Root, Inc. (ROOT) Earnings Call Transcript & Summary

Okay. Good afternoon, everybody. Thanks for joining us for the next portion of our conference here. We have a fireside chat this afternoon lined up with the management team from Root Insurance. I'm Mike Phillips, Morgan Stanley's property and casualty insurance analyst. And again, we're joined here today by Alex Timm, CEO of Root Insurance; and Dan Rosenthal, the CFO. So thank you both for joining us. Before we get started, let me just go through the disclosures, and then we'll kind of get right into it. I think we have about 30 minutes. So for important disclosures, please see the Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative. Okay. With that out of the way, Alex and Dan, thanks for joining us. Appreciate your time. I think we got a pretty good audience out there, and we'll go through some questions. And then we've got about 30 minutes. We can kind of jump around if we need to. We'll see if there's any questions from the audience, which I have here on my side. So thank you guys both.

You guys have a really unique company out there, and it's been -- create quite a lot of buzz. So I'm going to assume a lot of people in the crowd know about you. I guess I want to start by just talking about why, how it got started. What was the impetus for you waking up one day, Timm -- Alex, we know your background, but kind of just why Roots? It wasn't overnight, but you woke up 1 day and said let's do something different in insurance and personal auto. What was that impetus for that to happen?

Absolutely. Thanks, Mike, and thanks for having us. Yes. For me, I'm one of those weird kids where, as long as I can remember, I said I wanted to start an insurance company. So you can imagine I wasn't exactly popular in high school. But my father, he started an insurance company. He was an insurance entrepreneur; and at a very young age, I started in the business and then became an actuary. And I was working at Nationwide really when the idea for Root sort of came about. And there were 2 really big things. The first was, there was clearly a large distribution shift happening. Mobile was the fastest growing retail channel in the country; and yet, there weren't any real insurance companies that were actually well positioned to capitalize on it as a demand generation platform. And for the most part, it was an industry in turmoil, where you saw the exclusive agency channel has shrunk in half over the last 10 years. And really, consumers were fleeing to the Internet. And you saw GEICO and Progressive really just sort of enjoying the fact that they simply got to the Internet first. But consumers were still moving, they were on mobile, and yet we still haven't seen anybody move to mobile. And the second thing, and something that was really near and dear to my heart as an actuary, was through this change, we were able to actually use all of the rich sensory data off the phone to better price insurance. And there was this thing called machine learning or artificial intelligence that was starting to get talked about, and everyone was saying -- talking about big data and that, really, what I saw was the opportunity to fundamentally change the way that risk is underwritten and priced in a better way, that's better for consumers and better for society and would create a better business long term. So I quit and -- in 2015 and started the company. And the idea was we were going to build an insurance company that met consumers where they are, which is on the smartphone. And by doing business this way -- I mean these things are effectively supercomputers that people sleep next to and carry around with them everywhere. And by doing business that way, we will be able to collect a lot richer behavioral data base rates on how consumers are actually driving and by doing so, giving them better prices and creating a better book of business over time. And so that was the core idea. In 2015, we said we wanted to do this as an insurance carrier because -- operated inside of an insurance carrier. I also saw the benefits of actually rebuilding an entire insurance company from the ground up based on modern technology versus trying to outsource technology, which never seem to work very well for us. And so I said, we also have to do this as a carrier. And it was a wild idea at the time. We launched in 2016. We've been growing ever since and recently, as you know, Mike, went public in October of 2020 as we scale.

Perfect. That's a good intro and good background how you got started.

Thanks, Mike.

I guess, a large part of what you do is you mentioned the phone and technology is tracking how we drive versus the old factors of pricing personal auto insurance. So I guess before we really get into the details of it, can we just kind of make sure people understand what do we mean when we talk telematics, when we talk UBI, usage-based insurance? And there's other things that are kind of a subset of that paper mile, which is a little bit different, obviously. So just telematics and UBI, let's make sure people are aware of what we're talking about.

Absolutely. So you're right. Insurance is basically you're selling a promise, and that promise, you're pricing just based on math and predictive analytics. And so what we do, the bedrock of the industry is really data. And really when people say UBI or usage-based insurance, what we mean is pricing based on how much you actually use. What is the fundamental exposure underlying the risk, right? Is it miles driven? Is it how you're driving those miles? And so we do things like look at how many hard breaks an individual might take. How fast they might speed, whether or not they text and drive, which is a major cause of accidents. And we leverage all of that data to come up with a better price. Now as you had mentioned, there's an interesting billing feature that you can certainly build an offer, which is paying per mile. That's less of a pricing innovation and more of a billing innovation because it really just changes the lumpiness of a consumer's bills, if you will, versus the total amount. But that's also something that is sort of in this space.

