any detail you can give us on gross premium in force or policy mix by type, and if there's any changes in your average premium by type from information given in prior filings?
I can give a little bit of color on that. One of the things we noted is that our new business coming in is heading towards roughly 50% renters. Previously, that was significantly higher, and that's been declining. Not declining, but the home, pet, and life has been growing at a pretty healthy clip, and so now that mix is closer to 50/50. The overall book of business is not 50/50. That's the new business coming in. The overall book of business is heading that direction. Pet in aggregate is still running around 10% of the book, and homeowners still in 30% range. Think of it as probably roughly 40% of the overall is new products. Life, just getting started.
That's on a premium basis, or is it on a policy basis? Premium basis, not a policy basis.
Yes, that is correct. On a policy basis, the shift has continued. Roughly 90% of our customers on a customer count basis are renters, and the balance, homeowner and pet policies.
No material changes in average price per policy over the past few months, or six months, I should say.
No dramatic changes. Continue to pass trends and really the mix shift driving the bulk of the increase in premium per customer. There's always roughly a third of that premium per customer change has continued to be driven by greater coverage, more so than price increases.
Then on Pets, to what extent are you expanding the TAM, and to what extent are you having pet insurance customers be switchers from another carrier's policy?
I don't have the switch percentage at hand, and I'm not sure we've disclosed that yet, but we can maybe follow up on that if it's something we've shared. It's definitely a combination of folks who are first-time buyers of pet insurance, as well as some switchers. It's a much less penetrated market, so in the U.S., it tends to be a greater probability that somebody who's not had pet insurance before. That's all I have at hand right now.
Great answer. Thank you.
Thank you.
Ladies and gentlemen, this concludes today's question and answer session and today's conference call. We thank you all for attending today's presentation. You may now disconnect your lines and have a wonderful day.