Hello, and welcome to Booking Holdings third 2026 conference call. Booking Holdings would like to remind everyone that this call may contain forward-looking statements, which are made pursuant to the safe harbor provisions of the Private Securities Litigation Reform Act of 1995. These forward-looking statements are not guarantees of future performance and are subject to risks, uncertainties, and assumptions that are difficult to predict. Therefore, actual results may differ materially from those expressed, implied, or forecasted in any such forward-looking statements. Expressions of future goals or expectations and similar expressions reflecting something other than historical facts are intended to identify forward-looking statements.
For a list of factors that could cause Booking Holdings' actual results to differ materially from those described in the forward-looking statements, please refer to the safe harbor statements in the Booking Holdings earnings press release, as well as the Booking Holdings most recent filings with the Securities and Exchange Commission. Unless required by law, Booking Holdings undertakes no obligation to update publicly any forward-looking statements, whether as a result of new information, a future event, or otherwise. A copy of Booking Holdings earnings press release is available in the For Investors section of Booking Holdings' website at www.bookingholdings.com. Booking Holdings intends to use the investor relations page of its website, ir.bookingholdings.com, to disclose material information for purposes of the SEC's Regulation FD.
Booking Holdings encourages investors to monitor this website in addition to other public announcements and SEC filings, as information posted on that page could be deemed to be material information. We ask that you please limit yourself to one question so that we can give as many analysts as possible an opportunity to participate. Now, I'd like to introduce Booking Holdings speakers for this afternoon, Glenn Fogel and Ewout Steenbergen. Go ahead, gentlemen.
Good afternoon, and thank you for joining us today. I am pleased to report that our teams delivered another quarter of strong operating performance, with results exceeding the high end of our guidance across each of our key financial metrics. These results reflect the resilience of travel demand and the work our teams have done to strengthen our products and services. While travel trends can fluctuate in the near term, we remain focused on making travel easier, more connected, and more personalized. This quarter, we continued to advance several priorities that support that goal. Let me begin with AI, which is creating significant opportunities to better serve travelers and partners while strengthening our business. We believe this is the most transformational time for industries since the dawn of the Industrial Revolution. Particularly in travel distribution, AI is opening up new ways for people to discover, plan, book, and manage travel.
The technologies and the way people use them are evolving quickly. We are actively testing AI across the traveler journey, learning from real-world usage, and scaling the capabilities that create the most value for travelers and partners. We have made tangible progress across our brands over the last several months. For example, at Booking.com, we are expanding AI tools that help travelers explore destinations in greater detail at the top of the funnel. We also continue to advance our agentic capabilities, enabling customers to cancel a booking or get answers about their stay without having to contact customer service. Early testing is showing encouraging customer satisfaction results and fewer escalations to our service teams. Agoda launched its new AI Assistant, which is now live across 21 markets, and helps travelers move from an initial idea to a bookable trip using Agoda's live selection of accommodations, flights, and activities.
At Priceline, we continue to develop Penny, its AI travel assistant, and bring more of Penny's capabilities into the main Priceline experience so more customers can benefit from them. These experiences show the difference between recommending a trip and actually enabling someone to book it seamlessly. Whether it is comparing locations, photos, room choices, and prices, we can combine the ease of an AI assistant with the information, live inventory, booking capabilities, and service of our platforms. We are also excited to see the birth of new AI-native experiences beyond our existing brands. In September, we launched Lola, which helps people access unique travel, dining, and event inventory across platforms beyond our company through a conversational experience and a membership model. A few weeks later, we launched Vayo, a proactive AI personal travel agent designed to help travelers from initial planning through booking and beyond.
It can bring existing reservations from a traveler's email history into one place and use that history to create more personalized offers and suggest ways to fill gaps in an itinerary. Vayo also continuously monitors prices and stays engaged after booking, alerting travelers to changes and helping them adjust their plans along the way. These platforms are testing grounds where we can explore new AI capabilities, iterate quickly, and apply what we learn across Booking Holdings. Our established brands give us the scale to bring successful capabilities to our partner and customer bases. This combination allows us to pursue several approaches and direct resources towards the experiences that prove most useful to travelers and partners. Alongside these investments, our partnerships give us additional ways to reach travelers wherever they choose to engage. We continue to work with leading technology and AI companies, including Meta, OpenAI, Google, Anthropic, and Amazon.
OpenTable's native integration with OpenAI has become one of the brand's fastest-growing channels from a small base, and OpenTable recently debuted as the first dining partner with Meta's Muse personal agent. Agoda is live on Muse as a launch partner in Asia, connecting Muse directly with Agoda's new AI Assistant. Finally, AI is helping us operate more efficiently. We use it to help customer service teams handle routine requests, accelerate product development, and improve our internal processes. As an example, in the third quarter, customer service costs declined despite higher booking volumes, while satisfaction remained high. We are also seeing higher engineering productivity, while the cost per merge request continues to decline. These are real efficiency gains we are seeing today, and we believe AI can deliver more over time.
