Hello everyone and welcome to the Trainline HY 2026 results. My name is Emily and I'll be coordinating your call today. After the presentation, you will have the opportunity to ask any questions, which you can do so by pressing the star followed by the number one on your telephone keypad. I will now hand over to CEO Jody Ford to begin. Please go ahead.
Good morning everyone. Thank you for joining us today for our half-year results presentation. I'm Jody Ford, CEO of Trainline, and I'm joined by Pete Wood, our CFO. Let's first go through the disclaimer. Onto the agenda for today. I'll give an introduction, briefly discussing the progress we've made in the first half. I'll recap on the opportunity ahead and update on the regulatory backdrop in the U.K. Pete will talk you through our financial performance. I'll then update you on how we are progressing against our strategic priorities and finish with an overview of our Digital PayAsYouGo trial, which recently launched in the U.K. After that, we'll hand back to the operator for questions. As a reminder, we are Europe's number one rail app, delivering a market-leading user experience for our customers. We aggregate all major carriers and fares, offering a comprehensive range of value-saving products and features.
We focus on making the booking experience as seamless as possible, and we use machine learning and AI to supercharge the experience to get customers from A to B. It is our combination of value, convenience, and innovation that sets us apart, and that is evident in our scale today, with far more app downloads than any of our peers. The strength of our customer proposition is reflected across the group, with all three of our business units leaders in their respective markets. In U.K. consumer, we have the number one travel app in the U.K. In international consumer, we are the largest rail aggregator in Europe. In Trainline Solutions, we have the leading B2B rail platform across the U.K. and Europe, which now generates over GBP 1 billion of annualized net ticket sales. Looking ahead, we see significant headroom to scale all three business units.
In U.K. consumer, we will deepen our competitive moat while increasing the lifetime value of our 18 million strong customer base. In international consumer, we will deploy our proven aggregation playbook across France, Italy, and Spain. By 2030, these markets together should represent a TAM of around EUR 23 billion, including EUR 12 billion generated on aggregated high-speed routes. In Trainline Solutions, we will scale into the EUR 6 billion business travel opportunity across rail. Our performance in the first half demonstrates the progress we are making against this headroom opportunity. In the U.K., we delivered robust growth, reflecting continued strength in leisure travel alongside the ongoing digitization of rail ticketing. In international, we delivered positive early momentum on the French Southeast high-speed network, with sales up 34% following Trenitalia's expansion of services over the summer.
In Trainline Solutions, B2B sales grew strongly, particularly in Europe, with international B2B distribution up 55%. Operating leverage amplified our top-line growth, as did the benefits of our cost optimization exercise last year, driving a 14% increase in EBITDA. As a result, we have to date increased our profitability guidance for the year, as Pete will discuss shortly. This follows our announcement in September of an enhanced share buyback program, underpinned by our strong cash generation. The enhanced buyback implies we will have repurchased GBP 350 million of shares over three years. That's around a third of our market cap. Before I hand over to Pete, let me update you on the regulatory backdrop in the U.K. This morning, the government published the outcome of its consultation on the railways bill, with primary legislation to follow later today.
This will allow for the establishment of GBR as an organization, as well as the appointment of its key leaders. Since our full-year presentation in May, we have maintained an assertive stance with government, pushing them to deliver on the commitment to an open, fair, and competitive future retail market. While in parallel, we have sought to resolve existing examples where train operating companies self-preference today. Self-preferencing is where TOCs offer features within their apps, but we are prevented from offering and market them in ways that we are not allowed to do. This undermines fair and open competition. We've consistently put forward the case that these practices be rectified. We're making progress, as you can see on the left-hand side of this slide.
Previously, Trainline was prevented from offering some temporary fares that could be found on operator websites, and we were blocked from advertising in almost all stations and trains. Following our sustained engagement, the government confirmed earlier this year on both fronts independent retailers should not be discriminated against. This is a clear step forward. However, we still face blockers and obstacles when it comes to advertising at the station. Furthermore, notable examples of self-preferencing continue to persist. We are prevented from offering train operator loyalty schemes within our app, and we are unable to provide automated Delay Repay, a major pain point for our customers. We continue to engage with government stakeholders and the wider industry to remove these restrictions and, in turn, level the playing field. At the same time, we are innovating to extend our market-leading user experience and cement the loyalty and engagement of our customers.
With that, I'll hand over to Pete to talk through our financial performance.
Thanks, Jody, and good morning, everyone. Before I step into the financial performance for the group, let's briefly unpack the performance of each of our business units. Starting first with U.K. consumer, net ticket sales grew 8% to GBP 2.1 billion, reflecting continued strength in leisure travel and ongoing digitization of rail ticketing. Growth was supported by further market recovery, increased industry rail fares, and lapping strikes the year before. As anticipated, growth was partly offset by the first phase of Transport for London's Project Oval expansion, which launched in February and will expand further in H2. Turning next to international, where net ticket sales grew 2% year- on- year to GBP 594 million. Growth was led by domestic customers, particularly on newly aggregated routes such as the high-speed network in Southeast France.
