All right, now we officially start. Good morning, and welcome to the presentation of Rightmove's result for 2025 . I'm joined today by Ruaridh Hook, our CFO, sitting here, will be here in a second. First, a couple of takeaways. Our 2025 performance showed strong continued delivery in a competitive market, and we will step up the pace further in 2026. We continue to deliver compelling value from and across our platform to both core and other partners. We have a very strong position with consumers, partners, and our data, and with our AI capability, we're enhancing all of that even further. We continue to deliver our proposition. We're executing our strategy. We're excited about all future opportunities to further digitize the U.K. property sector. Now, we delivered some really strong KPIs for last year.
Revenue growth of 9% was supported by ARPA and membership increases in the core business, as well as contribution from growth in our strategic growth areas. Underlying operating profit growth of 9% reflects our revenue growth and ongoing investments in people, technology, and product delivery. Underlying EPS grew by 11%. We increased capital return by 21%. Finally, time on site at 16.8 billion minutes was the second highest on record, only beaten by the COVID exceptional burst in 2021. Said differently, the equivalent of 32,000 years of time was spent on the Rightmove platform last year. We made some strong operational progress as well, right across the platform last year. From the left, over 85% of that large audience came through direct and organic traffic, and we grew our app users by a strong 11%.
We continued to evolve to meet consumers wherever they are. We doubled our engagement numbers in social media channels. We saw strong penetration of our top packages in estate agency and new homes, as well as a very fast start for our latest and market-unique estate agency product, Online Agent Valuation. Our agency retention was the second highest in over 10 years. Third-party surveys showed record positive sentiment scores for Rightmove. We continued our strategic and operational progress and growth in the strategic growth areas. All of this was delivered through Rightmove's platform and leading data. We did over 6,000 tech releases. After a multi-year build, we now have 31 live strategic AI projects at year-end.
It's an increase of four on our November update. We tripled the number of data models used to process our proprietary data in the platform. This strong stance is down to purposeful work and investments over the most recent years and has a strong trajectory for future product delivery. Finally, on people, we have a world-class, engaged, and energized team. 89% of our team describe Rightmove as a great place to work. My sincere thanks to all hard- and smart-working Rightmovers for delivering our results of last year. It is a competitive market out there, but our position is stable, and it's strong, and that's because we keep delivering great value for both consumers and partners. We remain the leading place for consumers looking to make a move in U.K. property.
While facing various competitive dynamics over time, we have for years averaged over 70% share of portal time on Similarweb and over 80% on Comscore. In December 2025, we were at 75% and 89% respectively. That love and trust from consumers drives frequency, leads, and of course, a lot of data signals. Those enable us to drive strong outcomes and value for our over 19,000 U.K. estate agents and new homes partners. Now, I want to touch on that value point a bit. We operate in a competitive market, and we always gauge how we can do even better. We commission third-party surveys quarterly, with over 1,600 independent agents contributing responses. The top left chart here shows that the total positive sentiment scores from those surveys. There are two big takeaways.
One, just in absolute terms, we've seen a positive trend and a new record high, actually, by the end of last year. Market conditions and general sentiment out there often impacts survey responses. In the context of the weaker Q4 in the property market, through the U.K. budget hesitance, that's actually an excellent result. Two, in relative terms, you can see a 1.7x differential between Rightmove and the main portal competitors. Now, we ask for feedback at branch frontline, branch management level, and company management levels, and we also go deep on several subcategories. You can see that we lead across subcategories, across business results, value, and inclusive services at the bottom of this chart. We rate really well in what's a competitive market, yet we, of course, always look for opportunities to improve and for all partners.
Part of the value and those strong scores come from our Building Success Together program, which we launched in early 2024. We invest resource in supporting our partners' business objectives. We also help them to understand what happens in the market and what Rightmove can bring. As noted top right, this comes in many forms and at true scale. Dedicated account management in the field, our Rightmove Plus and Rightmove Hub tools, which are both available to all partners, regardless of package levels. We're sponsoring and collaborating with several leading industry organizations across the estate agency, new homes, and rental operators. We continue to invest in and progress these, too. Rightmove Plus, as an example, is the business management tool for partners. Last year alone, had new features and enhancements introduced over 25x .
Our partners' engagement value from Rightmove Plus is clear: 28 million sessions recorded in the year. In summary, we deliver Rightmove outcomes and value from a broad range of solutions, packages, products, data, insight, training, dedicated servicing through our account management and support teams, and we measure these results. Let's move to the property and markets for a bit. Within sales, top left here, it was really a year of 2.5 . H1 was strong, building on 2024, and with successive Bank of England rate calls, rate cuts. H2 was weaker year-on-year, due to the fears around the late autumn budget. If you take them together, 2025 as a whole, saw 10% more completions versus 2024, and that was in line with long time, long-term averages.
Looking at the year ahead, top right, there's been a clear post-budget bounce back in available stock, which is now at a 10-year high. This has caused slower price growth, which is, of course, supportive for buyers in the market. These elevated levels of resale stock is less helpful for new homes developers. On the bottom right here, shows new homes as a proportion of total for sale stock on our site. With approved planning action, applications at an all-time low, we don't expect a material recovery of the development numbers in the market in H1 this year. With the rentals, bottom left, increased supply and reduced demand continues to improve the more extreme imbalance seen in previous years, and which we have talked about.
The 2025 average of 10 enquirers per available property is still above the pre-COVID average of 6-7, though. Across all these segments, of course, mortgage rates is a key driver, and it continues on a steady downward trajectory. At the 31st of January, the average five-year fixed rate was 4.35%. That's 55 basis points lower than a year earlier, and that's per Rightmove daily mortgage tracker. With that, let me pass over to Ruaridh for more detail on our financials.
Thank you, Johan. Good morning, everyone. I'm delighted to present our financial results for 2025. Overall revenue grew 9% compared to 2024, with strong growth across the business. Starting with agency, row one in the table, revenues increased by 9% to GBP 305 million. If you look at the chart on the right, the light blue bars show that this growth was driven primarily by ARPA-led gains, which continued to be mainly discretionary. An additional contribution of GBP 6 million came from increased agency membership numbers. Moving down the table to new homes, revenues here also rose 9% to GBP 75 million. This was in spite of continued headwinds in the new homes end market, with new builds coming to the market remaining subdued.