Now, let me turn to the value we provide to our supplier partners, as the benefits of these AI improvements that we are bringing to our traveler customers have been built on the foundational relationships we have developed with our suppliers over many years. Independent hotels and alternative accommodation partners generate the vast majority of our room nights at Booking.com. Another way to look at this is the 10 largest global hotel chains account for only a low double-digit percentage of room nights at Booking.com. Many of our partners do not have significant resources. The value we provide to them is to improve their business by enabling them to have incremental travelers from around the world at lower cost, and by providing services and expertise they may not be able to develop on their own.
For example, small and even medium-sized accommodation providers are generally not able to efficiently market themselves globally. Our brand and marketing reach, supported by years of investment in machine learning and AI, helps connect partners' offers with travelers across markets and channels. Partners benefit from that reach without upfront marketing costs, paying us only when they earn revenue from a room night that might have otherwise gone empty. To help our partners, we provide them with sophisticated data and tools to better understand travel demand patterns and improve their performance. We provide different products, like Genius, that offer them opportunities to be more competitive. Our value goes beyond just showing a property to a traveler with advice on how to display it to improve conversion. We provide payment services. We provide customer service in over 40 languages, with both AI agents and people available around the clock.
Our dedicated partner service teams are on the ground across the world, working directly with properties to help them improve their operations. Local knowledge matters. Traveler preferences, payment methods, and regulations vary widely from one market to another. Payments are a good example. Travelers want to use methods they know and trust, even when those methods are unfamiliar to the property where they are staying. Today, we can process more than 100 payment methods in over 50 currencies. That helps properties welcome guests from around the world without having to manage that complexity themselves or lose bookings because they cannot accept a traveler's preferred payment method. This combination of global reach and local expertise supports our growth in markets where we see significant opportunity. The U.S. and Asia are two such important priorities.
In the U.S., we have mentioned in the past seeing significant room to grow over the long term. We have invested in product, supply, brand, marketing, and technology to build a stronger offering over the past several years, and our progress is clear. We have outperformed the broader market in terms of room night growth for the last three years, as we have deepened direct engagement and more travelers have chosen to book multiple parts of their trip on our platform. These results reflect the consistent execution of our strategic priorities and reinforces my confidence in the opportunity ahead. Asia is another important area of growth. Its many distinct markets call for local execution, and we believe Booking.com's global reach and Agoda's local expertise position us well to deliver it.
Across the region, we continue to invest in products, payments, marketing, distribution, and supply relationships that reflect how people travel in each market. As we discussed last quarter, that can mean adapting payment options, working with local properties, or reaching customers through the social and messaging platforms that they use every day. These investments help us better serve travelers and partners while supporting our long-term growth in the region. Across all of our markets, we are focused on building relationships that extend beyond a single booking. We want to help travelers with more of their trip and give them more reasons to choose us again for the next one. Our Connected Trip efforts and Genius loyalty program are central to that goal. Travel can be fragmented. A trip may involve accommodations, flights, ground transportation, and attractions, often with different providers.
By bringing these pieces together, we can engage travelers earlier, help with more of their trip, and seek to build a stronger relationship over time. Genius adds value for travelers and encourages them to return to Booking.com, and we are seeing progress on both. Connected Trip transactions, where a traveler books more than one travel vertical with us for the same trip, grew three times faster than Booking.com's overall transactions from a relatively small base. Those travelers also tend to return more frequently. Higher-tier Genius members plan further ahead, return more consistently, and book directly at a higher rate than non-Genius travelers. Over the last 12 months, level two and level three members represented more than 30% of Booking.com's active customer base and accounted for a high 50% share of room nights, up from a year ago.
In closing, our teams executed well in the third quarter, delivering solid results while advancing the priorities I discussed today. I want to thank them for that work. The pace of change in technology is accelerating, and therefore, the pace of change in travel is also accelerating. AI is creating new ways for people to discover and plan trips, compare various travel options, and do it all in an easier way. We are building quickly to continue to help shape that future. Through it all, though, our mission remains the same: to make it easier for everyone to experience the world. We have spent years building the brands, partner relationships, and services that help us deliver on that mission. We help travelers turn ideas into trips that they can book with confidence, and our services help our partners welcome customers from around the world.
However a traveler begins their travel inspiration and planning, we see opportunities to make the whole travel process better, easier, with less frustration and anxiety. In the end, it comes down to providing more value to both sides of the travel marketplace. That is the work our teams are focused on. It is why I am confident in the future of Booking Holdings. With that, I'll turn the call over to Ewout to discuss our financial results and outlook.
Thank you, Glenn, and good afternoon, everyone. I will now review our results for the third quarter and provide our current thoughts for the fourth quarter and full year. All growth rates are on a year-over-year basis, and the reconciliation of non-GAAP to GAAP financials can be found in our earnings release. I am pleased to report that room nights, gross bookings, adjusted revenue, and adjusted EBITDA all exceeded the high end of our guidance for the third quarter. We delivered profitable growth in a complex macro and travel environment while investing in our long-term priorities. Our results reflect the fundamental strength of our global business and the continued focus and discipline of our teams. Travel demand was resilient in the quarter, although we saw some moderation towards the end of the period.
