Welcome to the Angi Inc. Fourth Quarter 2025 Earnings Call. All participants will be in a standby mode. Should you need assistance, please signal a conference specialist by pressing the star key followed by zero. After introductory remarks, there will be an opportunity to ask questions. To ask a question, you may press star then one on your telephone keypad, and to withdraw your question, please press star then two. Please note today's event is being recorded. I would now like to turn the conference over to Andrew Russakoff, Chief Financial Officer. Please go ahead. Morning, everyone. Russi here, CFO of Angi Inc., and welcome to the Angi Inc. fourth quarter earnings call. Joining me today is Jeff Kip, CEO of Angi. Angi has also published a shareholder letter, which is currently available on the investor relations section of Angi's website. We will not be reading the shareholder letter on this call. I'll soon pass it over to Jeff for a few introductory remarks and then open it up to Q&A. Before we get to that, I'd like to remind you that during this presentation, we may make certain statements that are considered forward-looking under the federal securities laws. These forward-looking statements may include statements related to our outlook, strategy, and future performance, and are based on our current expectations and on information currently available to us. Actual outcomes and results may differ materially from the future results expressed or implied in these statements due to a number of risks and uncertainties, including those contained in our most recent quarterly report on Form 10-Q, our most recent annual report on Form 10-K, and in the subsequent reports that we file with the SEC. The information provided on this conference call should be considered in light of such risks. We'll also discuss certain non-GAAP measures, which, as a reminder, include Adjusted EBITDA, which we'll refer to today as EBITDA for simplicity during the call. I'll also refer you to our earnings release, shareholder letter, our public filings with the SEC, and again to the investor relations section of our website for all comparable GAAP measures and full reconciliations for all material non-GAAP measures. Now I'll pass it off to Jeff. Thanks, Russi. Thanks, everyone, for joining. I'd just like to start. We are fairly happy with where we've gotten to right now. Over the last three years, we've given up about $500 million of lower quality revenue. But at the same time, we've doubled our EBITDA and cut our capital expenditures in half, meaning we've swung from real negative free cash flow to real positive free cash flow. At the same time, we've moved our homeowner NPS more than 30 points. We've cut our churn by more than 30%. We've improved our customer success rates more than 20%. And actually, in the fourth quarter, we've turned our customer repeat rate positive about 10%. So we're pretty happy with the progress we've made. We're making a material stairstep improvement in our year-over-year revenue changes, probably 700-900 basis points. Actually, in January, we grew very modestly, although year-on-year, we don't fully expect growth in the first quarter. But we're pretty happy with where we are. We're very optimistic. On top of that, we've reset our margins. We've cleared the capital to invest in long-term profitable growth. And we're very excited about our prospects in the AI landscape. I'm going to talk a little bit about that. I think there's a few things to talk about. I think we should talk about LLMs, marketplaces, software, and agentic coding, different layers of emphasis there. First of all, when we look at LLMs, we see it as great opportunity. We're very happy to see LLMs enter and be places where homeowners and consumers generally, who have lower knowledge and maybe less frequent interaction, go to discover and explore. We've been very successful building an acquisition on Google, which is effectively the predecessor of the LLMs. Google obviously has its own LLM, where homeowners and customers go to explore, research, and discover. We've been very effective because we have built a network, a deep, broad, and skilled network, which Google still finds very useful as a partner to serve its customers. And we believe the LLMs will as well. We have started actively working with every LLM. We have had conversations and earned effective dialogue. We've announced a deal with Amazon's Alexa. We have an app submitted to another major LLM. We're talking live about two technical integrations based on the same technology we built for the app we submitted. And we feel very good about the opportunity there. We think that it's harder for LLMs to go out and build the deep and engaged customer base that we have. Again, we were able to do it and maintain it and sustain it all the time while Google tried to do the same. So we feel pretty good about our competitive position. We think we can serve as excellent partners to LLMs. In fact, we've deployed LLM technology in our SR path, in the core customer experience, which we are training with our own proprietary data and experience to make sure that we can land the homeowner in a better match. About 35% of our homeowners touch that part of our technology and experience. They convert about 3.3 times as well to a pro selection as the customers that don't. And so as we've trained that technology, we think that's positioned us better to interact with the LLMs. Our approach with the LLMs is that we can pick up the context and the conversation that the homeowner is having with the LLM at the beginning. When Russi says, "I have water on my floor, what do I do?" Or the LLM can have the full discovery with Russi and get to the point where we say, "Okay, we think there's a leak at the base of your toilet. We need a plumber." And we can take that information and bring the right plumbers. So we think we can do that effectively. Obviously, pros have separate marketing channels than we do. They go direct to Google. They do a number of things. We actually think longer term, we can help them there because we think we're probably at scale the best there is at finding homeowners who need help from pros. But we think that we will still exist and be able to grow in this environment. And we're just very excited to have competition at the top of the funnel and be able to diversify our channels. Let's just