Hello everyone, and welcome. Today, I'm delighted to have with us, CCCS and CEO of the company, Githesh, and Brian, CFO of the business. Githesh, Brian, thank you very much for coming. Glad to have you with us. As a starting point, could you discuss in some detail the amazing business that you're running, which I feel is somewhat less sensitive to a lot of the macro trends and headwinds that some of the other SaaS companies are experiencing?
Thank you, Alexei. Good morning, everybody. When I look at CCC and the way we've operated for several decades, it's actually a very stable underlying macro environment in which we operate. By the way, quick show of hands, how many of you drive a car? How many of you have auto insurance? All right, similar number. All right. If you look at it, auto insurance is mandatory and it's a very fundamental part of really what gives the stability to this industry and our business. The number of cars on the road has steadily increased over the last 2 decades. It continues to increase. Half GDP growth comes from population growth, so the number of cars keeps increasing. The underlying macro environment of auto insurance is what provides stability. That means our customers can also plan for the long range.
What we've also seen is through various economic cycles, whether it's 2008, whether it is 2000, we've seen a very steady underlying basis for auto claims and auto insurance. That's what has given us that stable base. Of course, we've been able to drive a tremendous amount of innovation on top of it, and that innovation is really what's driven the growth for our company.
I was very lucky to be at one of your recent events in Arizona. One of the things that I took away was the announcement or soft launch of the Estimate-STP. Could you discuss in a bit more detail what is it and how is it different versus what your customers are using today?
Sure. When we talk about the I in ISTP, it stands for intelligent, and the STP stands for straight-through processing. If you look at two macro things, vehicle complexity and process complexity, they're massive. Both are increasing pretty rapidly. Intelligent STP or straight-through processing is the ability to streamline, digitize, and automate a lot of these capabilities. Over the last decade of developing our artificial intelligence, the reason we call it ISTP instead of STP, is that the ability to apply artificial intelligence through our tech stack at every facet of this capability, we think allows our customers to deliver a very unique experience for that particular policyholder on that particular day for that particular damage in that particular geography. That is what intelligent STP does. Our customers are all seeing shortage of people. They don't have enough people.
They need more capabilities, it is really, getting a lot of excitement from customers.
One other element, that I thought was quite fascinating is, the level of trust that I felt you were able to establish between a lot of the parties that are involved in this very complicated process of this, auto economy, as you call it. Could you discuss in a bit more, what it gives you in terms of perhaps your NPS scores or your customer relations and how you've been able to build that?
Sure. Because, you know, it's over $250 billion are spent in auto insurance, in what is spent on auto claims, there are many different parties. Many of these parties are often can be antagonistic or have different priorities. Maybe that's a better way of saying it. You have insurance companies, repair facilities, parts providers, banks, car companies, so many different participants involved. Over the last couple of decades, what we really have felt is building a very deep network of participants where the CCC information becomes the standard of communication or the transparent way. For example, if you look at a, an estimate, the average repair estimate is $4,600. Insurance companies ultimately pay for the repair. Repair facilities repair the vehicle.
Car companies, parts providers, so many other people are involved. It has to be a trusted company and a trusted data set that all parties can agree. That has allowed over the last couple of decades, as more insurance companies, more repairers, more parts providers, the network has grown, and along with that, the network of trust, the trusted network has grown. As you know, you know, that's an to the square function when you have kind of that exponential growth. That's been very important to make the process more efficient.
That phenomenal NPS, like frankly, I haven't seen that high level of NPS across many of my companies that we cover. What helps you achieve that? I mean, what's the secret sauce behind it?
I think you're referring to our NPS of 82.
Right.
Right? Because the average software company has an NPS of 14.
Exactly.
14 to 82 is a big leap. We had always felt that we wanted customers to use our products because they love the product, not because they needed to use the product. Now it does both. There's about six or seven things that go into the NPS function. First, the platform itself. It's gotta be super easy to use, so we spend an enormous amount of time on the user experience, the tech stack, the delivery. We deliver 99.9%+ uptime. We delivered 1,500+ software releases last year. That means the pace of innovation that you're constantly delivering for customers, that they're getting value for their money, the impact in terms of efficiency, and at the end of the day, it's the references mean an awful lot.