Prices increased as we saw constant currency ADRs up approximately 3% year over year, primarily driven by Europe, but also by the U.S. and Asia. We saw a slight expansion in the booking window in Europe and at the global level, while the booking window in the U.S. modestly contracted. Average length of stay increased year over year for our business outside Asia, as Asia domestic travel outpaced international travel by the widest margin of any region. Now let's turn to our third quarter results. Room nights grew 7%, up from the 5% in the second quarter, driven by an improvement in international travel from bookers in Europe and Asia, while domestic room night growth was stable at a higher level. Globally, domestic room nights grew high single digits, and international room nights grew mid-single digits.
We continued to navigate the indirect impacts of the conflict in the Middle East, with elevated airfares and reduced capacity on certain routes weighing on long-haul travel demand. Looking at room night growth by booker region, Europe grew high single digits, with domestic room nights up high single digits. Asia grew high single digits, with domestic room nights up low double digits. The U.S. grew mid-single digits, driven by domestic room night growth. We are pleased to see another consecutive quarter of direct channel growth in the U.S. Rest of world grew mid-single digits, which was stable from the second quarter, as we continue to see growth from bookers in the Middle East.
Over the past four quarters, our B2C direct mix remained stable in the mid-60% range as room nights booked through the direct channel continued to grow in line with the rest of the business, while the mix of room nights coming from direct visits to our websites and mobile apps increased year over year. LLM referrals have accelerated over the last several weeks, but still account for significantly less than 1% of Booking.com's room nights. Organic referrals increased following the release of the latest ChatGPT models, which improved citation rates and drove more traffic to our platforms. Paid referrals are also growing as we continue to partner with OpenAI by expanding into additional markets and testing and refining our approach. The mobile app mix of total room nights over the past four quarters was in the high 50% range, up from the mid-50% range a year ago.
The mix of Booking.com room nights booked by travelers in Genius levels two and three was in the high 50% range, which also increased from a year ago. What is particularly encouraging is the behavior of travelers who engage with Booking.com across all three dimensions: booking directly, using our mobile app, and participating in Genius levels two and three. Room nights booked by this cohort grew meaningfully faster than those booked by other travelers, both in the third quarter and year to date. These trends reinforce our confidence in the investments we are making to deepen our direct relationships with travelers, and we believe our broader selection of accommodations, flights, and attractions, together with progress in payments, the Connected Trip, and more personalized and proactive AI-powered experiences will give travelers even more reasons to choose us and return for future trips.
Alternative accommodations represented approximately 37% of Booking.com's room nights, up about 1 percentage point from the third quarter of 2025, with room night growth of 7%, improving from 4% growth in the second quarter, as we saw an acceleration across all of our major regions. We also saw alternative accommodation listings growth accelerate to 9%, which is an encouraging signal for our future growth in this category. In our other travel verticals, flight tickets increased 19% year-over-year, a strong acceleration from the second quarter and ahead of the broader market, despite continued pressure from elevated ticket prices and reduced capacity on certain routes. Attraction tickets grew 24% year-over-year, also an acceleration from the second quarter. Importantly, Connected Trip transactions growth accelerated to almost 20% year-over-year and represented a low double-digit mix of Booking.com's total transactions.
This share has increased substantially over the past few years as we have expanded our supply across travel verticals beyond accommodations. We expect further increases as we deliver more personalized and gen AI-powered experiences and bring these offerings together more seamlessly through our Genius loyalty program. Our data shows that travelers who book with us across multiple travel verticals return more frequently, reinforcing our confidence in the Connected Trip opportunity. Our total merchant gross bookings represented approximately 75% of total gross bookings, up about 3 percentage points year-over-year. Our merchant payments platform is a key enabler of our Connected Trip vision, providing travelers with a more seamless and flexible booking experience, helping partners accept customers' preferred payment methods, and generating incremental contribution margin dollars for our business. Total gross bookings increased 10% year-over-year, or approximately 11% on a constant currency basis.
Constant currency gross bookings growth was approximately 4 percentage points higher than room night growth, primarily reflecting the higher constant currency ADRs and the contribution from flights and other verticals. Adjusted revenue increased 9% year-over-year, or approximately 10% on a constant currency basis. Adjusted revenue as a percentage of gross bookings was about 15 basis points lower year-over-year due to a higher mix of flight bookings and the timing of bookings versus travel, partially offset by higher revenues from facilitating payments. Marketing expense increased 11% year-over-year, slightly faster than gross bookings, driven primarily by faster growth in B2B affiliate marketing expenses. Excluding our B2B affiliate business, our B2C business had marketing leverage in the quarter as we continue to aim to grow our top-line metrics faster than marketing investments.
Adjusted sales and other expenses were 2.0% of gross bookings and were a source of leverage. Efficiencies in third-party customer service, supported by AI-enabled tools, more than offset higher payment expenses associated with our increasing merchant mix. Disciplined cost management continued to support operating leverage in the quarter. Adjusted fixed operating expenses increased 6% year-over-year, reflecting actions we have taken to manage our cost base. As our use of gen AI scales, we are applying the same cost discipline, including price caps on higher-cost models and model routing simpler tasks to more cost-effective models. For example, at Agoda, our blended cost per million AI tokens declined by more than 25% between June and September. For the company overall, developer productivity continued to improve, while cost per merge request declined double digits. These gains are helping us deliver new capabilities faster, including several Glenn highlighted in his opening remarks.