You can see the impact of this in the chart with the dark green showing revenue growth contribution of less than GBP 1 million from higher average membership increasing by 1%. The ARPA growth contribution remained strong, contributing GBP 5 million. At the bottom of the table, our strategic growth areas delivered another strong performance. Revenue increased by GBP 5.7 million, up 25% to GBP 29.1 million. Commercial revenues grew 13% to GBP 15.3 million as we continue to focus on customer acquisition, with membership increasing 29% year-on-year. Mortgages revenue was up almost 50% to GBP 6.8 million. This was weighted towards the first half of the year, mainly reflecting the timing of interest rate changes and hesitancy in the property market around the budget, impacting activity in H2.
Rental services made up of our Lead to Keys product, referencing and ancillary services, saw revenues up 35%, driven by strong growth across the Lead to Keys product. For completeness, the non-SG&A parts of other revenues, being data services, overseas, and third-party advertising, grew 2% year-on-year. Revenues outside the core represented 11% of group revenue, up from 10% last year. Compared to December 2024, across agency and new homes, membership increased by 225, up 1% to 19,272. This increase was due to growth in agency membership, which increased by 261, up 2% on December 2024. This was due to high agency retention of 90%, continued growth in agent formation, as well as current partners opening new branches.
Within new homes, we saw a year-on-year decline of 36 developments, down 1% at year-end. You can see in the bottom right chart a decrease of traditional developments in orange of 113, offset by an increase in housing associations in teal of 77. New developments coming on site remain low. We are not seeing a pickup in build rates and have seen traditional developments fall to their lowest level since January 2018. We do not see this changing in H1, but continue to be optimistic that developers will be encouraged to build more by H2 and in future years. Overall, ARPA increased by GBP 97 to GBP 1,621. 60% of ARPA growth was product-led, with similar percentage in both agency and new homes, as our partners chose to upgrade or purchase incremental product.
The remaining 40% of ARPA growth came from contract renewals, which all proceeded as expected. Given partner engagement with our strong suite of value-adding products, we expect a similar split this year. In terms of product ARPA growth, we saw upgrades in agency come from multiple sources, ranging from upgrades through the package ladder, from lower threshold packages, to new joiners joining straight into the top package. You can see this in the pie chart for Optimiser Edge joiners in the middle of this slide. The migration of the old top package, Optimiser 2020, has gone well, and will be fully retired by H1. Joiners in new homes to the Advanced package, shown top right, similarly came from upgrades and new joiners.
We had a new top package, Ascend, launch in May, with 818, 28% of developments live at the end of the year. We expect a similar split of upgrades going straight into this top package, but flag that the Advanced package remains highly attractive, especially for smaller developers. Expect to still see good inbound into Advanced next year. Taking these two pie charts together, you can see that key for both New Homes and Agency is that we do not rely on a single source of joiners to the top package. Expect penetration to continue to increase in both. The other driver of ARPA growth comes from incremental product purchase. You can see from the charts at the bottom for both Estate Agency and New Homes. ARPA increases at the initial upgrade in month one.
This is the column marked "Upgrade." We see ARPA increase across the first year and the second year. In both estate agency and new homes, you can see that ARPA keeps growing far past the initial upgrade. This happens as partners choose to purchase more of the same products or add additional products to their package mix. We have shown the previous top package in agency Optimiser 2020, and in new homes Advanced, to illustrate how we have seen this before, and that the initial months of the new top packages in both agency and new homes are performing as we expect and have seen previously. We know that continuing to provide great value and superior outcomes to our partners through continually evolving our new products, sees them choose to engage further.
At the end of last year, we added Online Agent Valuation, exclusive to Optimiser Edge partners, with an average price of GBP 170, providing both another reason to upgrade to the top package and also encouraging existing partners to increase their current product spend. Moving on to costs. Underlying operating costs increased by GBP 11 million year-on-year, resulting in a 70% underlying margin as we invested with discipline and within our cost framework. The main driver of costs remains our investment in people, up GBP 4.6 million or 7%. The other main cost component was our continued investment across technology, with an increase of GBP 4 million. In the year, there was GBP 9 million of internal labor capitalization, with total CapEx at GBP 10 million.
As guided in November, we expect to see an increase in labor capitalization in 2026, with total CapEx to be around GBP 16 million, less than 4% of revenue. In 2026, we will see investment as outlined last November, which will mainly be in people. We anticipate over 100 joining before the end of the year in roles across data, product, and engineering. A few of these roles will be through our new flexible resource provider, which will provide us with the flexibility of headcount over the investment phase. Other material increase in cost will be the AI-powered operations area, with work on the back office initial phase already commencing. All in, post-capitalization, this incremental investment is expected to total around GBP 12 million in 2026, as guided in November.
We remain highly cash generative, with a cash conversion ratio of 107% of operating profit. As we continue to grow the strong cash generation of our business, this leaves us well-placed to return surplus cash to shareholders. This year, a total of GBP 220 million was returned to shareholders, GBP 141 million via share buybacks, and GBP 79 million via dividends, an increase of 21% year-on-year. We reduced our share count by 2%, meaning over 40% of issued shares have now been repurchased and returned 6% of our year-end market capitalization in the year. This morning, we announced a final dividend of GBP 0 .659 , bringing the total dividend to GBP 0.1064 . There will also be a share buyback program of GBP 90 million until the 31st of July.
This will be funded by the growth in earnings, but also reducing cash reserves from December's GBP 43 million to around GBP 20 million by half year, which we see as sufficient to manage the working capital of the business going forward. Our capital allocation policy remains: prioritize investment in the business, evaluate value accretive M&A, and return all surplus cash to shareholders via a progressive dividend linked to earnings and buyback thereafter. Turning to financial guidance. This remains the same as set out in November. Looking at the right hand of this slide, revenue growth in 2026 will be between 8% and 10%. We expect H1 growth to be lower than the full year 2026 growth, with a higher growth percentage in H2.
This is due to the high comparator in H1 last year, particularly in mortgages, which saw significant activity in H1 2025 due to the stamp duty changes and falling interest rates. In new homes, due to the full year impact of 36 developments, fewer developments, contributing a negative revenue comparator of around GBP 1.5 million. For Core, we anticipate that membership will grow around 1%, and we lifted ARPA growth to between GBP 110-GBP 120. At an overall level for the SG&A, we anticipate growth to be around 20%-30% range. Underlying operating profit will grow by 3%-5%, resulting in an underlying operating margin no lower than 67%.