When one customer tells another customer, "This CCC's products are fantastic and had huge impact on me," that has a far better impact than any marketing or any other material we could put out. For us, NPS has been true north in that sense.
I think that it would be fair to say that CCC has been investing into something that now is on everyone's mind, AI, for a very long time, more than seven years now. I was wondering if you could remind us all of your AI-enabled products at this juncture, as well as what are the benefits that AI could bring to your customers?
Sure. you know, underneath, it's been about a decade worth of work that we've put in from building the very, very early AI models. We started deploying Estimate-STP, our first really complex piece of AI, where you can go to your car after an accident, take pictures, go click, click, and then the AI understands the three-dimensional complexity of the vehicle, the parts, structure, damage, et cetera. That's actually been a game changer, right? Since we delivered that in November of 2021, I think the vast majority, well, over 50% of the industry has adopted that solution, including seven of the top 10 carriers. We have also been deploying AI capability because we've been building it, AI capabilities, for almost a decade.
Estimate-STP, it's deployed in subrogation, it's deployed for repair facilities. We've also been using large language models. We will send 80 million text messages between repairers and consumers to keep them updated as to what's happening during the repair process. We also use contextual AI and have been using it for some period of time, but we see AI turbocharging our tech stack in line with existing workflows. We think that's a huge differentiator, is to be able to deliver AI inside the existing workflows.
Could you talk about some of the current penetration levels that you're seeing of that straight-through processing within your total volume claims?
Sure. Brian, do you wanna take that?
Yeah. As, as Githesh said, out of the top 10 carriers, we have seven that are using Estimate-STP. They're at different stages. Some are at, you know, full rollout, while others are at the early pilot stage. As Githesh highlighted, if you just look at the volume that those carriers represent, it's over 50% of the market. We're still really early, we highlighted that we're about 1% of claims are actually running through Estimate-STP, we certainly see the clients progressing through adoption and the ramp-up. When we think about the sizing, if you look at how we frame it over the long term or medium term, we, you know, there's over 20 million claims, 80% are repairable.
We think about looking at the repairable side, about 30%. We use mobile as a proxy. Today, we have about 30% of claims running through our mobile capability. If we get to 30% using Estimate-STP, that represents similar to our mobile channel. The price point that we've been talking about is for a no-touch claim is about $20. If you run through the math and say, okay, $15 million repairable and 30% times $20 as a price point, you know, that's over or that's around a $100 million opportunity that we see, and we certainly are seeing the progress as we start to ramp.
Perfect. Obviously everyone is concerned about potential disruption that AI could bring to some of the businesses, but I feel like something that people don't realize is within P&C Insurance, you know, all the crash data, it's sitting behind various firewalls, right? It's not readily available, while simultaneously you guys have $1 trillion worth of data available to you. Clearly that's a competitive advantage.
Yes. Part of the reason for that is that there's a lot of sensitive proprietary information in repair, in estimates, in vehicle information, and some of it is data we produce, some of it is from partners. But because of the complexity and the privacy issues. That's where the trust part that you're talking about comes in in a huge sort of way. Being trusted by all sides to be trusted with this information has been super important. We have about $1 trillion of historical data that's also allowing. We collect well over 500 million photos a year. That allows us to continuously update and tune and keep the AI very, very accurate. That's a huge part. That data set, not only is that a big data set, but the data set is changing every single day, every single minute.
Like, if you look at what we'll process in one day, that's a massive amount that we process in just one day. You have to keep updating and changing the model as well.
Right. That sort of, kind of takes me to the question of some of these competitive models. I don't think that many of them anticipate certain drifts that are happening constantly with these models. What is it in your model that makes your business stand out, and how is it different?
Almost every, you know, trust is, as we've talked about, an extraordinarily important part of what we do. When we produce our AI and we produce our solutions, we also predict how do we feel about the accuracy of the prediction we're making.