Adjusted EBITDA of approximately $4.6 billion grew 9% year-over-year. We delivered 9% adjusted revenue growth while expanding our adjusted EBITDA margin slightly to 47.0%, including stock-based compensation as an expense, a powerful combination of growth and profitability that underscores the strength of our business model and disciplined execution. Adjusted EPS of $4.66 per share increased 17% year-over-year, faster than adjusted EBITDA growth, helped by a 7.5% reduction in our average share count. We continue to execute on our Transformation Program and remain on track to achieve the full $650 million in annual run rate savings in 2027. We incurred approximately $23 million of transformation costs in the third quarter, the majority of which were excluded from our adjusted results. Now on to our cash and liquidity position.
Our third quarter ending cash and investments balance was $16.2 billion compared to $17.7 billion at the end of the second quarter. During the quarter, we generated free cash flow of $1.1 billion, which was impacted by a working capital decrease driven by a $3.6 billion seasonal unwind of deferred merchant bookings. We also returned $2.4 billion to shareholders, including $2.1 billion of share repurchases. During the first nine months of 2026, we returned $10.5 billion of capital to shareholders, including $9.5 billion of share repurchases. We remain committed to a disciplined capital allocation framework, investing behind the highest return growth opportunities across our business, while also actively returning capital to shareholders and maintaining a strong balance sheet. Moving to our thoughts for the fourth quarter, the travel environment is mixed.
As I mentioned earlier, we saw some moderation in global demand trends toward the end of the third quarter, driven by international travel. Domestic travel continues to hold up as we are seeing travelers take shorter trips and choose destinations closer to home. In October, a number of external factors have also affected demand, including disruptions to travel to and from the Middle East, the conflict's broader indirect effects such as elevated airfares, and localized unrest in certain European cities. Our guidance reflects these near-term disruptions and assumes that both the direct and indirect impacts of the Middle East conflict on travel persist through the end of the year. Our guidance also assumes recent FX rates for the remainder of the year, including the euro-U.S. dollar exchange rate at 1.12.
We estimate changes in FX will negatively impact our fourth quarter reported U.S. dollar growth rates by about 2 percentage points. For the full year, we estimate changes in FX will positively impact reported growth rates by about 1 percentage point for gross bookings and about half a percentage point for adjusted revenue. We currently expect fourth quarter room nights to increase between 1% and 3%, reflecting the toughest prior year comparison of 2026 and the uncertainty around the impact of recent events on travel demand. We currently expect gross bookings to increase between 2% and 4%, or between 4% and 6% on a constant currency basis, which is higher than room night growth due to higher constant currency ADRs and flight ticket growth.
We currently expect adjusted revenue to increase between 0% and 2%, or between 2% and 4% on a constant currency basis, slower than gross bookings due to flight bookings mix and differences in the timing of bookings versus travel, partially offset by higher revenue from payments. We currently expect adjusted EBITDA to be about flat year-over-year, or up about 2% on a constant currency basis, and to be impacted by increased planned investments behind our long-term strategic priorities, including expanding capabilities across non-accommodation verticals such as flights, attractions, and dining, advancing our AI capabilities, and enhancing our underlying technology infrastructure and platforms. We expect these investments to be partially offset by marketing leverage. For the full year, on a reported basis, we continue to expect gross bookings, adjusted revenue, and adjusted EBITDA to be up high single digits and adjusted EPS to be up low to mid-teens.
At the high end, this outlook is broadly in line with our long-term constant currency growth ambition of at least 8% for gross bookings and adjusted revenue, and 15% for adjusted EPS. In conclusion, our third quarter results demonstrated the resilience of our global platform and the disciplined execution of our teams in a dynamic environment. We continue to focus on key areas that we believe can further strengthen our long-term growth outlook, including expansion in key markets, deepening traveler relationships through Connected Trip and Genius, creating greater value for our partners, and advancing our approach to AI. As agentic AI evolves, we're taking a multi-pronged approach. First, building our own AI agent-powered experiences. Second, participating in new distribution channels that bring incremental demand at attractive economics. And third, using AI to improve customer service, developer productivity, and operating efficiency.
We believe Booking Holdings is well-positioned to win as the landscape evolves. Our confidence is grounded in the capabilities that matter when inspiration turns into an actual trip: extensive supply, real-time price and availability, trusted brands, attractive loyalty programs, payments and service capabilities, and longstanding relationships with travelers and partners. We believe these will create economic value both within our own platforms and through emerging channels, supporting sustained growth and attractive returns over time. With that, we'll now take your questions. Operator, will you please open the lines?
Thank you. If you would like to ask a question, please press star, then the number one on your telephone keypad. If you would like to withdraw your question, again, press star one. We ask that you please limit yourself to one question so that we give as many analysts as possible an opportunity to participate. Your first question comes from the line of Doug Anmuth with JPMorgan. Please go ahead.