Good. So I mean telematics -- and thanks for the definition so we're all aware. I guess it feels like telematics has been around for a while. I know Progressive's had it for a number of years, but theirs is different than yours through the plug-and-play device in the car. But telematics has been around for -- Progressive has had it for close to 10 years, I think, in a different way. So how different is what you're doing than what's already out there?

So yes. That's a great question. And when I started the company, everybody actually thought mobile telematics was impossible. And they said, the big problem was nobody could get consumer traction with it. You couldn't get consumers to try to plug in all these things to their vehicle. And as you had mentioned, actually, telematic goes back multiple decades, and there were even big black boxes that folks, I believe it was at Carnegie Mellon, started to put into cars a long time ago. And the problem with those is they were clunky. They were difficult for consumers to use, and the devices were very expensive. But the data you got off of those devices was also incredibly easy. It basically gives you a speed reading of the vehicle once a second. And if I were to show you that, you'd be able to -- I could show a human that. A human would be able to make sense of the data. It's very smooth. It's very natural. But there is a lot of substantial technology changes and differences between the sort of old school UBI, which has, like I said, been around for decades and mobile. And the reason for that is when Apple made the iPhone, they didn't say, well, how do I make a really great telematics device, right? They said, well, how do I make a really good phone? So not surprisingly, you have to do some pretty innovative things in order to actually understand how do you take phone data and then translate that into real risk insights. And if I were to show you mobile data on this screen, it would just look like sheer noise to a human. So you actually have to ping several different devices multiple times a second, upwards of 16 times a second and then use advanced machine learning to actually do signal processing. And that's very difficult. So even understanding whether or not something was a hard break versus not a hard break, you can imagine if the phone's sitting on the seat and kind of sliding around. Those hard breaks look different than if the phone's in your pocket. You can imagine driver passenger classification. I now have to know whether or not that was really you driving or whether or not you're on a bus. And so you have to use all of the really advanced analytics to be able to make it work, and you have to do all of this while creating a more modern technology stack to actually innovate on this. For instance, Android phones are very different from iPhones. You also have cheap model phones that have considerably different data quality off the phones. So it's very, very difficult, and you have to do it all while also not draining the battery. So the technology -- and that's why you've seen even Progressive, who, as you mentioned, has done this for over a decade, they built all of their old OBD-II models, the device models in-house. But then what happened when mobile came along? Well, they actually outsourced it because the technology is an entirely different ballgame. And so if you look at most of our competitors, they've actually outsourced the development overall -- almost all of their telematics applications. And that's why it's so different, and that's why a lot of people, like I said, didn't believe you could do this. And outsourcing doesn't work because if you outsource, you don't end up actually being able to tie all of the predictive data off of the phone back to actually the underlying claims data and incorporate it truly into your pricing model in a complete way. And that's why our predictive power has just been proven and seen -- it was validated by Milliman to be over 10x more predictive than any of the leading telematic service providers in the market.

Okay. I guess, speaking of competitors, and I guess, as you get on the list of the top 10 personal auto companies out there, we talked about Progressive already. But what's your sense of how commonplace telematics already is for, I guess, the top 10 players out there? Are they using it? Is it being used in pricing already? Is it being used accurately? Just how commonplace and well done is it for the top 10 peers out there?

So I would say everyone does it, but everyone does it very poorly. And it's not just a matter of the technology and outsourcing of the technology. It's also a matter of a very large innovator's dilemma. Now if I was to show you just the actuarial pricing curves for telematics, what you would see is that, just like most actuarial variables, the bad drivers are way worse than average. And they're really driving the cost for predominantly everyone. So the worst 5% to 10% of drivers probably account for roughly almost 50% of all costs in the system. And so really, we should be charging those drivers -- it's usually somewhere between 2 and 3x the average price of what they're getting charged today. However, if you look at our competitors, Root's the only one that's accurately pricing those individuals or just underwriting them out of the model. Progressive is the only one that actually, last we checked, has an actual surcharge for those bad drivers, and that surcharge is maxed out at about 20%. So no one's using the full curve. Most of these competitors, when you look at the top 10, don't even have any surcharge. They've got no surcharge. And if you don't surcharge those bad drivers, you can't afford to discount the good drivers to the extent that they deserve. So most of the industry has some sort of telematics application, very -- it's usually very small percentage of their book of business. It's sort of seen as a marketing gimmick, and nobody is fully utilizing the actual full power of the actuarial curve because they don't want to disrupt their existing business model.