talk briefly about then our role as a marketplace, some of the things that are being said about software out there and agentic coding. Fundamentally, let's focus on Angi as a marketplace. We are an agent. We have customers on one side. We have data and systems of record on the other side. We are effectively the execution layer and the UI layer in between. And we get the homeowner's job done, which is finding a pro who can do their home job well. And we get the pro's job done, which is finding a homeowner whose job they can do well. And we act as an agent. We believe that using agents will allow us to be even more effective at what we do and, again, use our proprietary data and systems of record and experience to be stronger and faster at development here, in addition to the existing network and customers and the resulting network effects that we already have. A competitor may be able to build an alternate marketplace technology, metaphorically, overnight in their garage now, but they cannot build our network nor our homeowner reach or our brand. So we think we're very well positioned. We think further that when you think about software, we think now we have the ability to extend our agents and actually integrate with all of the software out there that our customers use better. For example, we believe that we can act as the post-lead communication between the pro and the homeowner to clarify things for the pro, to book the appointment into the pro calendar, and perhaps even book the appointment into the homeowner calendar, send follow-ups, etc. We believe we can ultimately integrate also with ERPs and HR systems and anything else that helps the pro move through the chain to get the job done. And so we believe we're well positioned to actually extend our mission. Today, if five homeowners come to us with a job, three of them hire a pro, which is not that different than what we study when homeowners call a pro. We get to something more like like seven out of 10 hire a pro once they've made a phone call. But so that's pretty good. Of those three, only one hires our pro. We believe that by using agents, we can drive that up to two and then towards three, which will dramatically improve the value created, the retention, the repeat, and our ability to extend the marketplace. So we're actually very excited about this. And then the final piece, I'll just briefly state, obviously, there's a lot changed, even in the last week or so, with Agentic Coding and what's being written there and the possibilities. We're extremely excited here. Again, we can build something in our metaphorical garage over the weekend. We think this gives us great opportunities to extend our software by using agents and invest and regrow our whole network and business. So overall, we're very excited about the entire landscape. Let's talk a little bit about now. I'll talk a little bit about our business and revenue trajectory, and then I'm going to let Russi talk a little bit about margins, and then we'll take questions. First, I'd say we're roughly in the same place we were before, maybe modestly lower. Previously, we were talking about getting to a little bit of growth in the first quarter and getting to mid-single digits for the year. I think now we're looking at very modest negative growth, but still a material sequential acceleration in the first quarter and maybe low single digits for the year. What's the difference? The difference is, obviously, we had pressure. We discussed it on our last call from both Google SEO and our network channel in the third and fourth quarters. Between our November call and now, we think that that pressure is extended. So we have gotten more conservative on those channels in the year. What we've historically been able to do is take actions, do work on our product, etc., and actually change the trajectory of these channels. If you look at just Google SEO, we were down 35%-40% year-over-year in mid-2024. We brought that into the double digits, mid-double digits by the end of the year. We expected to continue that trend. We were metaphorically punched in the mouth again in the spring and fell to the range of 35%-40% again. Then we sequentially improved into the low- to mid-20s by mid-year. Then we got hit again in the late summer. Thus, we were where we were going into the year. What we've done is we've essentially said we don't think we're going to make that progress back again. We're going to assume Google SEO stays down at that lower level for the year. We're doing something similar with our network channel, where we basically have assumed we're not going to improve it in the rest of the year. We're just going to kind of get to the second half of the year and stay at that lower level we were in the second half of last year. So we've effectively gotten conservative. We think it's more prudent to look at our full-year revenue that way. Really, our focus is on our proprietary business, which, again, we grew 17% in 2025. We're expecting high single, low double digits in the first quarter there. We believe that that business can be a solid mid-single-digit plus, ideally double-digit grower long term. We've put a great deal of investment there. We've executed very well. Frankly, we've seen our repeat growth turn in the fourth quarter. We actually think that the high-quality branded traffic is coming back. With all of our improvements in the customer experience and what we see in customer behavior, we think it's time to lean back in to branded advertising, where we're running TV and streaming and social. We've done this effectively for years. We're basically going to return from the lower level we were at in 2025 to the level we were at in 2024, which is an effective level for us to spend at. We think we can do it well. Just talking about the quarters briefly, and then I'll hand over to Russi. In the first quarter, compares get more difficult, February, March. In our proprietary channels, we ramped two areas of Google last year, first in February, March, and then April, May, which was Google Display and then Google Search Partners. We got some effective revenue growth. But as we watch that traffic season, we actually saw lower win rates than the rest of our channels. And we effectively scaled them both down. That makes the second quarter in particular difficult compare and a little bit more difficult compare in