Progressive (PGR) Q4 2022 Earnings Call Transcript

Progressive (PGR -0.85%)
Q4 2022 Earnings Call
Feb 28, 2023, 9:30 a.m. ET

Contents:

• Prepared Remarks
• Questions and Answers
• Call Participants

Prepared Remarks:

Doug Constantine

Good morning and thank you for joining us today for Progressive's fourth-quarter investor event. I am Doug Constantine, director of investor relations, and I will be moderator for today's event. The company will not make detailed comments related to its results in addition to those provided in its annual report on Form 10-K, quarterly reports on Form 10-Q, and the letter to shareholders, which have been posted to the company's website. This quarter includes a presentation on a specific portion of our business, followed by a question-and-answer session with members of our leadership team.

The introductory comments by our CEO and the presentation were previously recorded. Upon completion of the previously recorded remarks, we will use the balance of the 90 minutes scheduled for this event for live questions and answers with the leaders featured in our recorded remarks, as well as other members of our management team. As always, discussions in this event may include forward-looking statements. These statements are based on management's current expectations and are subject to many risks and uncertainties that could cause actual events and results to differ materially from those discussed during today's event.

Additional information concerning those risks and uncertainties is available in our annual report on Form 10-K for the year ended December 31st, 2022, where you will find discussions of the risk factors affecting our businesses, safe harbor statements related to forward-looking statements, and other discussions of the challenges we face. These documents can be found via the Investor Relations section of our website at investors.progressive.com. To begin today, I'm pleased to introduce our CEO, Tricia Griffith, who will kick us off with some introductory comments. Tricia.

Tricia Griffith -- Chief Executive Officer

Good morning and thank you for joining us today. 2022 was among the toughest years of Progressive's 85 years. The continued effects of the pandemic, high inflation, and the largest hurricane in our history all played a role in the tumultuous environment. Throughout these challenges, Progressive people came together in unity and delivered once again, beating our goal of a calendar year 96 combined ratio while growing both written premiums and policies in force.

The pride that I feel in all of the people of Progressive is indescribable. We have met all challenges head on and continued to deliver the best-in-class results our stakeholders expect. Our ability to meet these challenges is embedded in our culture, which is built on our newly updated four cornerstones, which are who we are, which is our core values; why we are here, which is our purpose. We revised our purpose to be: We exist to help people move forward and live fully.

Our modified purpose builds on the legacy of our core values and our history of challenging the status quo to accelerate progress and equity. We modified the purpose to better unify and guide our organization. Where we are headed, which is our vision. We expanded our vision to include business owners, so it now provides a more holistic view of all the customers we are privileged to serve.

And how we will get there, which is the four pillars of our strategy. We've updated these two cornerstones to better reflect the evolution of our company and to ensure our culture adapts to the ever-changing environment we operate in. While our business will continue to progress, we will always seek to maintain the things that make Progressive Progressive. In today's call, we will once again focus on the strategy cornerstone and specifically will discuss the pillar of competitive prices.

Competitive prices means not just offering the best rate, but also the most accurate rate driven by pricing, accuracy, expense discipline, and industry-leading segmentation. As we have said for the last few quarters, we believe the major personal auto rate increases are likely behind us. In the fourth quarter, we continue to raise personal auto rates but at a pace slower than in late 2021 and early 2022. While some states are still waiting on filings to be approved by their respective regulators, we believe the majority of states are nearing rate adequacy, and our intent is to continue to increase rates commensurate with future loss trends.

Based on public rate filings, we know that our competitors are raising rates to address their own profitability concerns. And as they have done so, we've seen our relative competitiveness improve, which has resulted in improving retention, strong quote growth, and high conversion. We've now actually seen some competitors surpass our post-COVID rate take, which should help us sustain those improvement trends. As a result, we finished 2022 with the best fourth quarter for personal auto new application volume in our history, which contributed to growing our personal auto PIFs 3% in 2022 while also running a lower acquisition expense ratio than 2021.

We're continuously evaluating our media budget and will continue to use media efficiently as we enter 2023, and we will look to capitalize on this hard market. The higher direct acquisition expense ratio in January is indicative of our opportunistic stance, and it reflects the confidence we have of current rate levels in most states at this time. Of course, if the last three years have taught us anything, it's that the future is unpredictable. So, we will continue to monitor the business carefully and navigate as needed to grow as fast as possible while delivering a calendar year 96 combined ratio.

As I said in my annual shareholders letter, segmentation is a key facet of our competitive prices pillar, and nowhere is that more evident than in our investment and usage-based insurance products, which would be the topic of today's presentation. Progressive was a first mover in usage-based insurance. We have had a UBI offering since 1996 when we launched our first product called Autograph. This first effort was limited by the technology of the time and required a professional mechanic to install equipment in the customer's car at a considerable expense.

That first attempt evolved to TripSense in 2004, which was our first self-install option, and then in 2008, we launched MyRate, which is where we first employed cellular technology to upload data to our systems. And then, in 2010, we launched Snapshot, which we consider the start of our modern UBI program. We moved from our discount-only model to one that included the possibility of a surcharge in 2014. And in 2016, we launched the Snapshot app, which allowed the customer to use a mobile phone app instead of the plug-in device.

Our most recent addition is continuous monitoring, which began its rollout in the summer of 2022. In parallel to our efforts in personal lines, we were developing UBI for commercial auto products. This culminated in our first broad commercial offering of Smart Haul in 2018, which allows us to provide usage-based insurance to truckers in partnership with providers of electronic logging devices. We also expanded the Snapshot program for commercial auto with Snapshot ProView in 2020, which includes fleet monitoring services to small businesses.

Throughout our history of usage-based insurance, we have collected billions of miles of data and invested in a process of continuous improvement in our UBI products. Today, UBI is our most predictive rating variable, and it provides unparalleled rate accuracy to our customers. Through this, we have continuously educated our customers where, today, UBI adoption is at near historical highs. To highlight our UBI products on this call, we've invited two Progressive leaders who have played significant roles in our UBI development.

First to speak will be Jim Haas, business leader for our personal auto usage-based insurance. Jim is a 20-year Progressive veteran, and for the last five years, he has led our personal auto UBI team. Jim will discuss the advances we have made in personal auto UBI. Following Jim will be Cory Fischer.

Cory is a 19-year Progressive veteran. For nine years, Cory has been our business leader for commercial lines R&D. And as of last month, Cory accepted a new role to lead our agency distribution group. During his tenure in commercial lines, Cory led R&D during the successful rollout of both our commercial auto UBI products.

Cory will be highlighting all of the advances we have made in commercial lines UBI, once again giving Progressive first-mover advantage in this important technology. Again, thank you for joining us this morning. I will now hand it over to Jim Haas. Jim.

Jim Haas -- Usage-Based Insurance Business Leader

Thank you, Tricia. As Tricia mentioned, my name is Jim Haas, and I lead Progressive's personal and telematics efforts. Today, I'd like to tell you about how we've built on that long history in telematics that Tricia discussed, starting with how we've continued to improve our core Snapshot program, then moving on to discuss how we can now bring in data from outside sources to improve reading accuracy during the new business quote, and finally, wrapping up the discussion of some value-added services we'll be bringing to market soon. First, let's recap how our program has worked for a number of years.

Individual states may vary a bit, but I'll describe here how it works in most places. We are the agent present the customer with the offer to enroll to Snapshot during their initial quote. If the customer elects to sign up, they'll receive an immediate discount of up to 10% for participating in Snapshot that will be applied to that quote. So long as the customer stays in Snapshot, that discount stays on the policy during the whole first term.

During the quote, the customer also elects whether they want to monitor using a plug-in OBD2 device that we send them or via an app on the smartphone. The customer then has 45 days to plug in the device or get the app set up. That point, all they need to do is drive. We'll let them know how they're doing along the way.

Just before their policy comes up for renewal, we'll use the driving data collected during that first term to calculate the renewal discount or surcharge. That rate adjustment will be included in that renewal quote and will be applied to future policy terms. At that point, the customer can delete the app or send us back the device, and they'll be done monitoring. I mentioned that it could go down or it could go up, that is, we would apply a discount or a surcharge.

But that's not how our program started. In the early days, the program is structured a little bit differently, but a customer's rate couldn't get worse by participating in Snapshot. Customers could receive a discount of up to 30%, but their rate couldn't go up no matter how risky their driving looked. Many companies still employ a model like this today.

Of course, no rating variable and insurance only suggests giving discounts. It's about segmentation and a matching rate to risk. So, the data would suggest that some customers should see the rate increase while others should see a decrease. While only offering discounts might encourage more people to participate in the program, it makes the economics a lot more challenging and limits the accuracy of the overall pricing.