With no change to our longer-term target set out in November, we anticipate underlying operating profit growth in later years to be at similar levels to revenue growth, as we still see no reason for a margin lower than 67%. That concludes the financials. I'll now hand you back to Johan.
All right. Thank you, Ruaridh. Our investment case outline here will be familiar to most, and this summarizes our approach to value creation at Rightmove. On the left, Rightmove has exceptionally strong foundations. We have established a differentiated leading platform at the heart of the U.K.'s large and structurally growing property market. The platform is digital, low cost, capital light, driving high returns on capital. The subscription-based B2B model has a proven ability to deliver and generate value in all market conditions. Moving to the middle of the diagram, we're using powerful data and profound network effects to deliver that value to all stakeholders. With it, we're executing an expanded growth strategy with targeted investment and delivering data and AI-backed product innovation, and that is done through a high caliber and very energized team.
We're entering now our 27th year with confidence to deliver a larger, diversified, yet very connected Rightmove platform. All said, this will continue to deliver compelling financial outcomes. Now, our strategy is to develop the leading digital ecosystem for the whole moving experience, powered by exceptional data and network effects. Our people, data, and platform really are the foundations and strong differentiators for the three business pillars of Core Partner, Consumer, and New Growth. The property market is a huge economic activity, and we think there's a long runway to deliver more digital value and grow our business. That, of course, includes the use of AI. Now, there's been a lot of debate who the winners and losers might be, both for classifieds and more recently across a range of industries, really. I wanna talk to property classifieds, specifically.
In my view, there are really four components you need to win to compete effectively, also in an AI world: consumers, partners, data, and AI capability. We're really well-positioned across all of these. We were well-positioned before GenAI, we will be with the next generations of AI as well. Here's why: We're a technology company. We built up market leadership through deep knowledge, digital leadership, and deep layers of servicing our industry in the first three of these four components, and that's been done over 25 years and at an increasing pace. We keep doing that day in, day out, improving all the time. The numbers are leading, and they're deep. Now, the most recent components of these four is, of course, AI capability. AI models and tools, they're fast developing, it's dynamic, and they're not fully defined yet. Here's the thing, though.
Anybody can get a hold of AI capability. It's an enabling technology that you can buy, skills you can hire, and that you can learn to operate. We've done exactly that, and for several years already. What Rightmove has is a very, very solid performance and performant platform and business model. It's creating the fundamental attributes that are mentioned here, important to any business success. In turn, they all boil down to two things, which again, deliver true business results and sustained leadership: trust and vertical innovation. We obviously thought a lot about this. In our view, in the case of property classifieds, is that LLMs or startups running on LLMs are missing or quite far away on three of these four components that matter so much in this particular vertical.
ChatGPT has been around now for three years, referral traffic to us is still under 0.5%, actually, their U.K. app downloads and traffic has leveled off in the last five or so months. More so, I don't think they or other horizontal LLMs can or want to service our vertical as deeply and focused as we and others do, nor to innovate as relentlessly and deep in the specialty of it. I'll be very open-eyed and give the large LLMs the upper hand of AI capability and AI innovation overall. Remember, again, they actually enable and sell that capability to buyers like ourselves. As we add this AI capability to Rightmove, we combine it with the first three components that we already have and that are so strong.
We're in the best place of anybody to innovate and service this vertical in new and even better ways. I actually gonna go and cover these four components in a bit more detail because it's so important and so topical. Let's start with consumer and partner. You're familiar with network effects and how they're part of a great business like Rightmove and how we invest in them. I think it's very important to understand that in the case of home exchanges, there are three special aspects of these network effects, which make them even stronger for property classifieds, and certainly in the U.K. First, in the middle, property transactions, they're high value, highly personal, take a particularly long time in this country, and they're very often done in joint deliberation with another person.
There's also an incredible amount of browsing done on properties because of two things: homes, they're fun to dream of or to be inspired by, and also becomes, because finding the right one and really deciding when it's time to move is such a serious and important life decision that comes at a high price. The habit loops are therefore massive. This is very different from a number of B2C categories like e-commerce or research of different kinds, where AI or agents can provide an alternative and shortcut path. Secondly, to the left here, the same consumer actually plays multiple roles. If you consider the four key roles who use Rightmove and their multiple use, very often a buyer is also a seller, and a seller is also a buyer, in the same chain of events or at different points in life.
There are two and a half million private landlords in the U.K. renting to tenants, and those landlords, of course, themselves live and move. There are parents who help their kids with a rental or a first-time purchase, while they themselves might be downsizing or buying a second home. Here's the point. The individual gets value from the same property platform for many different needs. They've seen it in the past, they know what the quality is, and they are being in different roles. The platform is trusted, it's specialized, and it has all these different audience roles. In a way, this forms like a consumer-side individual, individualized network effect in itself, not just across to the other side of the platform. Again, that's very different to, for example, e-commerce and other verticals, where the consumer might only be a buyer.
Thirdly, of course, in the U.K. property vertical, there's a diverse nature of our partner base. Estate agents, new homes developments, developers, rental operators, commercial, and smaller niches. Even in a single branch, estate agency, you can have sales, lettings, commercial, potentially financial services, a business owner, and branch staff. The U.K. partner is very fragmented and with low barriers to entry, and there are many different roles that benefit from being on the platform. Agents are local property experts, they can access a highly effective audience platform and with a lot of services included to power their business goals. In our view, when you combine these three points, property complexity, consumer multi-use, and agent diversity, you realize that the trusted and vertically specialized UX of the portal will not be replaced by generic or horizontal AI interfaces.
Let's talk about the fourth component, AI capability. We've been building a great tech and data AI capability for a few years now, as we reported on several times since 2023. The simplified, and I know it's simplified, tech stack view on the left here outlines how our core platform is built on Google Cloud with logically connected enterprise tools like BigQuery, Looker, Model Armor, Vertex AI, and so forth, and is running AI models from Google, like Gemini, Nano Banana, and so forth. We have a close strategic and product team collaboration also with Google, and we are actually working together, and we have a good view on what's coming in the future. We have orchestrated the platform, the stack, the pipelines, to nevertheless be flexible, performant, and trustworthy.
We have relationships with, and we also use Microsoft, OpenAI, Anthropic, and a host of smaller solutions. Some of those smaller ones are pure-play AI, some of them are AI-enabled existing software. Our data science team, they can build and connect proprietary Rightmove data models or external models or a combination of them. In November, we showed you one example of the proprietary model and how it uplifts the results, something that is only possible for us because we're in the stack. The stack enables us to deliver more value and differentiated outcomes for partners and consumers, and of course, gain operational leverage and productivity for ourselves.