Thanks for taking the questions. I have two. First on the tangible AI progress that you are seeing in brands, obviously good to see the customer service cost efficiencies. Is there anything that you can share in terms of some of the top-line benefits perhaps that you are seeing around either conversion rates or frequencies or just other ways to quantify? On those accelerating AI referrals in recent weeks, are you seeing any differences in some of the newer agents, like Muse and Grok, relative to what you saw previously with the other LLMs? Thank you.
Hello, Doug. Good evening. This is Ewout. First, on some of the tangible AI progress. There are four key categories where we see some positive early indicators. Let me explain that and give a couple of examples around that. First, we are seeing higher engagement. Examples here are Booking.com's daily AI search visitors are up 10 times with trip search in the third quarter compared to the second quarter. A second example here is our new Vayo platform users are connecting their email, and we are seeing strong conversational engagements with Vayo at the moment. Second area is higher conversion. A couple of examples again. Priceline's Penny and Agoda's AI Assistant users have converted at approximately twice the rate of non-users. Third area, higher booking value. We see Agoda's AI Assistant users have an average booking value of approximately 20% higher than non-users.
The fourth category is incrementality. Again, some examples. OpenTable's native integration with OpenAI has become one of the brand's fastest-growing channels and had, in September, 190,000 seated covers. OpenTable's agentic interactions with Muse and other horizontal agents have brought a significant number of completely new customers to our platform. So four key areas where we see improvements and really positive indicators: higher engagement, higher conversion, higher booking value, and incrementality. It is early days, but we are encouraged by the momentum, and we really think these are really nice positive uplifts on these metrics. In terms of your question around referrals from LLMs versus the newer horizontal agents, what we are seeing that this traffic is still very small in aggregate. Nothing is really jumping out there in terms of one that is growing faster than another. So it is still very small.
Yes, we have seen some acceleration in the recent weeks, but it is still significantly below 1%. I also would like to point out that the vast majority of the traffic is coming to us, and that is based on third-party data sources. We are not missing out on the traffic, but it is still really small.
Your next question comes from the line of Mark Mahaney with Evercore ISI. Please go ahead.
Thanks. Two questions, please. I know you have got this partnership with Muse with Agoda in Asia, and just talk a little bit about the strategy of, you have got three sort of global brands. Why work with Muse with Agoda versus one of the others? Then the marketing deleverage in B2B, and I think that is been the case for a quarter or two or three. I think you are leaning into this as a channel. Could you talk about any signs of traction, of success, more momentum you are getting just in the B2B segment? Thank you very much.
Hi, Mark. Why don't I take the first one, and I will let Ewout take the second one. It is interesting which agreements with whom we get done first and priorities, et cetera. Look, I think the important thing to recognize is that we are working with everybody all the time. Some things are going to happen faster than others. Some things are going to get completed faster than others. You should not think just because we have Agoda getting the launch in Singapore is somehow more important than what we end up in the U.S. with Muse or anyone else. Please don't misinterpret that. I have confidence that we will have great relationships with all the players, and not just the ones that you know of now, because we know there are a lot more coming down the road.
Your next question comes.
Mark, with respect to your second question, with respect to B2B and marketing deleverage, let me point out the following. We are pleased, of course, with third quarter marketing leverage for B2C, and also, of course, pleased with the growth of our B2B business volumes. Specifically, your question about B2B, that is a business that is getting more focus and attention internally. I think you know about our plans that we are launching on January 1st, a combined B2B business, so you will also see then more details about their numbers, because most likely we will break that out going forward. What is happening there is if the business is growing and they are reporting more revenues, usually the commission structures are based on a rev share, therefore also the rev share and the commissions are going up, and that is flowing through the marketing line.
That is for a good reason, because if that is going up, it means that from an economics perspective, it is also very healthy for the company.
Your next question comes from the line of Matt Cost with Morgan Stanley. Please go ahead.
Hi, thanks for taking the question. I am sure you have seen data out there showing that a lot of the commercial activity that is happening on these early agents is skewing towards travel. I guess, to the extent that you are seeing that come to your brands, should we view that as something that is just a convenience for people who are going to book travel already, or is there an aspect here where by removing friction from a very complex purchase process, there may actually be incremental demand that is engendered by these tools? Thank you.
That is an interesting thing, and we are seeing the data you are mentioning. Lots of people using these personal agents or any form of AI to do travel discovery. It is easy, it is interesting, people like to do travel discovery throughout all the time. It does not end up with a lot of bookings, as Ewout mentioned. The referrals numbers are pretty darn small, and we believe we are getting a good number of the total number out there. What does it mean for us, though? Over time, it may end up being another perhaps significant way that we get business, the way we have gotten business in the past from other people who had a front door of discovery.
It may be making it easier for people, less friction to try and learn what they need to do, so perhaps we end up with things that are actually going to convert at a higher rate. We believe it is important enough that we are coming out with things like Vayo, like Lola, and all our brands coming out with things too, so people can come to us first. Look, what I really want to do is complete our Connected Trip using AI to make it easier, provide more value to the consumer, so that when they are thinking of travel, they are always wanting to come to us first.