We all know that pricing accuracy is critical in auto insurance. So, starting in late 2014, we began introducing a surcharge. We set our factor so that about a fifth of customers received a surcharge, a share that we've kept about the same over time. The surcharge was small at first but start us on the path to greater accuracy.

It's also when we introduced the participation discount, which is down 10% in most states. Over time, we've moved the price closer and closer to what the data would tell us, first, by increasing the maximum surcharge from 10% to 15%, then increasing it to 20%, and finally to where we are today. We offer larger discounts than we did in the past, now up to 30%. But we also have larger potential surcharges than in the past, too, now up to 40%.

We've come a long way since the days of only offering a discount. But it's important to note that about 75% of customers still receive a discount and only about a fifth receive a surcharge. Not surprisingly, we see better retention on customers who earn a discount and worse retention on those who earn a surcharge. Renewal rates for the safest drivers who are earning the biggest discounts are about 6% higher than average, while they're about 16% lower for the riskiest drivers who aren't receiving a discount.

So, the program systematically helps us retain lower-risk drivers and shed higher risk ones. And because of the changes we made to our Snapshot pricing over the years, we're more accurately priced than ever before on those customers that do stay with us. While these changes have certainly improved the accuracy of our pricing, there are still limitations to our approach. We only use data from a single term, typically, four to five months, to set the price for the life of the policy.

This made it so we wouldn't know about material changes in an individual's driving behavior would lead to less accuracy over time. We had experimented with a continuous program a long time ago, back in the late 2000s. At the time, though, consumer acceptance of telematics generally was still growing and collecting data on an ongoing basis was quite expensive, so we decided to move forward with a partial model. Things have changed a lot since then, and in early 2022, we started migrating to a continuous version of Snapchat.

The basic process isn't all that different. Customers still choose whether and how to participate in Snapchat during the quote, and they still receive a participation discount. They still have 45 days to plug in their device or set up the app, and then they can just drive. The difference comes after the renewal as, instead of sending the device back or deleting the app, we ask them to continue to monitor, and we'll adjust their rate at each renewal to reflect their more recent driving.

Additionally, that participation discount is larger now, at 15% instead of 10%, which encourages more people to participate in the program. Having more recent data also improves the predictive power of UBI, which means we can price even more aggressively. We've increased the size of the maximum potential discount to 45%, and the maximum potential surcharge to 60%. We're one of the few companies to combine a continuous model with the possibility of surcharges.

As of the end of 2022, we had deployed this new continuous model in 12 states, representing over a quarter of our net written premium, and plans rollout to most of the rest of the country during 2023. So, while we've been steepening these factors and requiring longer monitoring to make our pricing even more accurate, what's happened to take rate? The answer is that the share of our personal auto customers participating in Snapshot has moved steadily upward. In fact, it's up nearly 40% across both channels combined since January of 2019. So, today, more people participate in Snapshot than ever before, providing data for longer periods of time that we can use to price more accurately than ever before.

So, that's the update on what's been going on with our core Snapshot program, which provides great incremental segmentation over what is available via traditional rating variables on renewal policy terms. Now, I want to talk about how we're working to bring this pricing accuracy to where it's most useful at the time of the new business quote. Since, in this scenario, the customer is only now just coming to us, we'll have to use telematics data that was collected by others. Our long telematics experience has helped us learn how to use this data and how to make it predictive.

We currently have two initiatives in market. The first is a lead generation program that works by inviting good drivers to come and quote with us, while the other begins when a customer comes to Progressive directly to get a quote, and we find and incorporate driving data from third parties into the rate we present them. Let me start with the lead generation program. Here, a partner like Credit Karma invites their customer to opt in to collecting driving data to see if they could save money on their car insurance.

In a privacy-friendly way where we don't receive personally identifiable information, we can let Credit Karma know who's likely to receive a discount so they can invite them to quote with us. Only when the customer chooses to quote are we able to personally identify that data, are then able to apply that discount to that individual's, quote, improving pricing accuracy immediately. Our second program involves working with data collected by automakers. Over the last several years, more and more of them have been equipping their vehicles with technology to collect driving data.

They've been working to show the value of these programs to their customers so that they'll sign up to share that data with them. And we've been able to tap into that. When a customer comes to us to quote and their driving data is available, we'd ask the customer if they'd like us to use it to determine their price. They say yes, we bring that data in and apply the UBI discount or surcharge to their quote immediately, again, pushing that rate accuracy to where it matters most, the new business quote.

Because of our long experience with the UBI and working with both OBD and mobile data, and the fact that we attract so many insurance shoppers every year, we're well positioned to execute and benefit from this. To be clear, automakers are in various, often early stages, of getting the necessary equipment into their cars and making this data available to insurers. Additionally, just because new cars coming off the assembly line have the hardware takes a long time for the fleet to turn over. So, this isn't common yet, but we're excited about the opportunity it represents.

We're already working with two of the largest automakers, General Motors and Toyota, and it's clear that this population will grow over the coming years. Lastly, I'd like to share some exciting news that doesn't involve using telematics data to more accurately rate policies, but instead builds upon our telematics heritage to provide a valuable service to our customers. Over the last couple of years, we've experimented with offering a service to detect and respond to major accidents to some of our Snapchat customers to learn if they value the service and to better understand how it could be useful in handling claims. We've been encouraged on both fronts, as customers have consistently told us that this kind of service is something that they really do value.

And our claims representatives have seen that telematics data can help them settle claims more quickly and efficiently. Still, we know that despite how times have changed, there's a large segment of customers who don't want their insurance premium to be based on their driving data. That means that if we limit this just to our Snapshot customers, we'd be leaving out a lot of others. So, in March, we plan to start making accident response available to all of our auto customers, not just those who are in Snapshot.

We'll use data from the sensors on the phone to detect when a serious crash is likely to have happened. We'll reach out to the customer to confirm the accident and to see if they need help. If we don't hear from a customer at all and it seems particularly serious, we'll request that the police conduct a well check to make sure our customer is OK. Since we know the customer's location from the telematics data, we know just where to send them.

We know there are other accident response services available. We think what we're offering has several key benefits, though, that distinguish it. First, many services come at an additional monthly expense. Ours will come included in the policy.

There is no additional charge. This adds value to the customer's relationship with us and can become another reason to choose and to stay with Progressive. Additionally, while other insurers offer crash detection to their UBI customers, we'll be making it available to all of our personal lines auto customers, whether they're in Snapshot or not. Third, we're deliberate about dispatching EMS.

We certainly want to get EMS dispatch as quickly as we can when a customer needs it. But we also don't want to waste EMS resources and bother the customer when they don't. This can be challenging since the sensor data for a near-miss and an actual accident can look awfully similar. So, we try to contact the customer more than once using different methods so they can let us know if there was an accident and if they need help.

We only send help if the customer's request it or if we haven't been able to reach them after several tries and the accident looks severe. While most customers in these accidents don't actually need an ambulance, many do need a tow truck. We're able to build on our roadside assistance experience to meet this need. We understand that these tows are even more urgent than a typical roadside request.

We can also help customers notify their family members. Lastly, after an accident, one of consumer's top concerns is how to get their medical bills paid and how to get their car back on the road. As their insurer, we're the ones positioned to help them get there. By detecting these accidents and having this telematics data available, we're able to get their claims started more quickly and able to handle it more efficiently.

In fact, if an ambulance or tow is dispatched, we'll actually get that claim started on their behalf. Let me show you an example. This is a real claim that was detected via accident response. In the loss pictured here, our customer was driving on an urban boulevard when they hit some ice, lost control of the vehicle, first striking barricade to their left before striking the barricade on the right.

Fortunately, no other vehicles were involved. As you can see, the airbags deployed and there was pretty severe damage to the front of our insured's vehicle. In fact, this vehicle is a total loss. Additionally, our customer and two of their passengers were injured in the crash.

Within two minutes of the impact, we reached out to our customer. While they didn't respond to our initial push notification, they did respond to a subsequent call from a live agent, which occurred less than five minutes after the accident. That agent dispatched an ambulance and a tow truck to the scene. Only a few minutes later, the claim was filed.

Altogether, it took only 10 minutes from the time of the accident to when we had a claim in our system. We've surveyed some of our accent response customers, and we've gotten some great feedback. For instance, this customer shared that they were "unable to call for help at the time, but help called me." The service helps us help our customers when they need it most. It's not just the immediate response to the accident that can help customers, though.

Here's another example. This customer collided with another vehicle when changing lanes. We detected this accident as well, dispatched a tow for them, and have the claim filed within 12 minutes of the accident. What I want to highlight in this case, though, is that this customer had this accident just two days after buying their policy.