Yes, it does feel like a big marketing thing right now. If you pay attention to the TV ads that are out there, and we have our product just for you. And whether or not that's being done accurately or not is, from where I sit, to be determined.

Absolutely.

But you touched on this earlier, and the massive amount of data that you have is, I'm sure, is way beyond my scope of even knowledge. But one of the things that must come up is just the reliability of the data. You said at the car, if the phone slides into the seat or is in your pocket, you can still tell if it's hard breaking or not. And that's just amazing to me how you can do that. And I'm sure we can talk for hours just on that one topic. But any kind of summary you can say about how you can feel confident that you've got reliable data you're using to get the best price for your product?

Yes. That's a great question. What we do is we are always back testing all of our models. And the way that we -- why we have faith in this is we assign that risk score in that discount upfront, and then we reassess it at the end of a 6-month period because they're 6-month policies, and we also assess it at the end of 12 months. If the data substantially changes, we will adjust prices or underwrite out someone and less -- but that's less than 1% of our customers, so very few people. If gaming was a big problem, what we would see is that we'd really have to rescore a large percentage of our population, and we're just not seeing that. We're seeing that old habits die hard and driving's one of those. And that first 2- to 3-week period really is a good enough indicator for us to be able to price longer term, and that's shown in our loss ratio lift charts.

Okay. Yes, old habits are hard to die not just for drivers but also for insurance companies on how they price the product. I go back a number of decades when I was doing kind of personal loan pricing myself. And the factors that we used back then are, as you know, still kind of the same. Tell me about the make, model of your car and tell me how far you drive. And tell me about -- a little bit about the driver and then came credit score, right? And now credit score, well, maybe we can talk about that later, is starting to fade. But the point is it's decades that it's been since -- we're still using kind of the same archaic-type pricing. My point is that it seems like it's a industry that is slow moving to change. And I guess I would ask, how long do you think it's going to be for telematics really becomes a -- well, it's been around for a while. It's still not a dominant pricing factor. There's regulatory issues. But how long does it take cake until it really becomes a strong a piece of the overall pricing factor in personal auto.

I think as companies like Root continue to take share, it will eventually be untenable to not really have a strong analytics and data science platform that is powered by telematics. And telematics itself or big data itself is going to continue to change as mobility evolves. We are very leaned into autonomous vehicles and the changes in vehicle technology. And we also -- and we're aggressive, and we've got product folks working on those things. And as that changes, you'll see that this data and really truly operating on a data-first technology stack is going to become more important. And eventually, what you will see is probably similar actually to what you saw with credit score, where a lot of incumbents didn't start to react to Progressive using credit score for the first time until they actually saw adverse selection. And the remarkable thing about that is, to this day, Progressive has a competitive advantage because they're better at using credit scores. And credit scores are incredibly, incredibly simple compared to this technology. So I think it's certainly, we believe, a durable competitive advantage to just be better at data in this space.

Okay. A kind of related question, I guess, is to get to a different angle is how big the market is. We know personal auto, 270, 260, pick a number, billion dollars of premium or $260 billion of premium. But I guess when we slice that down, you were direct to consumer. Currently, that's about 25% of the market. That's basically GEICO and Progressive. But how big is the telematics market? When I think of those other pricing factors that are already been used that we kind of talked about, I don't have a choice in those. My car is my car. My sex is my sex, my marital status, but I can choose whether or not I want to use telematics or not. And I don't know how big that market is, and I assume it's going to grow. I think it will. But what's your sense today of, I guess, the uptick of consumers saying, yes, please sign me up for that one? Out of the $260 billion market, how big is that piece that consumers are saying please sign me up?

We believe it's huge. We really have not run into any sort of top of the market or any sort of pressure. Again, we're still early. And again, we know that, that market is also -- is growing as well. And so we don't think we're anywhere close. I mean with the $266 billion market, no matter how you slice it, it still ends up being quite large. And so we think that there's a massive opportunity. The other thing that's very interesting is, although, as you mentioned, that maybe 25% to 30% of the market today is through direct, that number is growing substantially. We believe new writings versus PIF that the new writings market share of the direct market is at least 50%, probably going more closer to 60%. And so when you see that kind of market share shift and all of those -- by the way, direct customers are much more likely to be UBI users. As that continues to happen and you start to see the distribution shifts, too, there's really a lot of tailwinds when it comes to UBI adoption.

Okay. Kind of one more, I guess, quickly, on this sort of rate topic is there's 2 maybe possible hurdles. That's regulatory hurdles. How big is that in your sense? And just privacy issues, too. Do customers -- the consumers can be pushed back on no privacy and then regulatory issues by state?