The 31 strategic initiatives, plus a whole host of many more AI tests across the business today, will soon be less of a number counting exercise, Rather, it's gonna be completely infused in an organic way of operating. We're perfectly set up to leverage AI capabilities. Now, I'll come back to a very crucial component, data. We estimate that over 90% of our data is proprietary. It's also interconnected, and we leverage it with human expertise and usage in mind. This data is not available anywhere else, and it keeps compounding inside our ecosystem. We've shown you many examples of large data sets in the past. Here, just outlining a few examples, but to illustrate how unique and valuable this data is. For property, as an example, we have over 28 million unique properties on our Rightmove optimized UPRN address framework.
Someone might say, "Well, that's all scrapable, isn't it?" Fact is that over 50% of the metadata underpinning a Rightmove listing is not scrapable from the face of our site. For partners, we have, for example, built 57,000 defined geographic agent patches. We dynamically optimize them with our data and also with input and tailoring from our partner agents. That drive unique insights, products, and great outcomes. For consumers, for example, again, the 69 billion first-party signals. They don't only provide that strong habit loop that I mentioned before, but they of course, convert to outcomes through moving auction strength of buyers and sellers. They also drive unique products, insights, and recommendations, and provide fodder for what we develop next. The real magic and protection is how those and many more data points are interconnected in the platform.
There are a few more examples in the middle. The data compounds and it fortifies. Finally, in the third column, but not to be forgotten, we overlay our human expertise to enrich this data. It's being completely vertically focused. We also make sure it delivers real outcomes and value for humans that is using the data. All said, we hold the living map of U.K. property moving. The value is not in AI itself; it's what AI can deliver when it sits on the best property data in the U.K. To sum it all up, we combine these four components. What we have is one connected ecosystem, already powered by data, and it's enhanced by AI. All right, over to some of the concrete product delivery that drove the 2020-2025 results, and a bit of a glimpse towards 2026 and onwards as well.
We increased the pace of delivery in 2025, with only a few of the features illustrated here. I'm going to talk to the renter's checklist on the left. It's important example, because it's part of our rental market solutions to digitally enable more of the moving journey. We've seen some strong growth metrics in 2025. A few of them are noted here. With this renter's checklist for consumers, we put all the tenancy admin in one place on MyRightmove, seamlessly integrating it with things like open banking and verifications on what to do next. The average user revisited their checklist 8x . The information is stored in their MyRightmove account, so it can be reused, and that, of course, builds a lifetime value opportunity for us.
Like many other products, this product also helps the other side of the platform, in this case, lettings agencies. They benefit from operational efficiency through the Enhanced Leads, and seamlessly have those in their CRM. Now, they can also operate the entire flow digitally in the Rightmove Plus environment, from referencing deposits and many more things, all the way to contracts. Quick step back to the outline from November of how we're accelerating the consumer demand going forward. Number one is that we are adding and enhancing ways of searching. Number two is that we're accelerating our services in a consumer home moving journey, what we call Beyond Find. Here are about two examples, what we're working on: the Move Journey Assistant set up for sales, and the expansion of MyRightmove into MyHome, a full service hub for homeowners.
Across the consumer domain, we have around 25 key releases or so planned for 2026, and for context, that's more than the entire platform, consumer and partner sites together, delivered in 2023. I want to expand a bit on conversational search, no surprise, which we launched only a few weeks ago to a limited amount of traffic. Here's just a demo of what it looks like. I'm going to talk over while you follow this. This experience and feature is built through our partnership with Google Cloud. It's using Gemini models. It's trained on and interrogates our listings, text and images, and we use our over 1 billion proprietary image databases, database and many attributes that goes into the listings.
As of today, it links straight into listings on the main site. We'll evolve this tool, led by the data that we see and our design expertise. We're going to make sure that we deliver a high-quality experience. Data so far, from thousands of conversations, tells us that users seem to have a pretty good idea of what they're looking for. They continue to explore and engage with tools in the main flow of listings and on the site. So far, those who engage with conversational search are almost 3x more likely to send a lead versus the control group. Overall, feedback has been very positive. I want to consider a little bit the conversational assistance in searching a bit more strategically. First, on the left here, this is really, in many ways, it's just a continuum of changes.
However you discover, we have you covered, right? We're entering another search modality or paradigm for consumers, and our position is the same as it has been with previous changes. Discovery is key, right? The classic behavior of visual scrolling and comparing properties, I believe, will always be there. Longer term, I also think this holds a real amplification opportunity for Rightmove. Conversational search will enable hyper-personalization and new utility for consumers on our platform that they couldn't get before. AI assistance will be useful up and down the funnel and seamlessly provide complementary information along a complex moving journey on the platform. This will drive two things: higher platform engagement and substantially more intent in behavioral data signals.
We can convert that data signal to increased value and targeting for core partners and for diversified revenue opportunities, just like we have done in the past. We have already started a few years back to build many more of these consumer features with exactly that in mind. You can see some of this in the graph and in the table metrics here. Impressive growth. A lot of that comes from well-defined features and, of course, the scale of the audience and traffic that we can apply them to. Every feature we build is research and data-backed. It brings utility, frequency, and data to us on an ongoing basis. With it, and it's noted right here, we create enhanced partner value and, of course, revenue opportunity for Rightmove. Some of these improve or enable new products for core partners.
For example, the Enhanced Leads to letting agents or the appointment bookings with the New Homes Ascend package. Others are monetized separate through commercial relationships that we have, like, for example, mortgages or ancillary letting products. Here's the thing: as we scale and compound this data, we just increase the revenue and profit opportunities. Over to the partner side. We released significantly more product and optimization source of partners in 2025. A few key ones are set out in this slide. I want to highlight Online Agent Valuation on the left, as Ruaridh mentioned before. It soft launched in the fall. It's off to a great start. This tool works on both sides of the platform.
It enables consumers to receive a digital valuation estimate from an estate agency with a quick turnaround. It's an opportunity for agents to start a new online relationship with a potential vendor through our platform. It leverages and reinforces our existing valuation domain of various tools, slotting in very logically with instant valuation, local valuation alerts, best price, and premium price guides, and so forth. Agents, in this case, can also choose to use an AI tool to support the responses in OAV. For those that do, we have seen so far in the data that the response times are 16% faster on average, and the cohort actually books 20% more visits.