Some of the first steps in handling any claim include establishing the facts of loss and determining if the loss is covered. Unfortunately, there are people who buy insurance after they've had an accident and then try to make a claim for it. This means that honest customers are inconvenienced as we need to take the time to verify that the accident happened when they told us it did. That was pretty easy in this case, however, since we detected the accident in the first place and arranged for the tow truck to pick up the car.

And we can see from the telematics data that the crash happened where and, importantly, when our customer said it did. Therefore, we were very confident that this loss did occur after and not before the customer purchased the policy, which let us move forward more quickly with getting the customer back on the road. So, it's not just the accident response service itself but its promise to improve the claims experience that can benefit our customers. They've told us they value accident response, and we're excited to make it available to them.

We'll start that process in March and expect to make it available to all of our personal auto customers over the next year or so. To wrap up, we're not just jumping on the telematics bandwagon. We have decades of experience here. We're clear eyed about the challenges but also see the great opportunities telematics offerings present.

But this is just the personal line side of the story. My colleagues in commercial lines are developing their own approach about how to use telematics data in their markets. And Cory Fischer, who's led commercial lines R&D for the better part of the last 10 years, is here to tell us about it. Cory.

Cory Fischer -- Head, Commercial Lines, Research and Development Business

Thanks, Jim. As Tricia shared, I'm Cory Fischer, and up until recently, I was the business leader for commercial lines product research and development, which includes responsibility for our commercial telematics efforts. Today, I'd like to share the progress we've made since introducing our commercial telematics programs, some early observations, and how we're thinking about future investments in this space. The chart on the lower left was shared in a prior investor relations call.

It's why we're excited and have been investing in telematics. Pricing segmentation is an important part of how we compete. The group I've worked with is charged with advancing our commercial auto segmentation leadership position in the market. Telematics is the most predictive rating variable we have by a lot.

To build on that segmentation leadership position, we need to be a leader in telematics. For context, we have two branded telematics programs. Smart Haul is the first program we introduced in 2018 after several years of data collection studies and pilots. It targets truckers that are required to maintain hours of service logs.

The launch intentionally coincided with the federal mandate requiring these truckers to move from paper logbooks to electronic logging devices. We recognized those devices provide continuous monitoring and capture very similar data to our Snapshot devices, including location information. With the customer's consent, we use the data from the trucker's electronic logging device to generate a score and apply a rating factor. Ideally, we access the data during the quote process, just like the process Jim discussed using data to price new business.

If we're not able to get the data, either there are newer venture without driving history or we don't have immediate access to their ELD vendor, we provide a participation discount on the quote. Snapshot ProView, on the right, is more similar to what Progressive personal auto has offered. Its target is essentially anyone that doesn't have an electronic logging device. We market Snapshot ProView as safety and savings.

For customers that enroll, we provide a participation discount at new business and send a Snapshot device that they can easily plug in. Given how small businesses have employee turnover, this program has continuous monitoring. We'll then use their driving data to generate a score and a factor that's applied at renewal. This factor could be a discount or a surcharge.

In addition, we provide access to a free driver portal. Our customer can see where their vehicles are and how they're being driven. It's a fairly simple way to keep tabs on their fleet and driving behavior. Since we deploy these programs, we've been monitoring a number of key metrics.

The chart on the right shows a version of take rate, that is percentage of new business apps and rolling into a telematics program. We've seen from prior program launches it can take a while and require a sustained effort to get meaningful adoption. Though we've identified opportunities for increasing take rate, we're genuinely happy with how quickly adoption has occurred and the trajectory it's on. In addition to adoption, we've seen that when quoters enroll into a program, they convert at a higher rate than the quoter that's eligible but chooses not to enroll.

We also see that our telematics book of business performs better from a profitability standpoint even after the discount is applied. That's not the longer-term goal. Our intent is to accurately price all segments of target. Between our significant footprint and the commercial auto market, healthy adoption rate and conversion lift, we built a substantial commercial auto telematics book of business.

That book would be a top 15 commercial auto insurance carrier on its own. In summary, we're really pleased with the growth and performance of our telematics efforts. But we've learned a lot since introducing these programs. One recognition: Getting into telematics is not easy.

It's not just as simple as adding a new rating variable. It takes broader and sustained effort, new capabilities, and investments. For example, we're asking our partners in the independent agent channel to take additional steps to enroll customers, which can include a supplemental app and working through customer reservations while sharing their driving data. That requires an incremental level of support and training.

With telematics, we're now dealing with hardware and data transmission. We're also in the logistics business, sending out devices, tracking inventory levels, and recently dealing with global supply chain issues. That's a different set of vendors, technical challenges, systems, and business processes that we've had to work through. I think the biggest difference is the amount of data that these devices generate.

We have to ensure that data is being managed effectively. It's imperative, as the data is being transmitted and stored, that it's handled securely. Also, as the largest commercial auto insurer in the United States, we have a lot of data and have developed effective tools, processes, and skills to manage and model those more traditional data sets. Telematics is an entirely different situation.

One truck alone might generate over a million records per year. Our traditional approaches aren't nearly as effective, transforming big datasets into variables that we can test and model off of. A big advantage to us is that personal auto, through its efforts and investments over the past 20-some years, has established a lot of these capabilities or learnings that we can leverage. For example, our Snapshot ProView program uses the personal auto Snapshot device.

Given the volume that personal auto has generated, we have a very competitive cost per device, and that device is already set up to transmit that data back to our secure storage solutions. In terms of big data, personal auto helped us move to their cloud solution, essentially a highly scalable cloud-based data storage and analysis service. This significantly reduces our processing time to analyze data and at a much lower cost than traditional alternatives. They've also provided support as we upskill our team to use these new tools.

To summarize this slide, getting into telematics requires a lot of effort. A competitive advantage that we have in commercial lines is that our colleagues and Jim's group have already laid a lot of that groundwork that we can either use or learn from. It takes a lot to stand up a leading telematics program and to maintain that leadership takes ongoing investment. I'll share some of those ideas with this slide.

Near term, there's a lot that we can do to improve the experience. We have data that shows where customers drop out of the funnel or hit operational hurdles, like granting data-sharing consent or installing the plug-in device. We have worked with our vendors and have new business processes that will improve both. We have several manual processes in place that support these programs.

By automating those, we will lower our program expenses and improve the customer or agent experience. There are a few segments that aren't currently eligible for one of these programs. We're looking to expand eligibility. One example is for our small fleet customers that have their own telematics service but not electronic logging devices.

This could be a service fleet like plumbing. Our intent is to enroll them into a program and access the data from their telematics service provider. Through these near-term plans, we'll continue to grow our telematics book and, with that additional data, continue to refine and improve our scoring models. We're also excited about new data sources and applications.

We're seeing better reception from segments that we target to having dashcams. We've been doing a pilot with dashcams for over a year. There's data from video that we don't capture today, like following distance, which could be valuable. And as Jim shared, we're interested in how telematics can help with the claims process.

Dashcams would be an additional input that could be meaningful for claims. We're also using the data to support our underwriting efforts. For now, that might mean verifying other information on the application like radius of operations or garaging ZIP code, which are important for us to assess where the vehicle is operated. Sometimes we need to follow up to confirm that the information on the application is correct.

We can use telematics data to better target which risks we should follow up on. Our telematics score is currently bolted on to our core model. Given not everyone is in a telematics program, we solve our core model first and then add a telematics score factor for those enrolled. Over time, we plan to move more of the telematics data into the core product, especially where we have those insights at time of quote.

Here's an example. This shows one truck's driving pattern for 60 days. On the application, the insured's address and garaging ZIP is Elizabeth, New Jersey, and they've listed their operating radius as 25 miles. Our model would use information closest to that address to determine a territory factor.

Some customers that garage there might never go into New York City. This truck goes lots of different places, but it turns out, there's a fairly regular pattern of going into the city. It's hard to tell from this math, but from the data, let's say 75% of the trips go from Elizabeth into New York City before turning around. A territory factor that puts more weight on the New York City experience would be more suitable.

Part of our road map is to integrate telematics deeper into our model. I don't see a time we have a telematics-only model. We know we get valuable segmentation from non-telematics variables, but there are variables like garaging address and radius that proxy driving location and where the actual telematics data would be a superior solution. Wrapping up.

We've got a great start to our telematics journey. We've got a significant benefit from the experience and investments our colleagues in personal auto have made, and I'm really excited about how we'll continue to use telematics to drive competitive advantage going forward.

Doug Constantine

This concludes the previously recorded portion of today's event. We now have members of our management team available for -- available live to answer questions, including presenters Jim Haas and Cory Fischer, who can answer questions about the UBI presentation. [Operator instructions] We will now take our first question. Stacey.