Yes. Privacy issues, what we see on privacy issues, first maybe, we don't see a lot of concerns there from consumers largely because the data we're collecting here, one, we're incredibly transparent about the data we collect; and we only -- we never sell it. And so we're also very -- and it makes intuitive sense to consumers that, oh, yes, maybe you should price me based on how I drive. And so they know exactly what we're collecting. We actually share them back insights from their data. They're getting real value. So it's not like some of the other tech companies where consumers may be unbeknownst to them were having their data sold and for purposes that didn't necessarily directly benefit them. So we haven't seen it from a consumer lens be really important. Now I think from a regulatory lens, you will start to see more controls there, and we are well ahead of that. We've been investing in our data privacy and protection for quite a long time. Regulatory hurdles, though, are real. Being managed at a -- on a state-by-state basis is difficult. And all the regulators are really created differently. And it's a material barrier of entry for most competitors. Each state has different laws, has different regulations and is nuanced. And so as we enter into new states, it really is a learning process. And it really is dependent on those regulatory time lines. Sometimes those regulatory time lines come in, and they're weighing better than our expectation. And sometimes they're weighed worse because it's very hard to predict sort of the winds of a regulator. I will say, though, that currently, with the political environment and where we are today, particularly trying -- with COVID and trying to get more affordable options to consumers, particularly when they're staying at home, there's more of a drive for a company like Root. So we have seen the regulators, and we're in 30 states, which is record setting. How fast we got into those 30 states is really unprecedented. And we're going to continue to expand, and we believe we can get into about 85% of the population by the end of this year.

Okay. Good. Let's turn to some of that then. Let's turn to maybe some numbers and some results for a bit. I think that since your IPO, clearly, there's been a little bit of pressure in your stock. I would say part of that -- how big a part. But a part of that is confidence in the data translated into accurate pricing, which then translates into loss ratio on par with the industry, which not -- they're not quite there yet. So you've talked in recent commentary certainly with your last results that you've got some seasoned states, and that's going to bode well for you with what you're doing there. But maybe before we get into that, what would you attribute your, I guess, early, what I would call, maybe struggles of getting the loss ratio reined in early on? You're getting it reined in now. But earlier on, what was causing that, do you think? And then we can kind of go into what we're doing from there.

Sure. I'll talk early on. When you start an insurance company, the hardest thing is you don't have any data, right? So you've really got to just sort of -- we hired external consultants to come in and tried to build us basically a vanilla rating plan. And the problem with that is it doesn't actually work to do it that way. You really have to build your own loss cost models internally in order to accurately price. It's the same reason why not everybody can just go copy Progressive's prices, right, and then have Progressive's results in the industry. It's not quite as simple. And so in the early days, we knew that we had to basically learn our way and pay the tuition of getting that data before we were able to really tune our models. So in the early days, we launched. We started to see data. It came in. And we saw sorts of places that we needed to adjust our models, both on the telematics side and also on some of the traditional variable side, where we saw significant opportunities to improve our segmentation and even at times improve our rate levels. And as we've done that and as we've focused on each individual state, we've brought a lot more of those markets in line, and now we feel very confident that we're at our loss ratio targets in many states in terms of the lifetime loss ratio of our customer. So the other thing that's important to note is new business loss ratio is around about 15% higher than renewal business. No direct carrier makes money on new business. It just doesn't happen. And so that's another thing. It's just not having a very young book will also apply pressure to that. But we feel, in many states and in many customer segments, we've got the pricing totally under control. And we're hitting our target loss ratios now. And really, it's a matter of rinse and repeat.

Okay. Good. Then I'll shift to what I said seasoned states. I took a term from your print last week, I guess, Thursday. Seasoned states, kind of what do you mean by that? Why'd you call them out specifically on your last conference call? And what are some encouraging signs you're seeing from seasoned states?