Mm-hmm.
We will actually upfront say, "We have very high confidence. We have very low confidence." We actually mathematically calculate the confidence levels for every single thing we do. With intelligent STP and routing, the routing can be adjusted based on the confidence level of the AI. That has also helped build a lot of trust over time because our big concern was people find it pretty exciting when they first try this out. They say, "This is something that required so much human ability to be able to do. How is an AI even able to do this?" Right? People went from, "I don't think this is even possible," to, "This is amazing." We didn't want people to actually over-index on that 'cause we see a long ramp, a very nice ramp over this.
The other thing we've done is that the data we collect every single day drift, you know, the drift that you're talking about. Drift is really the delta between what is accurate and how the data is changing. Because of the fact that we're in these flows every single day, we're actually monitoring, adjusting, and can deploy the models very, very quickly through our infrastructure. That keeps the accuracy really intact.
How are you able to apply some of that AI technology into your clients' workflows?
Because we're already connected into people's workflows. Take mobile. Our systems over the last 20+ years, we have very deep integrations between us and our insurance customers. On the repair side, we have, you know, 28,000+ repair facilities, and we are the platform in which they operate every day. Our largest repair customer has 35,000 employees. Our smallest repair facility has three employees. Our platform scales up and down through the whole thing, and every single person has work queues and the like. Because we're already integrated deep into the existing workflows on every facet of it, parts providers and the others, inserting and delivering the AI in line with the existing workflows makes it super easy.
Understood. Perfect. If I could maybe change gears slightly and talk about some of your new opportunities. I mean, it's clear to us that you've established quite a dominant position in auto physical damage. Could you discuss your current exposure and casualty and, you know, what makes you think that you can replicate the success that you've had in auto claims?
Maybe we'll do this in two parts, and maybe Brian, you can take the second part.
Yeah.
When you look at the physics of the accident, 15% of all auto accidents have a medical claim associated with it. There's a linkage and a connectivity. We actually have physics models and AI models that can actually look at the physics of the accident and then help predict what kind of injuries result, what kind of payment should happen. That's a very deep connection. Then maybe, Brian, you can take the other part.
Yeah. As far as the sizing, today, casualty is about 10% of our revenue, say $80 million. The auto physical damage side of the business is closer to $300 million. If we were at similar level of adoption, those would be similar-sized product lines. We think the opportunity is moving from $80 million - $300 million, and doing that over time. Today, we have 27 of the top 30 on our APD side, and much lower penetration rate on casualty. That's the opportunity we have in front of us.
Perfect. Thank you, Brian. Talking a bit more about one of the somewhat newer elements within your story, which is subrogation, could you talk about the opportunity there and the demand for that product that you're seeing?
Sure. Just for the audience, subrogation is a process of one policyholder hitting another policyholder, and then the person at fault collects the money from the other side. That happens behind the scenes. Today, roughly 15%, 20% of all claims are subrogated. The dollars involved are massive. About $20+ billion in subrogation. $2 billion- $3 billion is just administrative cost. What we've been able to do is to deliver a brand new product, and again, the artificial intelligence plays a key role into it, where the AI can now understand the claim-related information about who is at fault, which ones should you collect, and which ones you shouldn't collect. We've been applying a lot of that capability to subro.
We've just literally introduced the product in the last few months, and the receptivity has been absolutely terrific from customers. The heart of subrogation, the underlying information is a valuation, is a total loss report, it is a medical claim, it is a tow bill, it is a repair estimate, all pieces of information we understand very deeply.
Understood. If I may ask about some cross-sell and upsell opportunities. There are quite a few products that you discussed, diagnostics, Calibration. Can you talk about what role does upsell and cross-sell play in your long-term growth strategy and revenue assumptions?