February, March. So we expect the first quarter with the kind of 60-ish% network decline baked in to come in at -1% to -3%. We expect the second quarter to come in at flat, maybe a little bit down. And then we expect to get to mid-single-digit in the second half of the year as the network channel stabilizes, flattens out, and we're able to grow our proprietary and effective long-term rate. We're optimistic we can do better. But that is prudently where we want to guide right now. Again, looking out over the course of the year, we basically think low single digits, let's call it 1-3, that's impacted by a few hundred basis points worse of Google SEO and network outlook. It's impacted to the positive side by our brand spend. And then in the first quarter, again, there's a little bit of delay in the product roadmap that came with a RIF. Sometimes you have to make a short-term sacrifice for the long-term good of the business. The first quarter is going to be a bit negative at -1% to -3%. Again, I think we're overall very pleased. It's not quite as high as we wanted it to be. Again, 700-900 basis points of acceleration from Q4 to Q1, focused on growth in this year and very strong performance overall in our proprietary channels. With that, I will let Russi just talk about our margins and our EBITDA progressions. Great. So starting with Q1, as Jeff mentioned, we're going to be deploying dollars for offline marketing, which we had in Q1 of last year, we had pulled back on and spent virtually nothing as we were shifting to Homeowner Choice at that period of time. So that includes increasing our spend in the U.S. It also includes some spend internationally, where historically, TV has worked well in Europe in Q1. We had backed off kind of during COVID and after COVID. And now we're reinvesting back behind the brands. And it also includes $3 million of new creative. We've also begun to ramp up online pro marketing. All of that will drive revenue and profit, but on a lag with only part of that returning in the quarter. So quarter-over-quarter versus the fourth quarter, our sales and marketing goes up by about eight points as a percent of revenue. Then revenue increases seasonally as you get into Q2 and Q3, with some benefit as well from the Q1 spend flowing into the future quarters. Directionally, we should add $35-$40 million of incremental revenue into Q2 versus Q1, where we'd expect also to spend kind of $10-$12 million more in marketing to acquire the extra SRs. But we won't have any additional creative to expense in the second quarter. Both pro acquisition and fixed costs will be directionally flat on a dollar basis, but better on a percentage basis as the higher revenue comes in in Q2 and Q3. So together, that will deliver incremental EBITDA in the kind of mid-$20 million range in Q2 versus Q1 EBITDA and gets you to overall EBITDA in the mid-40s for both Q2 and Q3. Then as we go from Q3 to Q4, if we look at last year, our revenue declined seasonally by about $25 million quarter over quarter, if we assume a similar dynamic this year and roughly kind of 50% margin flow-through. And on top of that, we typically expect to pull back on offline marketing during the holidays, about $5-$10 million. That directionally gets us back to low $40 million range for adjusted EBITDA in the fourth quarter. Next, I wanted to give a little bit more context as well about the restructuring and how the savings flow through in the context of our overall guide for the year. So the way to think about the restructuring at a high level is that the objectives were threefold. So one, get the cost structure in the right place. Two, create room to make the meaningful investments we're talking about. While three, also delivering profit growth on a year-over-year basis. So the way to think about the $70-$80 million of savings then is that first, it's on an annualized basis. So that results in in-year savings in the mid-60s, with $25 million of that as cap labor. And the right reference point for that is what our total cost base was going to be for the year. So if you look at 2025, we had $223 million of fixed OpEx plus $60 million of CapEx. That gets you a total cash fixed cost basis of $283 million, which is how we kind of view our cost structure. Now, prior to the restructuring, our exit rate finishing the year would have had our fixed cost base increased by roughly $20 million year-over-year. Post the restructuring, we now expect that number to be approximately $40 million lower year-over-year, which means $60 million in total of reduction off of the pace. That allowed us to free up capital now for long-term ROI-positive investment and growth, while still delivering the $10-$15 million of profit growth year-over-year we've guided to in terms of higher Adjusted EBITDA and lower capitalized wages. The key investment areas are, as Jeff said, first, the brand marketing. It's an area where we took our foot off the gas in 2025. We pulled back pretty significantly from our historical trend levels. We are leaning in now with all the positive trends in the customer experience. Second, the online pro marketing, which will be LTV positive as we drive growth in the new pros acquired, but P&L negative in-year as we ramp that up. And then third, sales in our large pro segment, where we already have a nice business, but we're underindexed against the industry at large. And we have a big opportunity to grow there. So these investments are key pillars that unlock incremental revenue growth, which will offset the trends in network and SEO traffic that we've been discussing for the past few quarters and which we're forecasting conservatively, as Jeff mentioned, so that there are no expectations of turnaround in these channels embedded in our guidance of low single-digit overall revenue growth for the year. If you fast forward to the end of the year, we'll be a mid-single-digit grower with proprietary growth higher than that and comprising over 90% of the business, accelerating, and now with better cost leverage than 2025 so that the top-line growth can have more financial impact over the medium term. All right. So with that, why don't we open up the queue, and we can open up for Q&A. Yes, sir. To ask a question, you may press star then one on your telephone keypad. If you're using a speakerphone, we ask that you please pick up your handset before pressing the