OAV, I think, is a good example of where AI is an enhancer of an already great digital product with real value, but AI is not the entire product itself. Finally, with OAV, Rightmove platform also gets more data signals through up-to-date photos and property attributes supplied by the consumer, and this is before the property becomes a listing and gets put on the market. That, of course, can feed into our AVM, which is a business line on its own, and also powers many other things internally that we can build on for the future. Both 2025 and 2026 show how we are developing across several product lines and segments, much more in parallel than in the past, and with AI bringing more efficiency and marketing opportunity to partners.
Moving on from core to the strategic growth areas, these grew, as you heard from Ruaridh, by 25% as a group, and that's close to 3x the core growth rate. Operationally, we've taken some great strides forward in the year. For commercial, we added 275 new members to the platform. This year, we will launch our new search pages, and at that point, every aspect of the user web journey will have been completely overhauled to commercial-first experience. We'll also be launching our first chargeable product in the segment year and year. In rental services, revenues grew by 35%, and as we set out in November, we started to roll out the upfront modules of Enquiry Manager and Enhanced Leads to dual agents within their core subscription. It's a process that is ongoing over 2026. This is an exciting market penetration step-up.
It brings efficiency to agents, to landlords, and to tenant applicants, and it's at true market scale. In mortgages, we saw strong growth overall. You will have seen that we announced a new exciting partnership with NatWest, the U.K.'s leading digital mortgage lender, which will be introduced in April across both sites and our apps, and we'll also continue to build out the broker opportunities over the course of this year. Finally, again, an importantly reminder, the SGAs all strategically reinforce the core platform, drives user utility and frequency, and again, thus, the great data sets that we have. This slide is a reminder of the three focus areas that we described in November. We are positively stepping up the pace with an eye to the medium-term opportunity of a more diversified and technically advanced platform.
We're driving towards that larger digital opportunity in the U.K. property ecosystem. Also as a reminder, we set some really ambitious midterm target KPIs for these initiatives, and I'm glad to report that all of this is mobilized in one way or another, and the capabilities will build and realize throughout 2026. We're going to see results along the way. One example, of course, being the successful launch of conversational search already in the very beginning of the year. We'll come back to these areas and the KPIs over time. I hope you can see that we drive this business with discipline, high quality, and our goal is to deliver strong value and returns. In conclusion, here are the key takeaways I showed you at the start of the presentation, and I want to repeat them. We're happy with the strong results in 2025.
It was a record year for innovation for Rightmove. We look forward to an exciting 2026. As you can see in the graph, we're stepping up our innovation and delivery considerably yet again. We will grow revenue and profit in line with guidance, adding to strong financial returns in both the short and medium term. With that, we're going to go to Q&A. Ruaridh's going to join me up here. Please raise your hands. Yes, some already did. Say your name when you're passed a microphone, let's aim for two questions in the first instance. We can double back if it's fine.
Yeah, we'll come back. We got chance.
Great.
Jess, do you want to go?
Yeah.
Which one is yours?
Hi. Morning, it's Jessica Pok from Peel Hunt. Two questions, please. The first one, just on the ARPA guide, Ruaridh, GBP 110-GBP 120. Can you give us an idea of how we should think about that, agent versus new homes, given the trends that we've seen last year? The second one maybe on mortgages, the new relationship with NatWest. Any color on what triggered the change and what we can expect from that relationship in the near term?
30 a second. On ARPA guidance, GBP 110-GBP 120 is the blended ARPA guidance. Expect estate agency to be towards the bottom end of that, and new homes well above the blended rate. I would flag that in both EA and new homes, we expect their ARPA growth to be higher than they saw in 2025.
All right. On NatWest, yeah, we're very excited about entering a new partnership here. We've had a great partnership with our other partner for the last couple of years. NatWest is really the number one mortgage lender in digital channels. That tells you, I think, something about the vision alignment that we have. We continue to work deeply with one partner 'cause we're quite keen to both build a business, of course, give more consumers more utility on the platform, but really also try to innovate along the way in this industry, which is still very fragmented and analogous and offline and so forth. Those are really the few simple reasons behind it.
Will?
Hi, it's Will Packer from BNP Paribas. Couple of questions. Firstly, could we talk a little bit about agent relations? From today's update, the survey data looks very encouraging, although I note we didn't see the absolute numbers, but that'd be interesting. Retentions at record levels, you've got new agent additions. In contrast, if you read the trade press, it all sounds a bit grim. You've got the court case coming, and I think there's a perception that your relations with your customers are more adversarial versus some of your peers globally. How do we square that circle? Is it there's a few loud adversarial agents, but the median agent is getting happier? Could you just give us a bit of color there?
Secondly, your framing around the labor intensity of Rightmove is a little bit different to some of your peers within classifieds and other platform businesses. You're growing headcount aggressively. It sounds like that's gonna continue for a little while. Could you frame that for us? Is that catch-up investment because the previous management team didn't hire enough people? When can we see the labor force to stabilize? Any color there would be useful. Thank you.
Yeah, of course. I'll take two and you jump in. Look, the first one, you mentioned some of those KPIs, which I think are standout, right? High, second highest retention in a decade, highest take-up of our new product, OAV. We had record uptake of Optimiser Edge. That's shows customers are engaging with our products and really happy with the outcomes. You know, that, for us, is a real sign of strength in terms of the relationship we have with customers, of which over 80% are now with us for five years. They know us well, they know our products well, and we work with them to grow their businesses.
You're always gonna have, you know, a small minority who might be louder than the majority. I would say that those KPIs are what we look at to show the strength of our products and the value that we provide our customers. We also, as we showed today, do monitor sentiment, and we're delighted to see that sentiment not only much higher than competitors, but growing. You know, we don't rest on our laurels. We take it very seriously, and we keep our finger to the pulse in terms of how agents are feeling, and we support them as the property market ebbs and flows. Ultimately, for us, key, coming back to providing those great products, I think that take-up really shows it.
In terms of the labor intensity, yes, we're adding over 100, and those 100 people are gonna be building some fantastic products and fantastic assets. They're gonna make Rightmove stronger and on our path to higher growth. You know, that, for us, is a short-term investment. It's gonna allow us to build many of the things that will enable us across the domains that Johan talked about. You know, we've provided a flexible resourcing partner as well to help us, you know, accelerate or pull back in that recruitment as we see fit. For us, this is about driving higher profit growth, and this is about us building things that we're really excited about, that we see great ROIs from, and that requires some headcount in the short term.