This was offset by a 2% decline in foreign travel sales, primarily given changes to Google's search results page and a leveling off of inbound demand from the U.S. for European rail travel. I'll now step through the underlying drivers in more detail. This slide breaks down the different market segments across our international business, with each segment including domestic and foreign travel sales. You can see that the growth rates vary, reflecting how we are actively managing marketing investment and prioritizing routes with carrier competition. Turning first to Spain and Southeast France, we grew 11% across both markets as we positioned ourselves as the aggregator of choice, partly offset by downward pressures on Spanish rail fares. These markets now represent 22% of international net ticket sales. Elsewhere in France and in Italy, growth was more modest at 3%.
Within these markets, which account for two-thirds of international net ticket sales, we continue to manage marketing spend as we await further carrier competition. In Italy, that is expected to happen from early 2027, with SNCF granted slots to run high-speed rail services last month. In the rest of France, carrier competition is set to expand shortly thereafter, which Jody will discuss later. Germany and the rest of Europe were down 16%. While these markets represent longer-term growth opportunities for Trainline, we are actively prioritizing the markets that have liberalized or are set to liberalize. While changes to Google's search results page remained a drag on growth in H1, we are seeing encouraging early signs of traffic building from generative engines.
We are the number one cited rail app in ChatGPT across almost all our core markets, and we are leading in citations from Google's AI overview module, significantly ahead of other rail aggregators. As a result, sales from generative engines have grown exponentially, increasing 13-fold since Q3 last year, albeit from a low base. Let's move next to Trainline Solutions. Net ticket sales grew 18% in the half to GBP 529 million. Growth was led by B2B distribution and our fastest-growing subsegment, up 36%. This reflected growing demand from corporates shifting to rail travel, as well as our enabling travel management companies to scale their ticket sales in Europe, including the likes of Navan and SAP Concur. Jody will talk shortly in more detail about the strong momentum this business is generating. Bringing this together, top-line growth for the group was towards the upper end of our full-year expectations.
Group net ticket sales grew 8% to GBP 3.2 billion. Revenue grew 2% to GBP 235 million, with growth slower than net ticket sales given the previously announced reduction in the U.K. commission rate. Gross profit was up 6% to GBP 193 million, with growth outpacing revenue given lower cost of sales. Turning to costs, we delivered a reduction of GBP 11 million across cost of sales and other admin expenses, more than offsetting the impacts of the commission rate cut in the U.K. Cost of sales were down GBP 6 million, reflecting a reduction in the industry costs in the U.K. Other admin costs reduced by GBP 5 million, given the successful execution of our cost optimization plan in H2 last year. These savings enabled Trainline to deploy more marketing to Southeast France as carrier competition expanded. Our profit grew strongly, outpacing net ticket sales, with adjusted EBITDA up 14% to GBP 93 million.
This tracked above our previously stated guidance range for the year of between 6% and 9%, and it translated into strong earnings growth, with adjusted earnings per share up 27% to GBP 0.126. Underlying free cash flow generation in the half was GBP 79 million. The strength of our cash generation underpins our enhanced share buyback program of up to GBP 150 million, which we announced in September. This is our fourth consecutive program. Over the last two years, we have bought back GBP 215 million worth of shares, equivalent to 15% of shares issued at IPO. With the addition of our enhanced program, it would equate to GBP 350 million of capital return to our shareholders over a three-year period, reflecting our confidence in our outlook and the strength of our business. Altogether, I'm pleased with our performance in the first half, particularly our strong earnings growth and cash generation.
Looking ahead to the full year, we continue to expect net ticket sales growth of 6%-9% and revenue growth of 0%-3%. Given our profitability performance in the first half, we now expect adjusted EBITDA to grow between 10% and 13% for the full year, above our originally stated guidance of 6%-9%. Thank you, and I'll now hand back to Jody.
Thanks, Pete. Let's now talk about the progress we're making against our strategic priorities. Starting with our U.K. consumer business. As the U.K.'s number one travel app, our scale and user experience is unmatched. This provides Trainline with a competitive moat, which we are deepening, strengthening the loyalty and engagement of our customer base. First, we'll unlock value for customers through products like SplitSave and Price Prediction. Second, we solve for our customers' travel needs, including the launch of our new rail disruption features.