keys. To adjust your question, please press star then two. Our first question today comes from Eric Sheridan at Goldman Sachs. Please go ahead. Thanks so much for taking the question. And thanks for all the details in the shareholder letter and the prepared remarks. Just coming back to the broader discussion about AI, maybe two if I can. First, curious how we should be thinking about the rollout of AI features, as you discussed, on the customer side of the platform looking out over the next 12 months and how that gives you some visibility or confidence interval in some of what you're talking about with respect to a return to growth on the platform more generally. And the second would be, how does owning consolidated supply-side data sort of position you relative to what you want to accomplish when partnering with LLMs? Curious on that. Thanks so much. So I'd say on the first question, the main area where we put focus on AI in the customer path today is the AI helper in our SR path. What we are doing with that is we're continuing to experiment with how we have more homeowners use it effectively because what we're interested in doing is driving up the number of homeowners who connect with the right pro. Again, 35% of our homeowners currently do it and are 3.3 times as likely to actually choose a pro in our UX and UI. We'd love to drive that to 50% and 60% and 65%. And we're currently actively running tests. We recently ran a test that picked up about 5%. And so we're pleased with that. And we're going to continue developing there. We are looking at other applications such as what I referenced with post-lead communication, which we think can, again, help stabilize and get the homeowner to actually meet the pro and move towards a job done well. And we're looking at how we might apply it in other areas of the product as well. That is, as I said, sort of a backdrop to integrating with LLMs. The more effective we are there, we're working with a white-label LLM on our platform. And the more effective we are there, the more trained our data and our AI implementation is when we interact with the context that comes to us from an LLM that a homeowner has entered there. And your second question? What was the second question? About the supply side. Okay, the supply side. Sorry. Again, the way we look at the world is you have customers. You have agents, which have generally historically been human agents or software algorithms with, again, UX and UI in between. And you have a system of record or a data layer. The system of record or the data layer is what allows you to perform the agentic task well. If I have the data on the customers, I can be far more effective as an agent on the customer's behalf. If I'm a human agent and I don't know my customer, I can't really deliver my product or service well without understanding the customer. So we fundamentally already have a system of record about customer behavior, success, etc., where we can understand our customers. And so when we go and we take our pro customers and actually our broad homeowner experience, so when we go to an LLM and we see a set of context or searches or queries come in, we can take that and compare it to our customer data, effectively run it through algorithms, use an agent, and make the connection better than if we didn't have that. So that's a reasonable moat we have at the scale we operate at for the number of years we've operated at. And we think it puts us in a very good position to effectively partner with the LLMs in the same way that we've effectively partnered with Google by taking clicks with some context from Google and matching the homeowner successfully on our platform. It's worked well for them in terms of monetization. It's worked well for us. It's ultimately working better and better in terms of the customer experience. Great. Thank you. Thank you. And our next question today comes from Sergio Segura with KeyBanc. Please go ahead. Hey, guys. Good morning. Thanks for taking the questions. I had two. First, just hoping you can explain the rationale for tripling the brand spend this year and why it's the right timing to do that now and what kind of lag we should expect before that spend translates into incremental service requests. That's question number one. And then question number two is just on the proprietary channel, as you lap homeowners' choice this year, how should we think about the normalized growth rate for that channel? Thank you. Thanks, Sergio. It's Russi. So first, on the brand spend, if you put it into the context of what this company has spent on offline marketing over its history, we're really now just going to be, in 2026, returning to 2024 levels. So it's not last year, we took a step back as we were digesting the changes from Homeowner Choice. But we're not stepping we're not increasing to levels that are above anything where we've spent profitably in the past. I can talk a little bit about our approach and how we get confident with our ROI. So in TV, in particular, we have a data partner that is connected through on a decent percentage of TV sets across America. And we're able to actually pair the IP addresses of people when they see our ads and pair that against the IP addresses of people who submit service requests. So we have pretty good visibility into kind of the uplift from our TV spend. It's not as precise as other digital channels, but we've honed this over a couple of years. And we have pretty good visibility relatively to be able to measure the ROI from our TV spend. And then kind of at the back half of last year, when we were spending TV but at lower levels, we kind of dialed in, changed our strategy a little bit, and our channel and station day-part mixes. And so we're getting it to pretty good ROIs on that spend. So between that, the results, our ability to measure this, and having the strongest brand in the industry, we feel pretty confident that we can spend at these levels and be profitable. Yeah, I would just add it takes a little while to build. So your first-quarter incremental spend is going to pay back the least well, but it's going to ultimately pay back long-term. There's a tail of months on this stuff. And as we add, the incremental is taking a little more to pay back. So we're going to pay back, I don't know, three-quarters in year with a tail outside of the year. But the other point I