What you will see and what we look forward to bringing to you on a regular cadence, is some of the really exciting products that they're gonna build.
Thanks.
Will?
Thanks, Will Larwood from Berenberg. Firstly, just obviously integrating a lot more AI functionality going forward, consumer, with, like, conversational search, et cetera. How can we expect sort of the cost profile of the business to shift, particularly thinking about sort of using more compute going forward? Secondly, you mentioned it very briefly in terms of the mortgage broker side of things, but if you could provide an update on that'd be great.
Yeah. Yeah, I'll start with AI. Look, we obviously anticipate and budget for, you know, compute costs that didn't exist in the past because of this. I think there are a couple of important things to remember. Again, back to that slide of how we set things up. We set it up in a very organized, very orchestrated way, and we have, you know, fantastic control over this, just like we have on other costs. Here's the thing, it's a cost to deliver opportunity, all right? If you look at token cost overall, I mean, they keep coming down by 80%, 90% on an annual basis across the world, right? Both because models become more efficient themselves and because there's a lot of competition out there.
It's an item to keep track of, but it's not something that concerns us particularly, right?
Mortgages broker.
Yeah, mortgages, I'll go to that one as well. We are, I think we talked a little bit about this before. We have brokers on the platform, but it's a small part of what we do today. A lot of attention has been on the MIP product, building awareness with consumers, seeing what that does, and obviously deliver great results. What we did, you know, last year, was prepare a little bit more to be able to scale the broker side of the business, as opposed to one-to-one relationships with brokers, because there's literally 5,000 of them in the U.K. It's also really about, you know, looking at this as.
I think of this as an inevitable trajectory kind of thing. Because of who we are, the interest in properties, the fact that 2/3 of properties needs to be financed, us having some kind of service in this space makes sense, and that's evidenced already. It's a long-term thing to build. There's still awareness, there's still optimizing it. There's still or still, but what we're trying to do, again, is build a better experience and an experience that doesn't exist anywhere else. That takes some optimization. It's 2% of our revenue today. We're happy with the growth, but there's gonna be a, you know, test and learning as we go along with it, and we're executing on it really well. Over time, there will be broker options as well.
It's about understanding the consumer, and again, because of all the consumers that we have, what's their mindset, right? Are they close to transaction or are they really out shopping and still wanna get an affordability check? Segmenting that, and dissecting, and making very logical for them, and therefore funnel them to, you know, different opportunities for financing is important. That doesn't, you know, come just from saying, "We do one thing on the website," right? Again, fantastic opportunity going forward, and lots of money in this space, and I think we have a real right to play.
Andrew?
Thanks. Andrew Ross from Barclays. I've got two on AI. First one's about the conversational search you've rolled out on platform. What are you observing in terms of the, you know, the conversion rate from search into leads or any kind of outcome-based metric that you track from that? Kind of what impact is it having on clicks onto featured and promoted listings as part of it? That's the first question. The second one is, you guys have obviously applied to put an app into ChatGPT. Can you just give us some context as to what the thought process was as to why do that? You know, on the one hand, you're kind of feeding the beast. On the other hand, first-mover advantage is where the users are.
What are the kind of puts and takes, how are you thinking about them?
Yep. So when it comes to conversational, again, I outlined a few stats, right? We, because of our traffic, and in spite of having it on a minority of that traffic already, we've seen, you know, thousands of conversations, lots of messages, very good flow through in terms of people getting the results that they wanted, and also, you know, as expected, coming back over to the main site, and digging around and using different tools and so forth. We have seen that uptick of about, you know, 3x the sort of lead-sending propensity. You know, to be honest, is that cause or correlation?
It could be the most qualified users that have been on Rightmove before and so forth, or is it a novel way, and therefore they become interested? I think it's too early to say, and anyone who talks about these data points, I think it's important to get that kind of context. Now, again, I point back to this as an opportunity, right? The fact that, how consumers experience the site and the listings and what they do with it, first of all, this is a first version of the integration, and how partners, you know, show up in that will, of course, evolve over time, right? It depends on how much of attraction this will see from consumers.
You know, small minority or complement to for a lot of people to what they do, it is just simply too early to tell. Again, the opportunity, if you think about it's a much more personalized and engaged consumer in different ways doing this, and that further qualification of someone's behavior has value. The fact that there's potentially new or, for sure, different commercial opportunity around this is also there, and that goes through our heads, right? It's early days. The second one on ChatGPT, yeah, I think you maybe outlined it well, you know, puts and takes, consideration. Look, today, they're meaningless in terms of a feed or a platform for, you know, people actually looking for and, you know, going after homes.
As we said with those stats, right? I think most of the peers report the same numbers, very small. Look, it is a tool that lots of people use for, you know, different things. For us, this is a test and learn, right? We wanna be where some consumers are, see what we can learn from that. Very importantly, of course, it's an app that we created. It basically displays listings and consumers then go back and do much more to the experience, where they have all of that experience, and again, all the data and tools and their own history and so forth, on Rightmove, and that's what we expect going forward as well.
Cool. You keep all the data, right?
Yeah.
Joe?
Hi, Joe Barnet-Lamb, I'm from UBS. Two from me. First one, a technical modeling one. I think it's important for the interpretation of ARPA guidance. Historically, forecasting agency was simple, as ARPA times by the average membership. We now have a growing proportion of non-ARPA revenue within agency. Can you just clarify which revenue streams within agency are non-ARPA, how big they were in 2025, and how you expect that to change into 2026? The second question is just on buybacks. We see you're effectively, you know, restarting and spending excess capital generation beyond dividends, and spending half of the GBP 40 million that you've accrued whilst you weren't buying back. Can you just give a bit of color on why you aren't spending all of the excess cash to get you back down to zero?
A sort of general commentary on, sort of the merits of running a net cash balance sheet, given where your share price is. Thank you.
Sure. Two for me. Look, yes, you're right. ARPA used to be much easier, where you took kind of customer numbers multiplied by ARPA, and you got roughly our revenue number. There is a non-ARPA element, which is because we don't count Agent Accelerator in our ARPA calculation, because it's a program rather than a package, and also insurance revenue in the rental service as part of the business, because that's insurance to consumers and landlords, so therefore, it's not counted under the average ARPA. Those two together used to be almost zero a few years ago. Great to see them grow, and they're around about GBP 3 million. That's what you should add on, once you take your average ARPA times by your customer numbers.