Questions & Answers:

Operator

Our first question for today comes from Michael Zaremski with BMO. Michael, go ahead with your question.

Mike Zaremski -- BMO Capital Markets -- Analyst

Hey, great. Good morning and thanks for the presentation. First question. Progressive's personal auto accident frequency levels appear to be trending a lot better than the industry over the past year.

I'm curious if you agree with that statement. If you do, any thoughts on what's causing the better-than-historical average trend line? I know, clearly, a lot of this call is devoted to segmentation, Progressive as a first mover -- or one of the first movers in telematics. So, you know, that, theoretically, should, I guess, improve Progressive's trend line versus the industry if I'm thinking about them correctly too. But I would love to hear more about that.

Tricia Griffith -- Chief Executive Officer

Yeah, Mike, I think you're thinking about it in the right way. A couple caveats. When we report frequency, we report incurred and a lot of our competitors report paid. But we've been seeing this trend more negative in the industry for a while.

We can't speak -- speculate on a lot of things. Obviously, segmentation is a big piece, we believe. We also believe that mix of business could play a role, and that could be mix in states or mix overall from a preferred perspective. We talked in the fourth quarter and in my letter about growth across all of our segments that we started to see increase in Q4.

We have seen growth again across all segments but higher in the Dianes, Wrights, and Robinsons. Again, we still see it in Sams, but it's higher. So, it could be mixed as well. The good news is that we see this data change quickly.

So, if we need to change anything we're doing from a pricing perspective, we'll see it pretty quickly.

John Sauerland -- Chief Financial Officer

I might add a couple on that. So, I think Jim has a slide where he showed the retention of customers who are getting a surcharge versus a discount is a great, you know, case in point of segmentation at work. So, the folks who are getting a surcharge are going to be the higher-frequency drivers, and the folks who are getting a discount are the lower frequency. And obviously, in the marketplace, as that continues to play out, that will help drive our frequency more favorable than our competitors.

Additionally, we've talked about underwriting efforts we put in place as we were challenged by profitability. We've raised rates considerably, as we've shared. I will note, though, that we have retained our underwriting restrictions fairly tight. So, we've been able to manage that in this very competitive environment to a point where we're able to maintain the underwriting restrictions, which, generally speaking, are pushing off the higher frequency customers and are able to raise rates at the same time.

So, I think we're in a pretty good place, both from the rate level perspective but also continuing to guard against segments that might be underpriced.

Mike Zaremski -- BMO Capital Markets -- Analyst

Got it. That's helpful. My follow-up is regarding some of the commentary you made about the automakers. And I know we appreciate that.

You know, you're not an automaker and you're partnering with them, and it's in the early stages. But I believe you used the word excited about the partnership. And I believe a lot of investors have felt that the automakers could represent a new leg -- or a new kind of leg of competitiveness within the industry. So, curious if -- you know, if you're having to -- maybe you could shed more light on the partnerships and if -- only if Progressive is having to pay the automakers for this data, why is it a bit of a competitive threat as well? Thanks.

Tricia Griffith -- Chief Executive Officer

Yeah, sure, Mike, I'll let Jim Haas talk a little bit more about that. But we -- you know, we think competition across the board, whether it's with the OEs or with insurance techs, is great. And we also feel like when we have partnerships, like Jim talked about, it's really important, and it benefits, most importantly, the consumers. So, Jim, do you want to add any color to that?

Jim Haas -- Usage-Based Insurance Business Leader

Sure. Mike, you're right, we obviously pay for the data eventually against the automakers. It's a way for them to monetize that. They're putting all those hardware in the car that costs money.

This is a way for them to get some return on that. Some OEMs are going to want to get an insurance, as we've seen, some might not. And so, this is a different way for them to get value out of this -- out of that data from us.

Mike Zaremski -- BMO Capital Markets -- Analyst

Thank you.

Operator

Stand by for our next question. Our next question comes from David Motemaden with Evercore ISI. David, you have the floor.

David Motemaden -- Evercore ISI -- Analyst

Hi. Thanks. Good morning. Tricia, I believe, in August, you told us that you thought auto PIF growth in the next couple of years would be difficult to match the 70% type of PIF growth you guys saw in the six years since the start of the last hard market in 2016.

Just wondering if we could get an update on your thoughts on that front, just given the challenges in the marketplace that some of your peers are seeing, some of the losses they've reported, and just how you're feeling with your rates and what you're seeing on the new apps. I'm just wondering how -- you know, do you still feel that way, just on the type of growth that you think we could see over the next five to six years?

Tricia Griffith -- Chief Executive Officer

Thanks, David. I will tell you, you know, I feel a lot more bullish than I felt about four or five months ago. We went through, you know, reacting to severity trends that, frankly, we've never seen before. I can't predict what PIF growth will be because we -- you know, there's a lot of environmental things going on.

And what I've learned over the last three years is, you know, to anticipate that there's going to be things that we haven't prepared for, whether it's the used car prices, inflation, frequency, driving behavior, pandemic. That said, I believe we're in a really great position, as we said. So, we -- you know, we said we were going to start to take rate. I think we foreshadowed that late '21.

It took some time into 2022 to obviously start to earn into the book. We do feel good, but we're always watching what's happening, and there's been a kind of an ever-evolving, changing. So, on the private passenger auto side, as you know, we took 13.5 points last year. We still have about 3 points of that to earn in this calendar year.

And just for some color, we took 1.5 points in January. So, we continue to look at trends and the future trends and the rate need -- that we need now to match rate to risk. But that said -- and John Sauerland added this on to the first question from Mike -- and that is, you know what, we're in such a good position now because we have -- there's a lot of shopping out there, a lot of prospects out there because our competitors have been now raising rates. So, they're catching up to us and sometimes exceeding us.

But we're in just a different position because we got out in front of it. So, there's a lot of shopping, a lot of ambient shopping, which is really efficient for us. And what happens, with what both Jim and Cory talked about, as well as other variables, we believe we have superior segmentation. So, we're going to grab on to as many customers and obviously keep as many customers as we can for as long as we can however this time frame and this hard market lasts.

And of course, that segmentation and the adverse selection flywheel works in our favor because of our many, many decades of investing in things like segmentation, competitive prices, our brand, etc. So, while I can't foreshadow what our PIF growth is going to be, we are going to do everything we can to follow our long-standing goal of growing as fast as we possibly can at a 96 combined ratio and making sure we can take care of our customers.

David Motemaden -- Evercore ISI -- Analyst

OK. Great. That's very helpful. And maybe just following up there, you know, it sounds like the -- you know, the continuous monitoring that's -- you know, that's, you know, been an interesting development that, you know, I'm interested to hear, you know, what sort of impact that has had on underwriting profitability and growth in the 12 states where you've rolled it out.

And how we should think about that as you roll it out on more states? You know, is that something where we can maybe see a step change in terms of profitability or frequency? Or, you know, anything from that perspective just on just greater segmentation that you might get from that.

Tricia Griffith -- Chief Executive Officer

Yeah. I mean, I think from a cost perspective, obviously, continuous is a little bit more expensive, but those costs have gone down over the years. But I think, you know, what we're excited about and what Jim talked about, the excitement about is the services that we're going to provide, especially in some of the claims examples. So, you know, we will continue to evolve like we have since 1996 with our UBI products, and we believe that will continue to enhance our ability to out-segment our competitors.

What that means from, you know, the exact amount, you know, we don't know yet, but we'll continue to monitor. And I'm sure every so often we'll update you on where we're at with our UBI offerings.

David Motemaden -- Evercore ISI -- Analyst

OK. Great. Thank you.

Tricia Griffith -- Chief Executive Officer

Thanks, David.

Operator

Please stand by for our next question. Our next question comes from Alex Scott of Goldman Sachs. Alex, you have the floor.

Alex Scott -- Goldman Sachs -- Analyst

Hi. Good morning. First question I had was just on, you know, all the potential new business, you know, just looking at the application growth and so forth. How is -- how does sort of the growth penalty from new business coming on look in this kind of environment, maybe during, you know, prior periods?

Tricia Griffith -- Chief Executive Officer

Yeah, you know, I've been thinking about that a lot because in times where we have high growth, we often talk about the new business penalty. And as I'm thinking about it, the penalties from my perspective are something that is done to you because maybe you did something wrong. As an example, Sunday, I was watching our youngest son play lacrosse. He got a penalty for slashing.

He knew it. He was emotional. He did it. He took a knee.