want to make is we kind of went on defense last year. We made a material change to the UX. We were working on correcting our customer experience. We wanted to ride through that in the first quarter. And then we wanted to deliver our target Adjusted EBITDA last year, which we did. And so we pulled back our marketing spend to sort of make sure we would do all that. I think with the way our customer experience has moved and with the upside we now have with not only Homeowner Choice in, but we've rewritten most of the questions in our Q&A. We've implemented the AI helper in the Q&A. We've moved all our pros into a product where they can choose task and zip. Our new pros who are coming in online are looking at the jobs before they opt into them. We have a multiplier effect on the level of matching. And we believe we should go back on offense. Going back on offense just means going back to the 2024 spend where over time, we believe we were better than break-even, although, again, there is a tail on that. So we do feel pretty good about it. And it is baked into our overall revenue growth. All right. And then your second question, Sergio, is about kind of normalized growth rate for proprietary. So we're splitting out, and we're showing you our proprietary and network revenue on a revenue basis now. So Q4 was 23%. And for the full year of 2025, it was 17%. So that's good visibility into kind of the two pieces of our business. And for 2025, you have a grower like that, and you have a decliner on the network side that ends up combining to be a decliner overall. But as we kind of progress forward on the proprietary side, we're saying overall revenue ending the year kind of in the mid-single-digit range, probably be high single digits, that means for the proprietary business. Going forward, proprietary revenue, we think, is high single digits, continues there, or even low double digits, depending on where we can get with pro capacity and continuing to make progress in our paid proprietary channels and the impact of the branded marketing. Okay. Thank you. You can go to the next question. Sure. Our next question comes from Dan Kurnos with StoneX. Please go ahead. Great. Thanks. Good morning. Rusty, you actually just brought up the first question I had, which is since you guys are leaning back in now and we're starting to see this advancement in proprietary SRs, just curious since the network is still declining nominally, what is happening with pro capacity? And then secondarily, you guys flagged on the last earnings call and in the shareholder letter that you're doing a global platform consolidation. So maybe, Jeff, just give us an update where you are on that front, any disruptions that we might expect to see? I think you called out one in your prepared remarks, but just curious if there's anything else we should be expecting on that front? Thank you. Let me take the second question first. By cutting the organization by 40%, we think we've extended by a quarter to the timeline in getting to our final single platform. But we have built our timeline in a way that we do not believe there will be a disruption to the business. And what, in fact, we are doing is we are going to deliver in stages. So first up on the rebuild is our new homeowner experience, our current homeowner experience, which comes from the start of what we call the SR path where the homeowner enters the funnel and starts answering questions, choosing a pro, and then managing their post-selection project in a projects page. That's the core homeowner experience. That's the first thing we're going to get rebuilt. It's rigid technology. It's old. It's very difficult to iterate quickly and improve. And we are rebuilding in what we would call a componentized way. What the componentized and more flexible way is, is we can skip steps. We can pick up from different channels at different stages of the flow. If, for example, we had somebody in a hardware store, and we had a QR code, and they were looking specifically at a Mini-split to do heating and cooling in their addition, we could pick up right there that they're in the Mini-split aisle and be very clear, very quickly on where to drop them in the experience, which would boost conversion. It would boost matching. We cannot do that today. Somebody who scans a QR code with a Mini-split in front of them has to start with, "What's your zip code? What's your category?" So it's going to enable a bunch of things, including making even better any integrations with LLMs and other partners. That is the first thing we're going to deliver. Then we're going to move on to delivering the pro experience and so forth. Now, we could change order. I think we're currently looking at software we might build with Agentic Coding and how that's going to work. But that being said, we don't anticipate disruption to the business. We actually anticipate enhancing the business and the customer experience as we go. And maybe we're a quarter or two later than we initially anticipated. But there's a lot of green left to cover there. I'll cover pro capacity. So over the past couple of years, but in particular, last year, we've completely changed the way that we acquire pros and organize our sales force, especially with the single pro initiative where we're selling a single product or a single sales force on a single platform. We've changed up the prospect mix and the kind of offer strategy. So we're selling much bigger pros with bigger packages but less pros. So our nominal amount of average monthly active pros is still down year-over-year. But the capacity per pro is up. Revenue per pro is up. And the overall capacity of the network is actually up when you net those two factors against each other. Now, the complexion of that will change a little bit next year as we lean into selling more large pros. And we're talking large pros. We're talking quite large pros. So those will be fewer in number but much, much bigger. So they'll have less of an impact on our kind of nominal counts, but they'll drive a lot of capacity. And then on the completely flip side, as we ramp up online enrolled, we'd expect those to be kind of lower capacity, smaller pros, but we'll be able to acquire them at a much greater scale. So then when you look at our acquired pros, you can see we were down 23% this quarter. But versus Q1, we were down 41%. So those year-over-year