I'll give you more details on this shortly. Third, we build trust and loyalty, scaling products like digital Railcards. Fourth, we increase customer engagement, for example, expanding ancillary services we offer our customers. Let's discuss some of these in more detail. Starting with solving customer needs, where we are rolling out a set of new rail disruption features. The features will support customers when navigating disruption on the rail network, leveraging the power of our AI and data tools. It will include Travel Forecast, which provides personalized notifications to customers in advance if their train is likely to be delayed or canceled. Customers will be able to see the location of their train in real time, with a map interface powered by our Signalbox technology. And our forecasting capabilities will continuously improve, leveraging real-world data sources, including our base of 18 million customers transversing the rail network.
Delay Repay notifications, which alert customers when they are entitled to compensation. These will be an interim solution until the industry allows third-party retailers to offer fully automated Delay Repay. Our notifications will provide estimates of what each customer is owed, plus a punch-out to the relevant TOC website to complete their claim. Our beta tests over the summer enabled the processing of around GBP 1 million in compensation claims. Finally, our AI Traveler System, which offers customers a live native chat experience with real-time travel information personalized to their specific journey. Since launch, we've been selective in deploying the AI system within the app, yet it's already garnered strong levels of customer engagement. So far, it's had over a million conversations with customers, almost a third of our repeat users. It's answering most queries, with less than 10% handed off to customer service representatives.
We'll soon deploy the assistant more widely across the app, increasing the opportunity for customers to engage with it, while also expanding its breadth of real-time knowledge capabilities. Let's watch a short video that brings to life our suite of rail disruption features.
Every day, thousands of travelers face delays they can't control. That's why we've built a powerful set of new features to help customers when they travel: Travel Forecast, Delay Repay, and AI Assistance. It starts with Travel Forecast, helping you plan ahead before problems strike. If travelers are delayed, we'll notify them and do all we can to help them get their compensation. When things change, our AI assistant guides you to smarter alternatives instantly, because journeys don't stop, and neither do we. Trainline, your smart travel companion.
When it comes to building trust and loyalty of our customers, Trainline has cultivated strong brand affinity over many years. In fact, we are the most trusted brand in U.K. rail, and our brand consideration is at record levels, significantly outperforming all other retailers. This has supported Trainline's continued growth in the U.K., particularly when faced with notable competition, and it will become increasingly important in an AI-driven search world. One example of how we're building customer loyalty is through digital railcards. We've enhanced upselling within the booking flow, highlighting to customers how much they could save by buying a Railcard alongside their ticket. This has scaled our user base 12% to 2.5 million in H1. By doing so, we are enhancing retention of highly engaged customers who transact 4x more often than non-Railcard holders. We're gaining particularly good traction with younger cohorts.
Our share of the 16 to 30-year-old Railcard segment has now increased to 44%. We're increasing the opportunity for customers to engage with Trainline, broadening our range of ancillary products and services, and growing additional revenue streams. These include hotel bookings and insurance sales, which grew strongly in the first half. At the same time, we are optimizing how we monetize our existing products and services. For example, this year, we are focusing on enhancing advertising revenue and, in the first half, improving the positioning of ad placements within the app. Likewise, we are currently running tests for a SplitSave fee. This could present a long-term option to supersede the booking fee where SplitSave applies. Now, turning to our international business. We continue to position ourselves as the aggregator of choice as European markets liberalize. Over the summer, carrier competition expanded on the EUR 1 billion high-speed corridor in Southeast France.
In June, Trenitalia launched five return services a day between Paris and Marseille. This is already having a notable impact, with average fares down 27% on the route. In addition, Trenitalia almost doubled their operations between Paris and Lyon to nine services a day, and that's due to increase again to 14 services a day from December. We are positioning ourselves as the aggregator of choice on the French Southeast high-speed corridor. We are leveraging our highly rated mobile app to showcase all the fares from all the high-speed carriers, with features that help unlock value for customers, like TopCombo, which allows customers to stitch together different carriers for return and multi-leg journeys. At the same time, we're positioning ourselves as the partner of choice for carriers, driving customer demand and, in turn, supporting their growth. As you know, we paused brand marketing in France a couple of years ago.
With Trenitalia's recent expansion, we resumed our efforts to grow our awareness in the Southeast. We are sponsoring Lyon-based football team Olympique Lyonnais, and we are running large campaigns in online video and out of home at stations and transport hubs around Paris, Lyon, and Marseille. We already have good levels of brand awareness across France at around 28%. Our focus on Southeast France has significantly increased brand awareness in the region. Across Paris, Lyon, and Marseille, our blended awareness score was up 12 points this year to 48%. That's supporting strong net ticket sales growth on the Southeast network, including sales between Paris, Lyon, and Paris, Marseille up 34% in the second quarter. France is a huge rail market, worth about EUR 11 billion today. It is expected to grow to around EUR 14 billion or EUR 15 billion by 2030, of which around EUR 7 billion will come from aggregated high-speed routes.