Yes. Thanks, Mike. This is Dan. I can jump in with everyone on this, although it's been a great discussion back and forth with you and Alex thus far. Look, I think the biggest refrain we got leading up to the IPO was tell us about the TAM, tell us about the technology behind the telematics, tell us about the growth opportunity. And then we traded in the worst week in the market since March and late October, and some of the questions blah, blah, blah from more of the traditional insurance base of, wait a second, how are the unit economics trending here. And we got a lot of questions around that post the Q3 earnings. So we spent some time and said, okay, let's tell investors and let's show the data of what gives us conviction in our loss ratio improvements. We improved 26 points year-over-year in the loss ratio. Let's not only explain it but also segment it out into different categories as to why it happened. There were 6 points related to COVID. Let's set that aside, right? Hopefully, for society's sake, that's not going to be a recurring event. We're going to put that behind us. But what is in the special sauce? And what we realized is, if you look at Root's blended loss ratio, you can't see the special sauce come through because of what Alex talked about. There's just noise in some of the early states where we were using the me-too pricing. Then we gathered the data and learned the model. So the real question is how is Root doing when our pricing model and our telematics algorithm are in market. And that's what we defined as a seasoned state because not all regulators have let that be the case. For us, we started 2020 with only 3 seasoned states. Remember, we wrote our first policy less than 5 years ago, so this has been a work in progress in developing that data. We started 2020 with 3 seasoned states. Now we have 20 seasoned states of the 30 where we're active. And what we saw is, in the second half of 2020, when we had material premium in those seasoned states and you put COVID aside, the loss ratio was 15 points better than the unseasoned states. So apples to apples, we saw that real improvement when our pricing algorithm and our telematics model are in place. So for us, it was let's communicate that data to the market. Those are the early cohorts that you can see now playing out without some of the price shocks, without some of the changes. In the seasoned states, you can see the model working. And then from here, it allows us to say, okay, what do we do next. We season the 10 states that are currently unseasoned, and we march into new states with the learnings from the existing states thus far. And that's what we're excited about with our 2021 plan and beyond.

Okay. I think you just kind of -- you hinted to the next question that is part of that then is a bit of a slowdown in what your previously was national rollout, and now you're going to slow that down a bit. Clearly, that probably caused a little bit of stir on Friday with your stock, I think. So maybe talk about that. What does that actually mean? I think you touched on it, but we're going to season our state -- season the unseasoned states first before you go national. That sounds like what you're doing.

And grow thoughtfully. It's funny. It doesn't feel like we're pulling back on growth because we're guiding to 30% to 39% year-over-year growth off an already significant base. And we're adding 7 really important states to take us to 85% of the addressable market. But Mike, I get it. The market speaks. And if we were only focused on Friday's stock price, we probably wouldn't have done what we did. But we're focused on long term, not just Friday's stock price. And we fundamentally have conviction in the seasoned states. The data tells us that we're right, and so now it's a matter of investing our capital into that sustainable and repeatable growth and then adding new states because we do aspire to be a national carrier. We will be a national carrier in short order. We just don't need to be nationwide completely in 2021. We think adding the 7 states allows us to focus our resources in the right way, touch on 85% of the addressable market by year-end and then build going forward. So we're really excited about the plan. We get that there's a short-term hiccup. And frankly, growth investors want to see us grow, and we're excited to come back quarter-after-quarter and show that growth.

No, perfect. That's a good recap. We are very close to our time limit, so I have too many more questions to even go through, so we can't unfortunately. Let me throw the last one out there for both of you. It's the cheesy one, but I'm kind of just curious for both of you. Like what keeps you up at night worried about your company? What's the biggest thing, oh gosh I woke up in the middle of the night sweating about something? Is there anything that keeps you up worried? What would it be? To either one of you.

Dan, you could start, and then I'll throw mine in.

I just -- I think the opportunity here is endless. This is execution, and let's just make sure we get the right talent onboard. People matter here because we know we have a clear first-mover advantage. We have a great customer base thus far, and now it's a matter of just, frankly, leveraging that and going forward and making the right investments.

Yes. And I would agree, absolutely. The -- I mean, like we said, it's a $266 billion market. We, right now by premium size, are the largest property casualty insurtech out there. The amount and the speed at which we've grown has been exceptional, and we want to make sure that we can continue that for the next 10 years at least. We do believe -- and when I look at the industry, there's an opportunity here to build a market defining historic company and very well could be the largest undisrupted industry today. Now we know that, that's going to -- the road there is not going to be easy, and we don't expect that. We're really some pioneers and the fact that we're operating at this intersection of insurance and technology. And we know that, that can create some short-term noise. But just maintaining the long-term focus and maintaining our focus on execution and focus on the consumer through all of that noise, that's what I think about every day.

Oh, I agree.

And Mike, your question assumed that Alex sleeps, which is probably the wrong assumption frankly. So we're probably -- we're both up at night on various things. The opportunity here is superb.

Perfect. Thanks, guys. Really appreciate both your time. Alex and Dan, thanks so much for your time, and look forward to talking to you again. Thanks very much.

Thanks for having us, Mike.

Thanks for having us, Mike. Bye.

Thank you, guys. Take care.