declines have been narrowing. And as we roll out online enroll and ramp up that, we expect our acquired pros to flip over into year-over-year growth in 2026. And then that, on a lag, will result in the growth in our overall average monthly active pros in 2027. So that is how it all kind of comes together, where we have capacity growth already right now in the network year over year, just based on the mix-shift in the pro base. We'll get to acquired pro growth in 2026 due to online enroll and then overall kind of nominal network growth in 2027. Very helpful. Thanks, guys. Thank you. And our next question comes from Stephen Ju at UBS. Please go ahead. All right. Great. Thank you. So, Jeff, so instead of just thinking about AI as being a challenge, there's probably an opportunity for Angi to present a differentiated consumer experience going forward given the data that you have. So from a tech stack perspective, what do you need to build or change to take advantage and move up the marketing funnel and become that destination platform? And secondarily, sort of a macro question here. We're, of course, getting different cross-currents. So I was wondering if you can weigh in on what you're seeing. Thanks. Okay. I'll take the first question. I'll let Russi take a shot at the second one and add if I can be helpful. Look, I think basically, in terms of the tech stack, what we've said is we have legacy technology, which we've got to replace with modern technology as a single platform. We are doing this all, shall we say, AI-first, which means our intent is to integrate AI. We've always used machine learning and algorithms, but we can use effectively conversational AI interfaces and, obviously, the advanced capabilities of LLMs to improve the customer experience. So anything we do new, we are doing with the idea of being AI-first in the product. I think secondly, we are thinking about how we build new pieces of software with agentic code that may actually replace some of our legacy technology or augment. What we have to be able to do then is integrate effectively through APIs or otherwise to deploy that new software. So I think we have to put some thought into how we do that. But this is actually sort of timely. We are in the middle of shifting to a new modern platform at the same time that really high-powered agentic coding has arrived, and we are going AI-first. So everything we are doing is with the thought of, just as I described in response to Dan's question, we want to be able to deploy in a more componentized way to multiple surfaces and channels. And we want to be AI-first in the deployment in order to drive the right matching and actually ultimate job done well, which drives value and actually growth and long-term resilience of the business. Yeah. Then on the macro, reflecting back on 2025, there was kind of April Liberation Day volatility. We recovered a little bit from there. Then heading into kind of the end of the year, you can see in the consumer confidence surveys kind of down 20%-30% in the last couple of months of the year and pointing in the same direction for January. So what we've seen and talking to partners and competitors and such is a little bit of weakness and pressure on volumes. We see a little bit lower mix down in kind of job values and consideration overall is what we're seeing and what's embedded in our numbers and our outlook. Overall, what we tend to see if it gets into a recessionary environment is it gets a little bit harder to get SRs. It gets a little bit easier to retain the pros. Our business generally has a pretty material amount of ballast due to the fact that two-thirds of the business is in kind of nondiscretionary tasks, whether you cut it by service requests, leads, revenue, and pros. Thank you. Thank you. Our next question today comes from Cory Carpenter at J.P. Morgan. Please go ahead. Hey. Good morning. I had two as well. Just hoping you could talk a bit about the revenue per lead decline. I know you called out that in the shareholder letter. So just maybe expand on what you're seeing there. And then secondly, with share repurchases paused, I think you're not able to do share repurchases for a period of time going forward after the spin. So maybe just help us with how you're thinking about capital allocation in the coming quarters. Thank you. Thanks, Cory. Yeah. So on the revenue per lead, we mentioned that we're delivering additional leads to subscription pros. The way the subscription product works is that pros pay kind of a fixed amount, and we deliver leads up to that value. If we're able to kind of optimally just get it exactly that amount, that's not really how it works. If we have a homeowner that comes in, submits an SR, and the only pros available are people who are already kind of at their subscription caps, we want to deliver the best experience. And so we still will deliver that lead to these pros. So it's possible that subscription pros get additional leads that we're not able to monetize. So that'll show up as higher leads, even though we're not able to get additional revenue for it at the moment. Mechanically, that just results in downward pressure on revenue per lead. What's going on beneath the surface is that we have features and functionality that will be rolling out to allow us to monetize, better monetize some of those additional leads, similar to how the functionality and the product works in Europe. Just in terms of the phasing of how we rolled out single pro and the subscription product in 2025, it was on the roadmap, and it's just coming out over the next couple of months. Okay. Then capital allocation, I think as you pointed out, we've bought the prudent amount possible post-spin. It's usually a two-year window. So that would put us at next April one And I think there's a couple of things. One is we have $500 million of debt on our balance sheet coming due in 2025. So we're keeping our eye on that and thinking about where we finance. We think we're in great position with that. We think that between the cash flow we generate this year, our balance sheet, and then our credit line, we have that actually fully covered. So that's just a consideration in terms of capital structure and capital deployment. We would not be against value-creating tuck-in acquisitions, but we don't have any in mind. We would do them at appropriate multiples and