We frame long-term growth targets of 7%-10%, and we talk about that as an organic position. What we've said is, over time, we'll be moving towards about 20% of that will be coming from new logos. Historically it's been higher, because of the penetration rate we have with the carriers, we see about 20% in the long-term model coming from new logos, 80%, to your point, is coming from cross-sell, upsell. Then we've said about half of that will be coming from these newer solutions that we've more recently launched and brought to the market. That's Estimate-STP, as we've talked about, it's a subrogation product that Githesh just talked about. It's diagnostics, it's payments.
Those new emerging solutions will be about half of the cross-sell, upsell going forward.
Could you maybe elaborate a bit more why you think products like diagnostics could be important for some of your clients, and what is changing in the market that's driving that demand?
Yeah. If you look at, if I look at the data from, say, four or five years ago, you go back to 2017, less than a very small percentage of cars needed to be scanned. Today, a modern car after an accident, there's a lot of sensors that go off that need to be calibrated. You basically have to plug your plug into the port in the car, the car from a diagnostic standpoint tells you, "These sensors are broken. This camera is not working. This thing is not right." You have to scan the car before you start repairing it, knowing what to repair. You have to sometimes scan it during the repair process, you have to scan it afterwards. You also have to calibrate.
For example, if it's a adaptive radar, the radar has to be calibrated exactly right. Cameras have to be calibrated. This was happening in less than 1% or 2%, you know, a very small percentage, about four or five years ago. Today, the scan rates in vehicles are well north of 50%, and that's the only way to actually repair a car properly. Our technology allows repair facilities to integrate scanning into CCC ONE, but also gives verification capabilities to the insurance companies and the OEM that the car was repaired properly, the scan was actually done, the calibration was done. Again, back to the issue of trust that you talked about earlier, it provides transparency and trust across all of these different parties. That's the actual problem of diagnostics.
In terms of sizing, it applies to as, you know, it'll apply to virtually all cars down the road. I don't know if you wanna add anything to that?
No. Thank you.
Yeah.
No. You covered it well.
Perfect. Thank you, Githesh. Brian, maybe some financing question to you. What factors within your business model you think could help you get to some of your long-term targets? I mean, they sound very impressive, mid-40s EBITDA margin. What do you think is behind that?
I mean, we've seen significant margin progression over the past few years. If you look at the four years prior, we moved from 30% margins to 39% margins. We saw 900 basis point improvement over a four-year period. This year we're highlighting that we're gonna get to 40% EBITDA margins, and the long-term targets are getting to the mid-40s%. We've proven the ability and the operating leverage that's in the business. We see that will continue. We see gross profit margins moving from the mid-70s% to 80% over time. There's certainly leverage in sales and marketing, G&A. We'll continue to put significant investment into R&D, which we've done. We've put over $1 billion of investment in the last 10 years into R&D.
We'll continue to invest in R&D, and that will be our fastest cost growth. We feel really good on the efficiency of the business, the operating leverage, and our ability to scale. Mid-40s% is a target that we're putting out there in the, say, five to seven-year horizon. That's certainly not a ceiling. We see the opportunity to go beyond that, but we wanna put a target that we can track against in the medium-term range.
Again, slightly changing gears now, moving to some of the recently announced partnerships. I was quite impressed with your partnership with Verisk. Could you elaborate a bit more what this partnership could bring longer term?
Sure. At a very high level. As digitization increases, we also have to make sure that fraud protection is available to all our customers. Using some of Verisk datasets, our datasets and other datasets, allows us to deliver more fraud and other types of solutions. You'll see more and more partnerships as well.
What sort of partnerships do you think are still required to strengthen your portfolio of services to your clients?
You know, there's a number of them, right? For example, we brought on board almost all the OEMs that, you know, they play an important role in collision repair. We brought on board more and more parts providers. We're also bringing on board banks. The reason we're bringing on banks is that billions of dollars of cars, once they get totaled, the title is actually held by the bank. That needs to be electronically transferred back to the carrier. There's a whole process of efficiency, you know, the STP or straight-through processing concept for total loss valuations. We see partnerships across almost all facets of it, just continuing to increase.
Could you also talk about your aspirations in terms of geographical expansion? This business is predominantly U.S.-based. There is very little small China exposure, but how are you thinking about it longer term? Where do you see TAM opportunity for your business?