In terms of the share buybacks, great, you know, first thing to flag, we return all of our surplus cash to shareholders, and we don't see that changing. We've reduced our cash reserves from GBP 40 million to GBP 20 million, which we think is sufficient to run the business from a working capital perspective going forward. For those that have been with Rightmove for a long time, GBP 20 million was always the number that we used to have, and feel very comfortable that that's a manageable cash reserve for our working cap. Flag that. In terms of looking at debt for share buybacks, I think it's what you're asking, we're not philosophical about no debt on the balance sheet.
Same time, we see there's many pros and cons of having no debt on the balance sheet. It's something that we continually evaluate and discuss with our advisors and with the board. At the moment, we don't have plans to leverage up, but I would say, as always, nothing is off the table, and we'll continue to evaluate all of our options.
Thank you. Just one follow-up, maybe, on Agent Accelerator.
Yeah. Yeah, yeah.
Oh, sorry.
Yeah, yeah. Go on.
On Agent Accelerator, obviously, with what we're seeing, with new agent formation, is it fair to assume that the Agent Accelerator will grow faster in 2026 than the average of agency?
It's Agent Accelerator, low ARPA, don't get too carried away. Great to see the agent formation come back. Wouldn't expect to see that continually rising, giving its record levels. Just be cautious about that, but great to see that market open up. Marcus?
Markus Diebel with JP Morgan. Johan, just one question again on investments. Clearly, we've seen 2026 going to be a peak year again. We're going to guide for, like, 3%-5% operating profit growth. Given where the shares are and you're prioritizing, obviously, buybacks and those things, I mean, how critical is it really for you that 2026 is really sort of a one-off in terms of operating profit growth and things bounce back relatively quickly? I.e., do you feel that some investments that you clearly had in mind are now a bit more put on hold longer term? Is that the case? Just a question as well, what is the mood? How critical is to see a meaningful margin bounce already in 2027?
The second question, just in general, because you touched on this, value accretive M&A. Are we then talking about sort of, like, investments in tech? Do you feel there are some tech assets out there that you should get to? Any comments would be interesting, because it feels as if there won't be much. Just want to be really clear on this.
Yeah. I'll have a go, maybe, Ruaridh, you can fill in. Look, we when it comes to the investments, right? As we outlined, and I say it again, we have a great foundation, a great, you know, tech platform. We're doing this because we think there's more opportunity in this market. We look at the U.K. property market, our position, and what we can do together with others over the medium term. We want to step up that pace. That's what we're doing. In terms of how that's shaped, we guided to 2026, and what that means on both revenue growth and operating profit growth. We're not, you know, going down, as we said before, to be specific year by year.
Of course, you can assume that the profit growth will start aligning more to the revenue top line in the years following, right?
Yeah.
That's kind of all we can say, and as usual, you look at the business, and you look at the opportunities or sometimes challenges ahead, and you adjust after that. We're very happy with what we're doing right now, and off to a great start with it. Secondly, on M&A and maybe value, well, value creation and what kind of companies? Yes, I mean, look, there's always been a plethora of proptechs in the startup space, and now many of them come with adult AI after them. I can tell you in some conversations we've had with agents directly, some of them, of course, use AI already.
It's like, "Hey, here's a, you know, a quicker way to do admin," or whatever it is. They're start seeing some of the AI-enabled products that we actually equip them with. They're also inundated, right? They get so many pitches from that .AI and the other .AI on an ongoing basis. It's a little bit confusing, and as usual, there's a lot of promise. Again, as I said before, I mean, AI is but one thing, right? You actually gotta build it on something. It's a, you know, it's a filter and automation tool, right? It certainly doesn't provide the whole experience. That doesn't mean that there aren't interesting companies, and we keep a good eye on them. We have conversations with several of them.
For now, our re- organic growth path and with the capability we have is clearly how we operate mainly.
Maybe in this context, it's actually quite interesting. I mean, yes, we see a lot of startups, approaching agents very early, very small niche, yeah. Do you feel that the large players, the OpenAI of the world, also go directly to agents and asking them to upload and work closer together?
I-
Is there anything that you see or hear?
Nothing I would say particularly on, let's say, the big LLMs from an enterprise perspective. First of all, because our 16,000 memberships typically consist of very small, medium-sized businesses. The fact again, that many of them are interested in using tools, right? Whether it's a free user or paying GBP 20 a month. Some of them are of course, you know, more advanced and trying to figure out what's happening, either on their own or again, sold by someone else. I don't think that's a particular thing that we see, no.
Annick?
Hi, I think Annick Maas from Bernstein. The first one is on ChatGPT again. Can you tell us a bit more about how the user data is shared in between ChatGPT and yourself? At what point do you get access to the user and actually can follow them around and actually can collect the data exclusively? The second one is on OptiEdge. When agencies don't decide to upgrade, generally, why is that? Do they keep the money, they don't invest? Do they go for something else? Can you just tell us a bit through the challenges that you hear when you're meeting with agencies? Thank you.
Yeah. I'll take one.
Yeah, I'll take two.
Take two. Yeah, on ChatGPT, again, what we built is an app, and, you know, it has an endpoint, and it sits within or will sit within the ChatGPT environment, right? What the consumer will experience is to be able to do conversations that. Answers will come partly from ChatGPT, and in the case of serving up property listings that are relevant, that will come from us. What I think others have reported and what you can expect, it's a fairly simple outline, right? Yes, it's possible to find our brand there. You can find it today, but now you can find it in a slightly more organized fashion. You know, consumers will be very encouraged and already know where to go and find the full experience.
So that's kind of the outline right now. That means that the really valuable aspects of, you know, data on how people navigate and what they've done before and what they wanna do in the future will remain in the Rightmove platform. Of course, remember again, we're building a conversational interface on Rightmove, right? People already have that habit loop. It's like, "Hey, I can do all of this conversation, including complimentary information on Rightmove." Yet another reason, I think, to not worry too much about some other alternative, universe being built out. Again, interesting enough to test it. That's the way we view it.