We're a little bit different because the things that we do and the amount that we pay is all controllable as you witnessed when we pulled back on media and did some other things to make sure we reached our 96. So, from my perspective, what we are paying -- so you'll -- you know, you saw the increase in January in our expense ratio, specifically on the direct side because we front-load our acquisition costs there. It's really an investment in our future. It's an investment for our owners, investment for our customers.

So, it's something that we're excited to do in terms of how we manage the business. And again, we have the caveats with our ceiling of an aggregate all business lines, 96 combined ratio. So, it's hard to compare, you know, from the past where we'll go. But we know that this has been a really volatile environment, and we have the levers to pull back should we need to, and throttle forward should we need to.

So I can't tell you the exact amount. I will tell you that we're pretty excited about the position we are in right now. And again, you know, we're Progressive, so we're never going to be cocky and we're always going to be really paranoid. And we follow the data maniacally and talk about it all the time.

But we are really excited about what we believe this hard market, however long it lasts, brings.

Alex Scott -- Goldman Sachs -- Analyst

Got it. That's helpful. Second question I have maybe a little more applicable to the telematics. You know, when I think through this and the pricing advantage that it gives you over some competitors that are not doing it to the same scale or maybe even just to the industry broadly, you know, I appreciate you probably don't want to quantify some of these things, but could you kind of talk to us about, you know, directionally, has -- that pricing advantage that it provides, has it expanded because of some of the things you're doing or some of these things that you're doing, just allowing, you know, sort of your pricing advantage to be maintained as others begin to do more on the telematics front but not to the same degree? Or is it sort of diminishing, is more competitors just, like, more broadly adopting telematics? I guess just high level if you could talk about that.

Tricia Griffith -- Chief Executive Officer

Yeah, I'll talk about a high level, and then Jim and Cory can weigh in for their perspective on businesses. But, you know, one, it takes a lot to get history of driving behavior and to continuously learn on that. So, I think you have to have a head start, which is why we sort of did a primer on the history of where we've been. We were out ahead of it, obviously, nearly three decades ago.

A couple of points, though. I think the -- and one that Jim made is, you know, I think it's important to make sure that we're pricing rate to risk. So, although customers are never happy when they get a surcharge, when they do and if they go somewhere else again, that's adverse selection. So, there's multiple pieces of not just the N equals one but also what can happen from a competitive perspective.

So, to be able to have the UBI as well as our other segmentation, to truly understand that rate to risk and what John talked about with our underwriting acumen is also an important piece. I think is probably more important on the commercial lines side from the perspective of if I'm a good driver and I drive a truck, can I save money on my insurance, which is a big cost in the commercial lines. So I don't know if, Jim or Cory, you want to add anything?

Jim Haas -- Usage-Based Insurance Business Leader

The personal line side, yes, there's certainly more activity in UBI than there was 10 years ago, let's say. We're just continuing to move our pricing forward, price more aggressively, take more advantage of that, regardless of what anybody else is doing. We want to price as accurately as we can and get as many people into that program as we can.

Cory Fischer -- Head, Commercial Lines, Research and Development Business

Yeah. Thanks, Alex. So, for commercial auto, we are the largest commercial auto insurer and we're also No. 1 in truck.

And I think with that, we have a lot more data than anyone else. And so, we have -- we feel like a leadership position in segmentation. I think telematics extends that leadership position relative to the marketplace. So, there are some companies that are investing in telematics.

But I think between our size, the amount of data we have, we believe we're increasing the lead through telematics.

John Sauerland -- Chief Financial Officer

I would just add. So, over the history of the personal auto evolution in telematics, you know, early on, we weren't pricing to the full curve. So, we understood that a lot of people should be receiving surcharges and some people should be receiving bigger discounts. Because we were at the forefront and we're seeking consumer adoption, we didn't price to that full curve.

I think we are now very close to or almost completely pricing to that curve with the data we have today. However, as Jim was sharing, we continue to advance the size, continue to pull additional data elements that can continue to have us leading the pack. So, our segmentation game is never ending. We're always trying to continue to advance to find additional data to use that can keep us with a gap between us and the competition.

And as Cory mentioned in the prepared comments with the slides, commercial lines is at an earlier stage, similar to where personal lines was. We still have a significant loss ratio benefit in aggregate for telematics in commercial lines, which means we have further pricing options that we can play as competitors come to market with their offering. So, I think we're in a great place from a telematics standpoint. We are not at all at the end of the journey.

We're going to continue to evolve and continue to advance our competitive advantage when it comes to pricing and telematics.

Tricia Griffith -- Chief Executive Officer

I'll add one last thing, just because I think this is a subtle nuance but really important, and Cory mentioned it a couple of times, when he's thinking about telematics in his role and his prior role in commercial lines, you know, he's talking to Jim and their team. He referred to his colleagues and learnings that they have. This is really common and very specific when you think about our strategic pillars. One is our people and our culture.

Our culture is to go back and forth, whether it's talent and rotational jobs, but understanding learnings from the first movers within the company, that is really key. And I think it's really different because many of us -- I think Lori Niederst, prior to her role when she was in CHRO, she shared many years ago the amount of depth in our leadership ranks because we think it's really important for people to know many sides of the business. So, although it's subtle and it's hard to measure the way our culture is and the way our leaders work, and the sharing of ideas and successes and opportunities is really important, especially in something like telematics.

Alex Scott -- Goldman Sachs -- Analyst

All very helpful. Thank you.

Tricia Griffith -- Chief Executive Officer

Thanks, Alex.

Operator

Stand by for our next question. Our next question comes from Andrew Kligerman with Credit Suisse. Andrew, you have the floor, please.

Andrew Kligerman -- Credit Suisse -- Analyst

Thanks a lot. You know, really helpful answer to that prior question, but I'd like to elaborate on it a bit. And just thinking about the fact that, I don't know, maybe 80%, 90% top percent plus of the insurer personal lines population utilizes Cambridge Mobile Telematics. I understand that Progressive's data and analytics is incredibly robust, the segmentation, you talked about history and data use.

But, you know, the competitors can see all of that. So, my question is, you know, is this replicatable in a very fast fashion? What makes it so hard not to replicate given that you're all using a similar platform?

Tricia Griffith -- Chief Executive Officer

Yeah, I'll let Jim talk about that. More similar platform doesn't mean we're all using the same data or the same models are using the data in the same way. So, one is it's expensive to get into this. Over the last three years, we've invested a lot, so you just can't all of a sudden turn on, understand your data early on.

It's different in urban and rural, it's different with your other segmentation variables. But Cambridge is the platform, but we're not all sharing the same data and outputting the same exact thing. Jim, do you want anything?

Jim Haas -- Usage-Based Insurance Business Leader

Sure. You're right, Cambridge Mobile is our partner. We've had a great relationship with them. They provide the data collection mechanism that gets embedded in the app.

And so, we do get that data from them. And you're right, most people use them in the industry. I think what's different is we own our own algorithm. We use that data the way we want to.

And we've learned a lot over the years about how to make the most out of that data. And I think you mentioned, like, everybody can see that. That's not actually true. Most -- in the entire industry, most algorithms for UBI are filed confidentially in the States.

So, those are not shared with all of the competitors. So, we can't see in other companies algorithms, they can't see ours. So, we think we can -- we've developed an advantage there and we can maintain that. We also have an OBD device which we offer.

Some of our customers pick, which gives us another source of data, which we think helps calibrate some of that as well.

Andrew Kligerman -- Credit Suisse -- Analyst

Very, very helpful. And, you know, that's pretty amazing. Your accident response product where you got, you know, under 12 minutes to solve for some of this stuff on site. I'm wondering if you could share with us the fact that -- you know, the impact of an LAE and severity, any loss numbers, and anything you could share there to gain some insight into how that affects your loss ratios.

Tricia Griffith -- Chief Executive Officer

Yeah, I think that's just too new. We're just really dipping into that, so we'll probably have more on that in 18 to 24 months. Again, it's something that we're interested in. We've always been, since probably 30 years ago, want to make sure that we're there for our customers in their greatest time of need.

So, our -- you know, the biggest thing that we want to do there is to help our customers get help when they need it. And if we can use the data from the impact to start the claim going and get them back into the position they were before the accident, give them that peace of mind, that's pretty important. And that could relate to a lot of great benefits in terms of retention, etc. But again, way too soon to be able to share much on that, and we'll probably have more in a few years.

Andrew Kligerman -- Credit Suisse -- Analyst

Got it. And just real quickly, you mentioned 40% increased usage in UBI. Could you tell us the percent adoption as a percent of the overall book of business that you have?

Tricia Griffith -- Chief Executive Officer

We -- from the new perspective, which is all we share externally, it's higher and direct. Jim, correct me if I'm wrong, about 15% in the agency channel? A little bit lower? OK. A little bit lower in the agency channel then more in the direct. We offer it to everyone in the direct channel coming in, and obviously, we work with over 40,000 independent agents.