Sure. If you look at the largest pools of vehicles, China and U.S., based on your analysis, could either be number one or number two in terms of the installed base of vehicles. We also felt being in China as we've been in many years, allowed us to test our muscles competitively to know that we can enter a very complex market and actually be successful in terms of penetration, although it's very small revenue. The largest opportunities continue to be around Europe. There's about 35 million claims in just five countries in Europe, so Europe tends to be a large opportunity. Asia, our platform itself is more agnostic in terms of language dependency and other things that we've chosen the tech stack to be very independent of language and the like. We do see over time scaling to other geographies.
To be very clear, our internal plans that we laid out for the next several years, we can grow very comfortably and adequately with zero exposure. That's not our intent. With zero exposure to, you know, any external geographies.
Yeah. I would just add, I mean, the seven to 10 target we talk about is an organic position, it's U.S.-focused. When you think about expansion internationally, we look at that mainly through M&A or partnerships. We don't see ourselves going through a long organic build-out in an international opportunity. It will be partnerships or M&A.
Any more M&A opportunities in the U.S.? Is there somewhere where you feel like you could add some expertise?
Yeah. I think the subrogation example is a good example. We see product, the ability to buy areas, that are very, complementary to our solutions and integrate those and then cross-sell them into a bundle like we do today. Subrogation, as Githesh talked about, that's a very natural fit with our product portfolio. Instead of building that, we went out and we bought a business that was, you know, early stage, pre-revenue, but had really good AI models. We integrated that and built out the platform. We see kind of product tuck-ins or technology buys as a really good fit for us, from an M&A perspective.
Outside of M&A, could you talk about or remind us your capital allocation strategy?
Yeah. Right now we're about 1.5 x levered. We have a very efficient balance sheet. We have highlighted that M&A will be a focus for capital. We wanna deliver against our strategic priorities, but over time, we'll continue to evaluate the best use of capital and make sure that we're driving long-term returns for our shareholders.
Perfect. I'm wondering if there are any questions in the room. If there's anyone who wants to ask a question, please raise your hand. Question on pricing, how do you think about that?
Sure.
Sure. My question is on pricing. How do you think about that?
Yeah, sure. Our pricing at the carriers is largely focused on an ROI basis. We look at the value that we drive against our clients and price the products on a return. It could be a 3:1 return, a 5:1 return, or a 7:1 return. That's how we establish pricing. It's not contingent pricing, we look at the value that's being driven from the solution, the efficiency it allows, and then we set the market price based on those set of ROI examples.
It would also, 80% is subscription, multiyear subscription revenues.
That's right.
Hi, how are you? Just with the, with sort of this, you know, what's happening with generative AI, would there be any risks that, let's say, one of your large insurance customers, would there be any opportunity for them to say, "All right. Well, let's try to build our own AI model using generative AI"? Or does it still sort of play into your favor 'cause you're sort of cross everyone in the industry? I'd just love to hear your thoughts on that.
Sure. Two or three 3 points around that. First of all, many of our largest customers have multi-billion dollar IT budgets, right? Before they made decisions to use CCC and adopt Estimate- STP, and some of our AI, they have been building their own AI and will continue to build their own AI. In this particular area, the accuracy of what we can deliver really complements what they do. Where we see this developing is that integrating unique customer experiences that our customers want to deliver married with the unique AI that we deliver. For example, they'll have a lot of behavioral data about their policyholders and how they need to be serviced. Marrying that with what we deliver we think is a great fit, we've architected to be able to do both of those.
Now, when it comes to actually delivering generative AI for, you know, an estimate or repair or any of these things, having an independent third party that is a trusted party across all of these different parties, we think continues to be extraordinarily important, especially when you're dealing with hundreds of billions of dollars.
Great. Thank you very much, everyone, for coming. Githesh, Brian, appreciate your time today.
Alexei, thank you so much.
Thanks for having us.
Thank you for having us.
Thank you.
Thank you.