make sure that they weren't creating too much complexity and creating. So we'd never rule that out. And then I think we have to see where things play over the next year and where we get to in terms of next April first. And I think long-term, we would obviously, with our ability to generate cash, if our stock stays at the levels it's at, we would still think about buying in the stock. And I think you could never say that a dividend is off the table either. So there's nothing imminent on that. Again, we're more than a year away from doing more share buyback. But I think we're in a pretty stable position, and I think that's how we're thinking about it. I'll just jump in for one sec just because, Jeff, you said that the bonds are coming due in 2025, but there's. 2028. But there's 2028. Oh, I missed it. Yeah, yeah. I just wanted to. Sorry. Correct the record. Sorry. Yeah. I'll get some 2028. Thank you. Yeah. No problem. Thank you. Our next question today comes from Brad Erickson at RBC. Please go ahead. Yeah. Thanks. I had a couple follow-ups. Sorry. On the first one, I may have missed this, but can you just quantify what the current exposure is to the SEO headwinds at this point and then just kind of how to think about how that evolves over time? And then second, on the Google competitive front, can you just remind us sort of or describe a little bit what's having kind of the most acute impact, whether it's just kind of the usual run-of-the-mill algo changes and promoting content versus maybe Google advantaging some of their own service provider customers or maybe a bit of both? Just help us zoom in a bit closer on kind of what's happening there and how you manage that. Yeah. So on SEO, we're currently at around 7% of SRs, leads, revenue is coming through SEO. So that's kind of the current exposure that's obviously been coming down over the past couple of years. And as we mentioned, the way that we're thinking about it going forward is that we'll continue to treat that as a source of homeowners that we want to be able to continue to acquire. But generally, Google is incented to continue to capture as much of their own real estate as possible and not make it available to everybody else. So we're planning the business accordingly to be able to take as much of that share as we can. But we're also focused primarily on growing our proprietary sources of traffic through every other channel. And that's how we've been able to continue to grow our proprietary revenue 17% overall this past year, notwithstanding that we've had kind of a piece of it, which was this SEO headwind working against us. Yeah. And I think if you look forward, you have to understand that there's a couple points of drag on our proprietary the next couple of years in our mid-single-digit-plus outlook for proprietary that we gave for the back half of the year. And we're hoping to exit higher. And obviously, that number shrinks every year as the percentage goes down. I think the way we look at this is that unless there's some external intervention, we don't think Google has any incentive to give anybody any free traffic. It's obviously how they built their platform and their business, but they have somewhat aggressively moved away from it over the last period of years. So I think, in general, the free real estate has receded a great deal, and that's Google. I also think there's been some algorithm changes that have moved back and forth. I'm not sure that they've net impacted us a lot more than others over the last couple of years. And those are sort of always going back and forth. And we have a team who's always working to try and make sure we stay on the right side of those, understand them, and react. But in general, our approach is to put out high-quality pages that get good engagement. And we think ultimately, Google's algorithms are designed to reward that. That being said, we do not think they will increase free real estate. And not only do they have a disincentive to do so, but now I think they have outside competition. I think secondly, everybody knows 10 years ago or so, they moved more aggressively into the local services advertising space in addition to their map product. And so they actually created a product, which a lot of our pros use alongside of us. I don't think it's more effective than us. Some pros would argue we're better. Maybe other programs would argue they are. But we've still effectively built our business with that going on, but they took more of the SERP that way. They've also pulled more paid ads up in the SERP. And then I think finally, obviously, AI Overviews are a different matter. We're actually surfacing really well there but not getting the clicks. Again, Google doesn't have right now a lot of incentive to have people leave the AI overview or the Google AI mode ecosystem. They are working towards selling ads there, and we are actively engaged in understanding how to buy ads there. I referenced their AI Max product on the last call and how we're expanding gradually wherever it makes sense into using that product versus the tROAS or the tCPA or some of their other bidding products. And they would tell us that it gives us better exposure to potential paid ads in AI mode. And so we continue to lean in there. Again, we've been extremely effective buying on Google. We grew our SEM well over 50% in the last year. And we think we can be effective. We don't think they're taking ads away. So we do think that we'll be able to continue to buy. But we also think they have no incentive to let anybody drive down the highway for free anymore because they are trying to grow and be a business. That's great color. Thanks, guys. Thank you. Our next question comes from Youssef Squali with Truist. Please go ahead. Thank you so much. Maybe just to follow up on that last question around AI and LLMs, can you just remind us what are the various LLM platforms you're integrated with today, in the process of being integrated with, and any early learnings or any early insights into kind of how that traffic is kind of behaving and kind of the cost of customer acquisition through that, again, understanding it's pretty early? Then Russi, remind us again of the difference in margin profile of service requests and leads across proprietary versus network channel, please. So Youssef, we're not going to name names publicly until we name names publicly. We