Some feel more comfortable making that offering. Not all do it. But we've seen an increase since 2019, which makes sense. And it's a perfect fix.

If you believe that you drive safely and maybe you're driving less often for new business across the board as other inflationary pressures hit consumers.

John Sauerland -- Chief Financial Officer

I just had that. So, in the direct channel, this take rate is substantially higher than in the agency channel. They also offer that in the commercial space. The demand through the agency channel is very robust.

So, we're talking about, especially with Smart Haul, which is generally larger trucks that have higher average insurance costs, the benefit of enrolling in a program like this is substantial. So, agents and consumers, or business owners, in this case, are way more likely to select that as an option. And, as Cory said, it's highly predictive in that channel as well.

Operator

Please stand by for our next question. Our next question comes from Elyse Greenspan with Wells Fargo. Elyse, you have the floor.

Elyse Greenspan -- Wells Fargo Securities -- Analyst

Hi, Thanks. My first question, Tricia, goes to some of your initial comments, right. You said that, you know, future rate would be driven by loss trend. I mean, I realize that it's, you know, a very volatile environment these days.

But what are your expectations, you know, for loss trend when we think about 2023?

Tricia Griffith -- Chief Executive Officer

You know, we will just continue to watch it. That's why I wanted to give you a little bit insight in January, taking 1.5 points. What we know is that we closely monitor a lot of these trends, and I mentioned a bunch of them. It's been such a volatile several years that we look at, you know, our internal data, UBI data, macroeconomic trends.

All I can tell you is that we will be able to react quickly. And that's sort of embedded in everything that we have done in the last 85 years. So, when we start to see those -- or see those trends and experience those trends, we'll react quickly. And that's why we feel like we're in a good position because we were able to react quickly to the severity trends that we started to see in late 2021 and why we're sitting in the position -- the enviable position we are in now to be able to profitably grow.

Elyse Greenspan -- Wells Fargo Securities -- Analyst

And then, on the UBI side, I thought most of this was opted into, right, for new business customers. And I know that's where you just gave some of the percentage. If you're a current Progressive customer at renewal or just during the length of your policy, can you opt in and choose to use a Snapshot product if you, you know, want to try to lower your rate?

Tricia Griffith -- Chief Executive Officer

Yeah. And we actually did, you know. It was in, I think, 2021 maybe -- or '21. I'm not sure exactly when we did, but we actually sent out material to our current customers when trends start to change to make sure that they had that offering.

And we had a small group accept that. But, yeah, we want to be able to provide that if it makes sense for each driver. But it's normally at new business.

John Sauerland -- Chief Financial Officer

Yeah, that was when we saw COVID driving down materially and made an offer to our existing customers to save through Snapshot. And not many took us up on that offer. So, it's open and available. Customers can do it.

They just haven't taken us up on it significantly. 

Elyse Greenspan -- Wells Fargo Securities -- Analyst

And then one more quick one. So, you guys mentioned the new product is continuously monitoring. So, the states that you're active in with that Snapshot product, anyone that uses Snapshot has to be -- continuously monitor or can they opt into a different monitoring period?

Jim Haas -- Usage-Based Insurance Business Leader

So, when they opt into the program, it is -- there's only continuous. You don't get to pick at that. Obviously, we can't force them to continue to monitor forever. So, we've got -- sort of the program is designed around that and accommodates when people stop giving us data, if they do, but we certainly encourage them to continue to grow that data.

The biggest discounts are available that way, and the program's designed that would be you monitoring continuously.

Operator

Please stand by for our next question. Our next question comes From Tracy Benguigui with Barclays. Tracy, you have the floor.

Tracy Benguigui -- Barclays -- Analyst

Thank you. Even though you feel like you've approached rate adequacy in most states, I'm wondering, since you're growing quite quickly with lower ad spend, could you get away with getting even more rate? I mean, you did mention that some of your competitors are exceeding you guys with rates. Do you feel like it's fine to be cheaper than some given your risk selection?

Tricia Griffith -- Chief Executive Officer

Well, we want to be competitive and we want to price to future loss trends. So, we continue to look at that, and we will increase rates should we see severity or frequency trends change. And that's why I did give that data point in January. So, we'll stay on top of trends and we'll continue to grow.

And like John said, we haven't, you know opened up our aperture in terms of not looking at things that are important from our underwriting perspective. So, we still have more levers from that perspective. But now to -- you know, for me is all about using the data and pricing accurately as quickly as possible to match each rate to risk.

Tracy Benguigui -- Barclays -- Analyst

Got it. And could you also touch upon your initiative to raise retention among multiproduct customers? That was mentioned in your filing yesterday. What are the key tenets to achieve that?

Tricia Griffith -- Chief Executive Officer

You know, we know that our customers and I think most consumers with insurance want stable rates. And so, we want to be able to have, if possible, even in this kind of volatile environment, what we call small bites of the apple to not have to increase rates like we had to in 2022 to stay ahead of trends. We also have very many offerings. So, we have, you know, whether it is in our Progressive advantage agency where we have a stable curve of home, on affiliate partners we work with to have more and more of those bundles rates, and Robinsons.

We'll continue to do that. And we really seek, you know, long term, to be sort of the destination company. So, whether it's on our paper or not, we think of that in terms of our pillar of being where, when, and how our customers want to shop and buy, so broad coverage. So, on the personal line side, we know we sell other people's home company when it doesn't necessarily fit where we're going to go.

On the commercial line side, we also use our products as well as others for small business workers comp, GL, etc. So, it really is about following the customer and the more we're able to offer them, even with the Progressive Advantage agents to where it's -- we work with a lot of different partners ranging from life insurance to travel insurance to jewelry insurance. The more that the customers have and the more that we've earned their trust, the longer they stay. So, retention continues to be the holy grail.

We're seeing positive improvement. I will continue to report on that as the quarters unfold. But we do know two parts of retention are really about having stable rates and having great service.

Tracy Benguigui -- Barclays -- Analyst

Great. And is there a segment you're focusing more on? You know, agency, will that be more focused versus direct?

Tricia Griffith -- Chief Executive Officer

No, we focus across the board. You know, so, they're a little bit different in terms of, you know, control. So, an agent can decide if they want to shop it on behalf of the consumer. But they're looking at the same thing as consumers are, and that's making sure that our mutual customers have stable rates with a company with a great brand and great service.

But, you know, our goal is to extend PLEs across every segment individually and in the aggregate.

John Sauerland -- Chief Financial Officer

As Tricia was saying, Tracy, you know, we're continuing to broaden the offerings to our consumers. So, you talked about multiproduct households, as Tricia was implying, those households stay with us a lot longer than single-product households and much longer than, you know, our historical core customer of Sam, who we call the inconsistently insured customer. So, our book does continue to shift more toward the Robinson end of the spectrum, which is naturally the longer-retaining end of the spectrum. We have, as Tricia also noted, you know, seen a setback as we've had to raise rates aggressively.

That is coming back. Our PLEs are starting to trend positively at least on the trailing three and of the -- our metric. And the trailing 12 is not yet fully turned, but certainly, the trailing three is the leading indicator of the trailing 12. So, we are definitely focusing on more multiproduct households as we roll out new products in the personal auto space.

Increasingly, we find segmentation that can make us more competitive for those preferred customers. Our book continues to shift that way, and as we get to a more stable rating environment, I think you'll see our PLEs continue to grow as they had been before we had to take a lot of rate.

Tracy Benguigui

Thank you.

Operator

And our next question comes from Josh Shanker at Bank of America. Josh, please go ahead with your question.

Josh Shanker -- Bank of America Merrill Lynch -- Analyst

Yeah. Thank you very much. Could you please give us some insight that -- you have the policy conversion and new application numbers for the full-year 2022, but it seems like something's inflecting in the fourth quarter. Can you give us some indication what's happening fourth quarter versus fourth quarter 2021 in quote conversion and new policy apps, and maybe into '23 as well?

Tricia Griffith -- Chief Executive Officer

Yeah, we don't publish new apps on a quarterly basis, but we did want to show it for a year over year in the opening slides we had. You know, basically what happened was shopping started, the hard market started. We were in a good position, and there's a lot of prospects that began shopping. And we had, you know, great coverage at great rates and we were able to convert.

So, that's -- it's really sort of where we started to see things turn. We were in a good position for that from a staffing perspective on the sales side and the claims side. But yeah, that's really kind of when we saw the hard market churn. John, you want to add anything?