On Optimiser Edge, you know, we actually don't want all customers on Optimiser Edge. We cater packages for all different types of customers and different types of businesses. We want them to have choice, and Optimiser Edge doesn't suit all customers. Low stock, low value, depending on where you are in the country, depending on competitiveness, funding, lots of different reasons. The strength of our account management team is knowing what products work for which customers. The way that they start the conversation isn't about which package to be on, but which packages or which products are gonna help you grow the business. Depending on that product mix, is what then will generate a recommendation of which of which package to be on.
For some, Essential is absolutely the right package to be on, and we don't expect them to move. Others, we'll see them move from Essential to Enhanced to Optimiser, and others will come straight in. That was a little bit of what I wanted to show earlier was the variance of how we see the inbound into the Optimiser Edge package. The other stat I would flag is that over 50% of our customers are choosing to purchase products above their committed levels. Again, they can engage and see value in our products without having to move up the package ladder.
For us, it's about coming back to offering a plethora of different products that suit whatever needs a business has, but also fit whatever the property market is doing, 'cause the property market, as we all know, in the U.K., can change a lot. We want products that suit them, whatever is happening in the property market. Sean?
Morning, guys. First question. Johan, I was really pleased to hear you describe ChatGPT as meaningless at the moment, given their 0.5% of your referral traffic. First question from me, from both a technical and sort of market power point of view, if it came to it, would you have confidence in blocking LLMs, not just from scraping data for training, but also for the grounding process in search? Sort of what would be the puts and takes, and how would you look at that decision?
Secondly, where you've rolled out, model capabilities, for example, in conversational search, are you finding that the major LLMs are good enough off the shelf, or are they requiring quite a bit of, fine-tuning customization to work with the data that you've got and that Rightmove, effectively, only Rightmove has?
Got it. Look, on the first one, technically, you can choose to be in an environment, you can choose not to be in an environment. I think that option is already there. Again, it's an interesting environment to test and learn in. Probably a very small meeting at the moment. It might grow, it will be relevant to be there. We'll see how that goes over time, simply, the optionality is absolutely there. I think on the conversational side that we've done ourselves, again, we operated the current version with Gemini models from Google. Again, it has the benefit of it's all, you know, very tied up through our stack.
But we have also built that capability to switch that out for literally any other large LLM. We have those relationships and conversations as well. So it's off the shelf in the sense that the general LLM is there. Now, as you know, every week or two or whatever, there's another dot something version coming out. The three things that we optimize for is it's not just cost, right? Again, that's kind of a tailwind over time because it's gonna continue to come down. It's cost, it's quality, and it's performance, right? Quality is very important, and performance as in, you know, speed and response rates. Already today, and even as a consumer, at least if you pay, right? You can see for yourself how the models act a little bit differently.
Of course, we have, you know, fantastic platform and capability in the teams to judge these all the things, right? We built this stack where we can plug and play on the side, and then we decide what we take live, and we run concurrent, what's called evaluation models. Models that evaluate the models on an ongoing basis. It'll continue to go on along that way simply. Maybe the last point: yes, of course, the generic LLM capability is one thing. The really interesting thing to create a fantastic experience and relevant experience for the consumer is to combine it with the data that we have. Again, the more people actually use this and or any other personalization features on our sites, the more tailored that experience can be.
A lot of that comes, or the vast majority of that really comes from our own platform.
Giles?
Thank you. Giles Thorne from Jefferies. Back on mortgages, please. The attributes, Johan, you used earlier to describe what pulled you towards NatWest, I'm pretty sure were things that were used to describe Nationwide when the MIP product was first developed. I'm still a little bit none the wiser as to what went wrong with the Nationwide partnership and what NatWest now solves. I wanted to push you on that a bit harder. The second thing, still on mortgages, is just to hear your latest thinking on how you solve for the problem of the broker product only appearing after a failed MIP, if that's even still the case. An update there.
Okay. Thank you. I'll leave you to judge your own wiseness, Giles. You know, as I said before, we've had a great relationship with Nationwide. What we are looking at now, where are we now? What are our own plans? What have we learned from all the data? We have selected NatWest as our partner going forward for what we think are really good reasons. On the second question, yes, the broker path to a large extent, has been because we have been focused so much on understanding the MIP path, has been focused on, okay, who doesn't get a MIP, and for what reasons?
Over time, of course, as I said before, we want to expand those choices for consumers through our segmentation, seeing what they do on the site and potentially what they're outright requesting. Some of that experimentation has been going on already, and that's gonna continue in the future.
Just to follow up, where is the remortgaging product? I think that was due to be second half of 2025. I forget the exact date, but I am pretty sure we passed the original signal around when you were going to launch that.
It's launched. It's on the site. Again, it's not the main focus. Remember that we have a lot of first-time buyers, of course, on the site. You know, for lender partners, often, you know, they want to try to get a hold of new customers. The remortgage product is absolutely there, has been there for a while, but it's sitting, as we have said before, logically connected, closer to the home valuation tools, for example, where people might be in that mode of, "Hey, I'm tracking the value of my property. That might be because I'm thinking about selling, or I'm thinking about refinancing because I'm staying." That's where that is.
Again, over time, that's an opportunity to obviously, you know, build that out further, but it's gonna come with, you know, in the right placements and as we see fit.
Right. Well, I think that's. Well, we'll squeeze you in, Andrew. Last one.
Let's squeeze in one more.
Go on.
Another one on AI and about kind of agentic, and appreciate there's a whole separate conversation about whether you'd actually want your personal agent to be searching for a house. In a future world where that could be possible from a technology perspective, what's your view about whether you'd let agents be searching on your site? How you kind of set up the technology to do it? Do you let them crawl and do whatever they want on any sites? Do you make sure you have a commercial relationship where it has to be through your flow? Like, just how are you thinking about the agentic journey?
Yeah, a little bit, let's say, early. Clearly, you know, the agentic opportunity, you know, keeps growing. Again, just what you said yourself, remember property, particularly. AI is a filter, an advanced form of a filter. Humans make decisions, right? It goes for a lot of processes. The level of filtering assistant, obviously, taking out, you know, admin tasks and so forth, big opportunity in AI. Humans need to be in the loop still for a lot of things, and even more so for other things, including this one. We'll see how that evolves over time. I really can't talk to, you know, the technology of it or who we might have a relationship with. There are interesting precedents on Amazon shutting down. I think it was Perplexity's agentic roaming around.
I don't know where that sits, right? It's something that we'll deal with over time, just like we deal with other opportunities.
I'd say thank you all for your good questions today, and wish you the best of the day.
Thanks, everyone.