Now, we know we are not going to get everyone, but we absolutely believe that we should, with the technology capabilities that we have, with the capital we have, with the resources we have, with the relationships with the suppliers, so we can help provide more value to the consumer. Of course, that customer service that is so important, being able to make people able to trust that when they come to us, they are going to be able to have the best experience in their travel. Those type of things are really important. I believe this is actually a wonderful thing, that there could be much more of a fragmentation of the way people are beginning their discovery, so as to enables us to actually be able to obtain customers at a lower cost. We could see that happen. Ewout, anything you want to add to that?
Just to say one more thing is strategically, of course, these new channels are important for us because it gives us an opportunity for diversification of third-party channels, and that ultimately will drive an economic advantage to us. We have been talking about that in the past, but I think that is already really showing up in practice, because clearly there are many AI platforms now, many AI tools, agentic tools out there. They are still figuring out their economic model. They haven't really come to a conclusion there because the competition is very high. We get the vast majority of the traffic because they like to connect to aggregators. That is much better for them, and that is good for us. From an economics perspective, the diversification looks like it is going to be a positive for us at the end.
Your next question comes from the line of Alex Brignall with Redburn Atlantic. Please go ahead.
Thank you very much. I will ask a follow-on on that previous question. It is something I have asked in the past about fragmentation. You are obviously seeing some B2C leverage. If you could talk about the components of that would be very helpful. Then, sat over here in Europe, we have obviously, and specifically continental Europe, seen the DMA impact on Google's search results, which have thrown a bit of a spanner in what they can display when people search for travel. I wonder if you could give us some thoughts on that and how it might affect your numbers. Thank you.
Alex, with respect to our B2C leverage, nothing is different than what we have always done. We are managing that in a disciplined way. We see two advantages there flowing through. One is we are always trying to drive up higher ROIs in our paid channels, and then also what we are seeing is that the direct traffic is up at the same time. Both in combination are helping us to achieve B2C marketing leverage in the third quarter. Also, we are looking at the possibility to deliver on marketing leverage in the fourth quarter as well. This is always our standard, what we are going after. Glenn, do you want to take the question about Europe, the DMA, and Google?
Yeah, I'll talk a little bit about that. Right now, we don't see any great advantage from this change. Obviously, it's early, et cetera. We haven't seen anything in terms of numbers change. I know some people have been talking about that, "Oh, it's going to be such an advantage for the OTAs." I hope it is. I haven't seen it yet. We'll see.
What's interesting, though, since you mentioned DMA, it gives me an opportunity to talk about something I don't see a lot of people writing about, and that talks about the regulatory framework in many parts of the world, particularly in Europe, and how these horizontal personal agents are going to be able to do things or not, what sort of access they're going to be allowed to have in terms of personal information, what the requirements are going to be, what's going to be allowed in being a commerce platform along with being an advertising platform. It's very complex. There's a lot of regulatory stuff out there, whether it be the DMA or the EU AI Act or the DSA or the VLOP laws, or you get into the U.K., and I saw something by one of the senior executives, I think it was out of the Treasury.
It was regarding the FCA, I think, in terms of being able to allow to do monetary things in news or not. I think there's parts of the world where the adoption of these horizontal personal agents may be much slower than is anticipated, and we'll see what that may mean for ourselves. We don't know. But it is interesting that I don't see a lot of people talking about that.
Your next question comes from the line of Justin Post with Bank of America. Please go ahead.
Great. Thanks. A couple on agentic. In addition to economics, one of the concerns is they might be preferencing direct sites versus OTAs. Are you seeing any evidence of that, and what can you do to really improve your relevance in AI results? Then second, just thinking about the cost increases in Q4, there are some questions on how you are thinking about investment in 2027 and what that could do to your algorithm. Any help with that would be very helpful. Thank you.
I will talk a little bit about the first question. I will let Ewout talk about investments going forward. It is interesting you say that, because I have heard some of that, but I see other things too where the OTAs are actually getting a greater advantage right now. The real story is, and this is the way we believe we need to play it, you have to provide more value, whether it be to a human traveler or to the agent, such that we want that agent or that person to be using us, looking at our stuff, going through us, transacting with us. If we are able to do that, in the long run, we will have a competitive advantage, and that is really the way it is.
Now, one of the things that we provide is to the small independents, and Ewout mentioned a little bit about that, how a very small percentage of our business comes from the big hotels, and so much of it is the small, medium-sized ones who do not have the resources, who absolutely are afraid that in the new world, they will not be seen at all. But that is okay, because we are there to provide them what they need. That is what we have been doing for the almost 27 years I have been at this company, is providing value to our partners, being able for them to get incremental travelers because they could not, whether it was 20 years ago, go online and have a website, or then they do not have an app, or now it is, "How do I show up in an LLM?
How is my agentic horizontal going to find me, see me, or whatever?" We are there to provide it for them, and we do it at a risk-free rate. Essentially, they do not have to pay anything unless they get a head in the bed, and that is what we are able to do for them. So, I absolutely believe that we have an advantage. This is the great time for us now, being able to use our technologies, all these things to be able to provide even more value to both the travelers and the partners, and I see the future is very bright. So, Ewout, do you want to talk about investments needed to achieve all this?