John Sauerland -- Chief Financial Officer

Sure. Joshua, we do provide year-to-date numbers through the third quarter and the fourth quarter. So, if you're just assuming equal weighting, you can come to a number. But we also in the presentation showed you actual charts of those.

So, you can't see exactly what the number is but you come pretty darn close by looking at those charts. So, the short story is conversion was up considerably in the fourth quarter. Prospects were up considerably in the fourth quarter. And obviously, new apps, as a consequence, were up a lot.

Unknown speaker

John --

Josh Shanker -- Bank of America Merrill Lynch -- Analyst

Thank you very much. And question for the commercial side. Since acquiring Protective, can you talk a little bit about your large fleet offering and how that's changed?

Tricia Griffith -- Chief Executive Officer

I can't hear it. Didn't hear the question.

John Sauerland -- Chief Financial Officer

Can you repeat the question, please?

Josh Shanker -- Bank of America Merrill Lynch -- Analyst

I'm sorry. The question was how has large fleet changed as part of your business since acquiring Protective?

Tricia Griffith -- Chief Executive Officer

Oh, thanks. You know, now substantially, yeah, I mean, we're still learning a lot and the integration is going really well. And it was really something that was additive to what we were growing from our small fleet. So, we continue to be excited about the acquisition of Protective and we'll continue.

And I met a couple of times with Mike Miller, who's running Protective. He came from the commercial lines side and is now running that, so, you know, he'll be excited to have his growth plan in play. We've been focusing in the last year or so really on integrating the company within Progressive at large.

Josh Shanker -- Bank of America Merrill Lynch -- Analyst

Thank you.

Operator

Stand by. Our next question is coming from Meyer Shields with KBW. Meyer, you have the floor.

Meyer Shields -- Keefe, Bruyette and Woods -- Analyst

Great. Two related questions. I think this was on Slide 26 with the 15% participation discount. The first is I was hoping you could talk us through what underpinned the decision to raise the discount.

I would have probably naively guessed that incremental information as less marginal utility. And the second question is all else equal, If I'm a bad driver and I get a surcharge, then I get a bigger participation discount. And how does that factor into, I guess, retention and profitability?

Tricia Griffith -- Chief Executive Officer

Yeah, I'll let Jim talk about that. I think that the larger discount is what was John was saying. We're repricing to the curve, to the full curve. And so, we know those customers are have great driving behaviors and we're going to add them.

And of course, the adverse selection is on the drivers that are not so good in the surcharges. Do you want to add anything, Jim?

Jim Haas -- Usage-Based Insurance Business Leader

Sure. Yeah, we raised the participation discount part. We're going to continuous -- we're asking customers to do more in that case, and we wanted to encourage them. And the participation discount creates a larger incentive for that.

As for how it would relate, if you ultimately got a surcharge, on the renewal term, the participation discount is only for the first term. So, it's removed when you get to the renewal. And so, you know, a bad driver is now getting a larger surcharge and they're getting a larger discount removed off the first term, which would get us priced more accurately on them in the subsequent terms, which is the goal. So, the participation discount is there to encourage people to participate.

And then, we just price as accurately as we can on the subsequent terms.

Meyer Shields -- Keefe, Bruyette and Woods -- Analyst

OK, fantastic. That's helpful. Thanks for clarifying that. Second question.

I'm thinking of California, but there may be other regions where rating flexibility is limited. It seems like the economic benefits are -- of UBI monitoring are enormous. Is there a chance of getting this incorporated into the rating algorithms for more risk [Inaudible] states?

Tricia Griffith -- Chief Executive Officer

We'd love to be in California. That's kind of a history from a lot of the proposition, one of three things that came up many, many decades ago. We want to grow in California. We talked about that a lot as the most populous state.

But in the meantime, we will continue to work with the department to get the rates that we need to make sure that we reach our target profit margins.

Meyer Shields -- Keefe, Bruyette and Woods -- Analyst

OK. Thank you.

Tricia Griffith -- Chief Executive Officer

Thanks.

Doug Constantine

Next question?

Operator

Our next question comes from Paul Newsome with Piper Sandler. Paul, go ahead with your question.

Paul Newsome -- Piper Sandler -- Analyst

Thanks for the call. I was hoping you could talk a little bit about if the telematics, particularly the continuous use telematics, are allowing you to effectively change prices faster if the environment changes without having to actually file a rate change. It sounds like that's possible, but I don't know enough about how the formulas work for telematics products to know if you can effectively react to environmental changes through the telematics product without actually changing the filings.

Tricia Griffith -- Chief Executive Officer

Good question, Paul. And I'll let Jim take that.

Jim Haas -- Usage-Based Insurance Business Leader

Paul, you're right, if a customer is continuously monitoring and their driving changes dramatically, like, because of a pandemic or something like that, the telematics reading on the renewal would pick that up over time. In terms of the effect of that, you have to just remember we just started to roll out continuous last year for the first six months. It looks just like it did before. It's not until the renewal that it changes.

A lot of the states we elevated, we're toward the back half of the year, so they're just coming up on their first renewal, so. And then you need to build up, you know, a bigger -- that will grow over time to be a bigger share of the in-force book. But right now, it's still early days of that. So, it'll be a little while before that's a really effective mechanism for that.

Plus, we would need to get right on all people learning UBI who do that through traditional means.

Paul Newsome -- Piper Sandler -- Analyst

Right. Can you talk about how you can -- how quickly you can change the algorithm as well? Is that something that means it's something close to a filing or you effectively just change the algorithm for both the continuous and noncontinuous if you see an environmental change?

Jim Haas -- Usage-Based Insurance Business Leader

Typically, that's through a filing. You know, we file the algorithm with the states confidentially, and, you know, that's typically how that would be done.

Paul Newsome -- Piper Sandler -- Analyst

Great. Thank you. That was my questions. These are wonderful calls.

Thank you very much.

Tricia Griffith -- Chief Executive Officer

Thanks, Paul.

Operator

And standby for our next question. Our next question comes from Gregory Peters with Raymond James.

Greg Peters -- Raymond James -- Analyst

Hi, everyone. Most of my telematics questions have been answered. And, Tricia, in one of your comments, you talked about media spend. I know you guys go through an annual budget.

Can you talk to us about your perspective on your advertising media budget for '23 versus '22?

Tricia Griffith -- Chief Executive Officer

Yeah. So, I mean, we absolutely do a budget, but then, ongoing, we assess it based on the efficiency of the spend. We look at our targeted acquisition costs, our cost per sale, and then kind of where we're getting the business from. So, I talked a little bit about ambient shopping, which, of course, is really efficient.

We're in a little bit more of that position now. But what we will do is, as the year progresses, we will continually assess what we need to do to spend more if we believe it's efficient or pull back, if we believe we're butting up against our 96 ceiling. And those are conversations that happen continuously. Some things we -- you know, to get sort of advanced discounts, we pay for, you know, up front.

So, those are kind of the spend. But we have a lot of flexibility, especially in the digital side, to bolt back should we need to, like you saw it in the last year or so. You know, we have a budget. It's fluid, and, over the years, it can go up or down depending on what we believe is in the best interest of profitably growing the company.

Greg Peters -- Raymond James -- Analyst

Does that answer mean that the budget is higher for '23 versus '22? I'm not sure what it means.

Tricia Griffith -- Chief Executive Officer

You know, at this point in time, it could be higher. I mean, things change. So, it was higher last year. If you would have asked me this last January, I said it would be higher than the year before, but it didn't end up being that because we had to make some flexibility.

So, you know, we budget for what we think can happen with a lot of flexibility.

John Sauerland -- Chief Financial Officer

A couple adds there. So, it's -- I would characterize it as dynamic and opportunistic. So, to the extent we feel we're adequately priced, we're going to spend a lot more on advertising again with, as Tricia said, an eye toward our targeted acquisition costs that we priced into the product and our lifetime economics on those policies. But if you look at our expense ratio, especially our direct expense ratio for January as an example, you can see it was up.

So, clearly, through the latter part of last year, as we were working hard to ensure we hit our 96 combined ratio for the year, you saw that expense ratio drop, which was driven, to a large degree, by lower advertising costs. In January, you saw that pop back up, which is a great indicator that we're spending more on advertising right now.

Greg Peters -- Raymond James -- Analyst

OK. My second question was on reinsurance. The market's hard that you've had -- you know, obviously, reinsurance is an important part of your story last year. Can you talk to us about your perspective on reinsurance and how it might change this year versus last year?

Tricia Griffith -- Chief Executive Officer

Yeah, absolutely. We have Brandon here who runs our reinsurance organization, so I'll have him weigh in.