have literally had some dialogue with every one of the major players. We've submitted an app to one of them. We're working actively on an integration with another one. We did make an announcement about Amazon Alexa, who is in turn talking to another LLM. And we've talked to the others. So we are looking across all of them. We do not have anything live right now. And we are getting a little bit of modest traffic free from some of the platforms, but it's sort of hard to parse. And it's performing the same way as other organic traffic, I would say. So we don't have a lot to report, either naming names or we don't have much data because we don't have much actual flow. We are actively in the mode of getting our app up and working. We've been able to test those in controlled environments. We think it's going to work very well. We think the best proof of concept there is what's happening when we deploy with an LLM on our site where we 3.3x our conversion to an actual pro selected. Yeah. And then on the profit profile of the SRs through the different channels, it previously was the network channels were more profitable prior to Homeowner Choice. By introducing Homeowner Choice, part of the dynamic intentionally was that we want to bring that experience to parity with our proprietary experience, which involves some intentional, an extra step of choice where you have to choose the pros. And it makes it by doing that, we reduce some of the profitability of the network experience. And now it's pretty comparable between the two channels, maybe a little bit higher on the network channel, but it's pretty comparable. Okay. That's helpful color. Thank you both. All right. I think we have time for one more question. One more, though. Not a multi-question. Yes, sir. Our next question comes from Matt Condon at Citizens. Please go ahead. Great. Thank you so much. I just wanted to ask maybe a follow-up on an earlier question, just the leads per service request. That increased pretty meaningfully in 4Q. And I was just wondering if you could help explain the underlying dynamics there. And then maybe just a quick follow-up just on consumer marketing expense. I know that we're leaning into brand spend in 2026, but 4Q also saw a pretty big step-up or acceleration in consumer marketing expense. Just wanted to hear any thoughts or anything that you guys were seeing that maybe led you to lean in in 4Q. Thank you so much. In terms of the fourth quarter, I think we saw 200 basis points of acceleration as a percent of revenue. It was actually consistent with the second quarter. I don't think it was a material acceleration either. It was a decline in total spend and a modest increase, but consistent with the second quarter. I would say overall, through the year, as we lose SEO and we lean into our paid channels where we're effective, we have seen an increase in marketing as a percent of revenue just as you follow through the year. I think that's how we think about the third to the fourth quarter. Then the first question. Yeah. The leads per SR, Matt, it's very similar to the response to Cory's question where we have additional leads that we're sending to subscription pros, right? So when the homeowners come in, we have subscription pros on the platform. And they're available even though that we've kind of capped them out and they're maxed out on what their contract values are. We're continuing to connect them to the homeowners. And that just results in kind of mechanically more leads on each SR. So look, let me just wrap up with a couple of key points, which is when we started this year, I think we said revenue growth would be -12% to -16%, really driven by Homeowner Choice. We landed the plane, gave up over $250 million of the network revenue. We landed the plane at -13%. The center of our range was kind of $140 million-$145 million of Adjusted EBITDA. That did include two high-confidence $5 million one-time income items. We delivered $140 million without those two 10s. As we look out to next year, our $145 million-$150 million excludes those two 10s, which we still think are coming in. So if you added them back in, we'd be at $155 million-$160 million. The other thing I just sort of point out in terms of profitability is we finished last year with $140 million of Adjusted EBITDA and $60 million of CapEx. The delta is $80 million when you take the CapEx away from the Adjusted EBITDA. And we're going to $145 million-$150 million minus $55 million, which means we're going to be solidly at mid-teens growth on modest revenue growth. But we are returning to revenue growth. We've actually done it in January. We're just not forecasting it because of comparisons and product slippage in the first quarter. And then we're going to proceed essentially on the same path with some more conservative expectations going forward. So we entered the new year having taken the action we took in January with the reduction in force and the restructuring with more durable margin, back to historical investment in our long-term brand asset. We are the leading brand in the industry. We let the investment slip last year for very specific reasons. But we're going back on offense because we feel extremely good about the movement we've made in our customer experience. We believe we have a tailwind with all of the change that came in through the year. We believe we have material opportunity on the large pro side of the business. If you look at pros with 10, 20 employees or more, they're two-thirds to three-quarters of the market. We're under 1% penetrated there. We're 4%+ penetrated in the small pro. We think we have very significant opportunity, and we're investing there. And we think we have all kinds of opportunity. I won't go through my whole opening remarks on AI. And I think we're super excited and optimistic. We think we're on the same trajectory but stronger in terms of profit and cash flow than we were before. We think we have a nice, solid, durable business here that, as an agent, as we've always been, can really accelerate in the AI world. So with that, thanks, everybody, for coming. Appreciate your listening. We look forward to working with you and talking to you in the quarters to come. Thank you. That concludes today's conference call. We thank you all for attending today's presentation. You may now disconnect your lines and have a wonderful day.