Yes, Justin. Let me give some commentary about the fourth quarter, and you also asked about 2027. First, about the fourth quarter. Our philosophy is to make disciplined investments for the long term of the company while also protecting profitability in the near term. From a big-picture perspective, I think everyone knows that we have already industry-leading EBITDA and EBITDA margins, and that is on a fully loaded basis, so we are not capitalizing a lot. We have a fully loaded expense base, including stock-based compensation. We do not understand how people are pulling that out. That is a real cost of running your business. At the same time, we also have strong execution and discipline regarding the delivery on our productivity programs, efficiency targets, operating leverage, and that is then offset by increased investments in strategic initiatives, and those were planned in the fourth quarter at a higher level.
But obviously, we have attractive business cases underneath. I also would like to point out that our full-year EBITDA margins, we expect to be stable. With respect to 2027, it is a bit too early to provide commentary. We will do that in February with our fourth-quarter results. But what I can say is that we remain confident with respect to our long-term growth outlook, and we will stay disciplined how we are planning to invest in the business also next year and in the future. Maybe one thing to throw out for 2027, and this is not related to cost, but more on the commercial side, it is a very early directional indicator, so it is still subject to change, like cancellations and others.
But if you look at our on-the-book statistics for the first half of 2027, that actually looks healthy, so that gives us confidence in the outlook for next year.
Your next question comes from the line of Ron Josey with Citi. Please go ahead.
Hey, thanks for taking the question. Glenn, I wanted to go back to your comments on just applying the learnings from Lola and Vayo and using Booking's platforms as a testing ground. Talk to us about the metrics insights you are looking for from those tools and the other AI tools that you have before applying the learnings. Then we just talked about it, but I would love to hear your thoughts on just the value of the supply that Booking has as agents become a larger partner, potentially larger partner, or a new front door for the internet overall. Just the value of Booking supply. Thank you.
Hi, Ron. In terms of what we use in terms of learnings and why we started up two greenfield operations to be able to come up with a new agentic way to do our travel business in two different ways, two different companies doing different. I mean, they are not the same. You are probably reading about what they are doing. The way we do this is seeing what customers like, what is converting well, what people are signing up for, whether or not, for example, at Vayo, will they take an email? Will they have the whole email history in the thing or not? We have learned that actually, yes, most will, which is really good to know. That is how we do it. In terms of the actual specific metrics, I am not going to go around and get into the details of how we are going to do it.
But obviously, when we see things that are working, that is such a value to bring to the much bigger companies, Booking, Agoda, Priceline, and being able to transfer that so that they can then take the advantage of doing that. It really is a wonderful place to be when you have the capabilities, the capital, the resources, the number of people that you can afford to do this type of agile, creating new things and being able to test it out, and then test bed type stuff, come bring it back when it is working or do not bother, and it did not hurt you because you tried it on a much bigger platform and it was not working. That is on that one. In terms of supply, what is the value? Make sure I got you right, Ron.
If you are asking what value do we provide to the supplier in a new AI world? Was that the question?
More like the other way around. When we do our searches on Muse, we see Booking show up, whether we are looking for a hotel in Panama City, Panama, or in Paris, for example. Booking inherently is always there, and I think that is because of the supply, but I wanted to get your thoughts there.
Why are we showing up more there than others or that? That is a hard thing to say. I think we have a lot of theories about why the models are choosing us. We believe it has a lot to do with the level of the content, the reliability of it, the value that the way the models go look through and see what value this content provides. Truth is, though, there is no handbook out, how do you show up more in an LLM? How do you show up more in a personal agent or not? In the end, what we really want is what I said earlier, though, and this is being able to provide the more value.
The agents, when agent to agent, our agent is talking to the personal agent, the horizontal, and really being instantly giving all the reasons it is better, and it is not just price, of course. All the elements that we have talked about over the last 40 minutes or so about what we provide to the traveler and to the partner, that that is the best place to do this travel in such a way that the horizontal agent recognizes it and is recommending or saying, "Do it here because this is best." That is what we want to achieve.
Your next question comes—
If I can add a few elements there, Ron. I think brand credibility is a very important element. I think that is why those horizontal agents want to work with us, because it gives them instant credibility being associated with our brand. As I said before, they like to partner with aggregators. It is going to be very expensive for them to connect to 4.7 million properties directly and get all the information about real-time data, pricing, availability, so they can get that through us and immediately have also the possibility in the future of a potential monetization model with us. Those are a couple of reasons why we think that we are showing so much up on those platforms.
Just to add to Ewout's point, if these horizontal agents do go down the route of some type of monetization that is not dissimilar to what we have been doing over the last two decades, we will be even in a better position.
Your next question comes from the line of Benjamin Black with Deutsche Bank. Please go ahead.