Brandon Hopkins -- Risk and Reinsurance Business Leader

Hi. Thanks for the question on reinsurance. I think we, particularly with how we adapted to the placement at 1/1, will carry that forward to how we look toward June placement, which is really the core part of our program. We tend to be fairly conservative in how we want to retain property cat losses.

I think we'll continue that. We'll have to be flexible in what the market is willing to offer as far as how we can tailor that to what we decide to purchase.

Greg Peters -- Raymond James -- Analyst

Does that mean you're going to have a higher deductible or a lower deductible?

Brandon Hopkins -- Risk and Reinsurance Business Leader

It's a little early to say exactly where we're going to land on that. Certainly, the bottom end of reinsurance programs have seen a lot of pressure, not just this year, but over the last couple of years. We will be looking at other creative options, if need be, to manage our retention.

Greg Peters -- Raymond James -- Analyst

Got it. Thank you for the answers.

Operator

Stand by for our next question. Our next question comes from Michael Ward with Citi. Michael, go ahead with your question.

Mike Ward -- Citi -- Analyst

Hi. Thank you, guys. Good morning. I do have a question on telematics, specifically the OEM data in vehicles.

Just curious if you have any data on the capabilities of the existing vehicles on the road in the U.S. It just -- it seems like that might be limited to more newer vehicles with cellular data connections, which I get the sense might be -- at least currently might be a fairly small chunk of the vehicle population in the U.S. just given turnover. So, just wondering if you can elaborate on that at all.

Jim Haas -- Usage-Based Insurance Business Leader

Yeah, you're correct. The framework we're talking about depends upon vehicles with cellular connection. So, we were referring to OEMs over the last several years have been investing in that. They're in really different places depending on the OEM.

Some are -- have had in there for a number of years. Some, you know, are still working on it. So, it's -- and it does take a while for the fleet to turn over. So, it's very focused on a few OEMs and the most recent model years.

But as we mentioned there, we've got relationships with GM and Toyota, which are two of the larger ones, They've got more of an installed base. And we expect other -- almost every OEMs moving in this direction. So, we expect that this population will grow over time.

Mike Ward -- Citi -- Analyst

OK, great. And then, maybe could you discuss kind of the take-up between OEM data versus using the mobile phone app route? It just seems like there's more hurdles involved in the mobile app and that maybe that's why take-up wasn't great initially. And so, sort of curious about how you see those two trending.

Jim Haas -- Usage-Based Insurance Business Leader

I mean, adoption in the whole program overall has been up, and we're happy with the adoption, both mobile and OEM. So, we're happy with both of those programs. We just offer different options to consumers depending on how they want to share that data. And so, we're just trying to get the biggest footprint we can, and both have been increasing.

Mike Ward -- Citi -- Analyst

OK. So, maybe just elaborate on that. I think you guys mentioned that the take-up or the adoption wasn't great for existing customers. And so, I'm just curious what -- why might it be great for new customers? Then, is part of it just the competitors raising rates kind of earlier, the shopping's up?

Tricia Griffith -- Chief Executive Officer

Yeah, I would just say, you know, it's one of the things that new business when you're asking for a bunch of data to kind of assess the rates risk. When we -- when I talked about the current customers, we were doing that in 2020 because we knew people were struggling with their finances,. Everything was going up from groceries to etc. So, we did sort of a marketing campaign and just most people didn't do it.

I think you get sort of, you know, just like anything, whether it's, you know, life insurance or health insurance to sort of make that decision and then you stick with it. It's something you don't necessarily want to do again. So, it's just easier to do it in new business when we're acquiring all the information we need for an accurate rate.

Mike Ward -- Citi -- Analyst

Thank you.

Tricia Griffith -- Chief Executive Officer

Thanks.

Operator

Our next question comes from Ryan Tunis with Autonomous. Ryan, go ahead with your question.

Ryan Tunis -- Autonomous Research -- Analyst

Hey, thanks. Good morning. I had a couple of them on just the UBI. So, the first one, just thinking about the collision detection.

I'm curious if maybe that could potentially have some kind of impact on frequency, because I would think that there'd be a lot of customers with kind of smaller dollar claims that wouldn't want to tell their insurance company they got in the crash and just to keep their rates where they are. So, is there any possibility that, you know, you knowing that they gotten the crash could potentially lead to you actually getting more claims in the door?

Jim Haas -- Usage-Based Insurance Business Leader

Yeah, thank you, Ryan. Good question. In this particular case, the way the crash section algorithm works, it's looking really for more severe accidents. It's not trying to pick up fender benders.

So, these are claims that we fully expect would have gotten reported either way. These aren't small dollar claims where somebody might just choose either not to repair or to handle out of pocket. These are larger claims. So, we don't think of it.

We think we've gotten these claims anyway.

Ryan Tunis -- Autonomous Research -- Analyst

Got it. Got it. And then, yeah, I guess just thinking about some of the telematics. So, one question I had was, you know, for the drivers that you find giving really big discounts to be like 40, what -- if you look at like kind of the traditional insurance rating variables you use that are non-telematics based, ZIP code, FICO, whatever, can you draw any conclusions about, like, when you're giving really, really big discounts to customers? Which of the variables tend to have been pretty shakiest predictive indicators? It's the first part? And then, something else I'm just curious about is when you find yourself giving a surcharge for the continuous -- you know, like, six months later, you gave the discount, now you're giving a surcharge, what are the main reasons that you're finding that you're having to give a surcharge after giving a discount?

Jim Haas -- Usage-Based Insurance Business Leader

So, on the first part of your question in terms of how it -- you know, what are the traditional variables and what informs, the whole point of UBI, that's incremental to all of that. So, if we could have seen that these variables without UBI, you know, were inaccurate, we would have fixed it. UBI gives us new information about that customer that the traditional variables tend to let you segment further, not just live in this ZIP code or they have this credit score, but this person within that is different than the other person that otherwise looks the same. So, that's the point: It's improving the segmentation.

In terms of a surcharge, the discount we give is a participation discount, which is just for signing up. And so, if they move from that discount surcharge simply because we've then gotten that driving data in continuous model, theoretically, people could move from one to the other. But in terms of why someone would get a surcharge, our model, you know, we kind of share on the website, it uses things like hard braking, how much you drive, when you drive, variables like that. So, folks who tend to get a surcharge, you know, drive more and drive -- have more hard brakes.

This is a sign of how aggressively they're driving. Those are typically some of the variables that figure into that.

Ryan Tunis -- Autonomous Research -- Analyst

Got it. And just, I guess, following up. So, I hear you, it's incremental. But what about just in terms of conclusions, is it usually -- can you say anything about, you know, those who get 40% tend to largely be much more preferred, or those who get to be kind of in the 40% discount tend to be more mixed or tend to be a little bit more on -- a little bit less standard?

Jim Haas -- Usage-Based Insurance Business Leader

I think the UBI is kind of showing it's incremental to all of that. So, they -- intentionally, we don't want it to look that way. We want it to be -- if we knew the preferred customers were more likely to get the discount, we'd just give them a bigger discount in the first place. So, we try to use UBI to kind of sort out this preferred customer deserves an even bigger discount, this other preferred customer deserves maybe a surcharge.

So, it's really -- it's segmenting beyond what we already have.

Ryan Tunis -- Autonomous Research -- Analyst

Thank you.

Doug Constantine

We've exhausted our scheduled time, and so, that concludes our event. Stacey, I will hand the call back over to you for closing scripts.

Operator

That concludes the Progressive Corporation's fourth-quarter investor event. Information about a replay of the event will be available on the Investor Relations section of Progressive's website for the next year. You may now disconnect.

Duration: 0 minutes

Call participants:

Doug Constantine

Tricia Griffith -- Chief Executive Officer

Jim Haas -- Usage-Based Insurance Business Leader

Cory Fischer -- Head, Commercial Lines, Research and Development Business

Mike Zaremski -- BMO Capital Markets -- Analyst

John Sauerland -- Chief Financial Officer

David Motemaden -- Evercore ISI -- Analyst

Alex Scott -- Goldman Sachs -- Analyst

Andrew Kligerman -- Credit Suisse -- Analyst

Elyse Greenspan -- Wells Fargo Securities -- Analyst

Tracy Benguigui -- Barclays -- Analyst

racy Benguigui

Josh Shanker -- Bank of America Merrill Lynch -- Analyst

Unknown speaker

Meyer Shields -- Keefe, Bruyette and Woods -- Analyst

Paul Newsome -- Piper Sandler -- Analyst

Greg Peters -- Raymond James -- Analyst

Brandon Hopkins -- Risk and Reinsurance Business Leader

Mike Ward -- Citi -- Analyst

Ryan Tunis -- Autonomous Research -- Analyst

More PGR analysis

All earnings call transcripts