Great. Thank you for taking my question. Genius has become a much larger part of your transaction mix. Can you dig into how you are thinking about the next evolution of the program and how you can deepen loyalty without maybe becoming more reliant on supplier-funded discounts? Then maybe a follow-up on the eight, 15 framework. As AI begins to drive real efficiencies across customer service, across engineering, across other parts of the cost base, does 15 still seem like the right long-term kind of EPS algorithm, or is there a scenario where AI-driven operating leverage allows you to outperform that framework without even driving faster top-line growth? Thank you.
Nice, Benjamin. I will take the first, I will let Ewout take the second there. Do me a favor, repeat that first one again. It gives me a little more time to think exactly how I want to answer it.
Yeah. It is really about Genius and how you are thinking about evolving it from here. It has already become a much larger part of your transaction mix.
Yeah. All right. Here is the deal. I have kind of teased this out a little bit in the past, I just want to be very careful what I say. We are going to improve Genius. It is going to be an even better program down the road. I absolutely believe, and I said it earlier, I said it several times now, the issue, the importance of providing value, providing value to that traveler, and it is also value to the partner. Genius works right now very well because the partner is funding a lot of it and is getting a good ROI out of that. Of course, the traveler is getting a great deal too because they are getting a lower cost for their holiday. It works both ways. There are a lot of other ways we can improve upon that tremendously.
There are other things that travelers want, and maybe the partner would be willing to give it up either at no cost, because there is very little marginal cost to them, or perhaps they would be willing to do it at a lesser cost that we could help work with. These are ways we have to think about down the road is, how do we continue to make sure that we are giving more value, and in a way that travelers who use us get even more value because they use us, loyalty, spend more with us, stay with us more, do things more with us, get more benefit, continue to do that in a way that it creates a flywheel effect and continues to grow.
Now, not going to give away when we're going to bring out what or what they're actually going to be, but I absolutely recognize the importance of continuing to come up with a differentiating value that enables us to continue to show up as the better way for the traveler to use us and for the partner to participate. Ewout, I think the question about the algo is yours.
Yeah, Ben, with respect to the eight, 15 and the benefits from AI that might flow into that, I think it is a little bit too early to tell you exactly how we will see that changing our long-term algorithm. What I can say is that we're already moving from experimentation to measurable operating value across the company coming from AI. So, in a couple of areas. First, we are fundamentally improving the travel experience for our traveler customers. Think about engagement, think about conversion, think about booking value, think about incrementality, and some of the examples that I gave earlier during this call. We have been speaking before about the benefits, for example, in customer service, where the average costs per booking are coming down and the customer satisfaction remains high.
Also in our engineering groups, where the average costs per merge request are coming down in a significant way. The third area is our headcount growth has been very minimal over the last two years, and that has been mostly in lower-cost locations like our Center of Excellence in India, where we also can find great talent. So, we're pleased that we're already seeing some tangible results, but it is a little bit too early to exactly determine in which way that will impact our long-term algorithm.
We have time for one last question, and that question comes from the line of Meredith Jensen with HSBC. Please go ahead.
Good afternoon. Thank you. I just had one question a little bit related to the last, but potentially a smaller piece to the pie. But if you might talk about a little bit of the changes in the distribution landscape and how it might open some new doors for Booking and hotel suppliers in ways that they could collaborate, particularly around loyalty programs and guest recognition. If you could, I know, Glenn, you've spoken a little bit about this in the past in interviews, and I would just love some updated thoughts. Thank you.
Meredith, so I make sure I understand your question, you'd like to know a little bit more about how, in terms of recognition for the suppliers in the process, or were you talking more for the traveler?
Just if there's the loyalty program connectivity between some of the hotel programs to, if they book through Muse and Booking is the connector, how that kind of could play out over time because everything's moving so quickly.
It is. It absolutely is. And that's one of those interesting questions in terms of networks and how do you create a loyalty program so you have more and more people who want to participate in it, gives more value. We all know the value of a network, and the more nodes you have, the exponential increase in the value of that network. Absolutely talking with hoteliers, talking with every travel player and non-hotelier players, non-travel players, people outside the travel ecosystem, and figure out ways, again, in these types of loyalty programs that you can actually, though, give more value to both the person who is participating, the traveler, and anyone who's on the other side supplying some type of service or product.
We absolutely believe that is an important part of helping build out and ensure that the Genius program remains one of the most valuable programs there is, and the reason people will want to come. Again, I do not want to give away too much of what we are building down the road, but we absolutely understand the need to continue to provide more value to both sides of that network. Is that helpful?
Yes. Thank you so much.
Thank you. Well, I guess that was our last question, so let me conclude. First, as always, I want to thank our employees for their dedication and strong execution. I want to thank our shareholders for their continued support. Of course, I have got to thank the travelers and the partners who place their trust in us every day. We together, we are so excited about the role AI is going to play in travel, and I am absolutely encouraged by the progress we are making across our strategic priorities. Look, we are going to continue to navigate this near-term uncertainty with the discipline we have always had, and we are going to be investing thoughtfully in the opportunities ahead.
Our people and the capabilities we continue to develop gives us the confidence in our ability to continue to play a leading role in shaping the future of travel and delivering lasting value for our shareholders. Thank you very much, and good night.
Ladies and gentlemen, this does conclude today's conference call. Thank you for your participation, and you may now disconnect.