We see the Southeast network as a gateway for growth elsewhere in France as carrier competition expands over the coming years. This includes Proxima, who will operate under the Velvet brand. They will run trains between Paris and cities in Western France. This will include Paris to Bordeaux, France's second busiest rail route. In addition, Le Tram, ILiSTO, and Renfe are all due to launch domestic services in France too, while cross-border carrier competition is set to arrive in 2030, with Virgin Trains set to launch. In Spain, we're evolving the balance between growth and profitability. Spain has been an ideal market to hone our aggregation playbook, where carrier competition expanded across its EUR 1.5 billion high-speed rail market. We invested behind our user experience and our brand awareness.
In turn, we have scaled our net ticket sales, giving us a considerable lead versus other market aggregators, and we continue to see runway for further growth. At the same time, we are increasing our focus on driving profitability in Spain. We are normalizing marketing spend while placing more emphasis upon customer engagement and growth of transaction frequency. Likewise, we are finding new ways to help carriers to grow. We recently launched Sponsored Journeys, a paid service that allows carriers to increase their prominence within our search function. We launched our first paid campaign with Iryo, which has been a real success, notably increasing customer demand. Across our international consumer business, increasing ancillary revenue remains a growth opportunity. Having made good progress in hotels last year, in H1, we bolstered our insurance offering with the launch of our new trip insurance product.
Offered alongside our existing Cancel for Any Reason product, this drove a material increase in insurance revenues. Moving on to Trainline Solutions, our fastest-growing business unit, and now generating over GBP 1 billion in net ticket sales. Business travel is our main growth opportunity here and represents around 50% of Trainline Solutions' sales. This is generated through our own branded channels as well as through our B2B distribution business. B2B distribution allows travel management companies and other business travel platforms to offer rail tickets to their respective customers. Primarily a U.K. business, we increasingly support our partners to sell tickets from multiple European carriers as well. They can do that all through one simple, seamless connection on global API, rather than tackle the complexity of connecting to multiple different carriers. As a result, international B2B distribution grew 55% in the first half. This business has good momentum.
Many of the world's largest TMCs and travel platforms are now connected to our global API and primed to grow ticket sales. In September, we expanded our partnership with the world's largest travel management company, Amex GBT, giving us confidence in our future growth. Finally, let's discuss our Digital Pay As You Go solution and its new trial in the U.K. As a recap, Pay As You Go travel provides a convenient option for short-distance journeys, also known as contactless or tap-in, tap-out. It's well established in cities like London, where it is frequently used by commuters. However, prepaid ticketing generates most of the passenger revenue for the U.K. rail industry, supported by price discrimination. It's far more suitable for long-distance journeys, where passengers tend to book in advance to get cheaper fares and reserve their seat, and we don't expect that to change. As you know, we've developed an in-app pay-as-you-go solution called Digital PayAsYouGo.
It leverages our geolocation technology from Signalbox and offers capabilities way beyond traditional tap-in, tap-out systems. For customers, this includes real-time pricing, integrated railcard discounts, and the ability to buy group and family tickets. For the industry, it requires no gating infrastructure, reducing the CapEx outlay and the time needed to deploy pay-as-you-go networks. We therefore see Digital PayAsYouGo as a better solution, which, when rolled out, can increase our scope to serve commuters and travelers booking on the day. In September, Digital PayAsYouGo trial went live on the East Midlands rail network. This represents a strategic opportunity to test our solution and demonstrate the benefits of Digital PayAsYouGo in a live environment. Of the four trials awarded by RDG, the East Midlands trial is the most complex, given it encompasses three different cities, Derby, Nottingham, and Leicester.
While it's still early days, we are learning fast, and feedback that it worked flawlessly is highly encouraging. Before we wrap up, let me play a video showing our Digital PayAsYouGo trial in action. Before I hand over to the operator for questions, let me summarize the key takeaways from the half. We've delivered a robust operating performance, improved profitability, and strong cash flow, underpinning our enhanced share buyback announcement. Today, we've increased profit guidance for the year, reflecting our disciplined approach to cost management. Looking ahead, I see sizable growth opportunities for our three business units, all of which are leaders in their respective markets. In U.K. consumer, we are deepening our competitive moat, launching new rail disruption features and scaling Digital Railcards. In international consumer, we are positioning ourselves as the aggregator of choice in Spain and Southeast France as carrier competition increases.
While in Trainline Solutions, we are supporting B2B travel partners as they expand their rail travel sales across Europe. Finally, as you've just seen, we're off to a great start with our Digital PayAsYouGo trial in the U.K., increasing our scope to grow sales of commute and short-distance travel. Thank you very much for listening. I'll now hand over to the operator for questions. When asking, please state your name and your organization.
Thank you. We will now begin the question and answer session. As a reminder, if you would like to ask a question today, please do so now by pressing star, followed by the number one on your telephone keypad. Our first question today comes from Gareth Davies with Deutsche Numis. Gareth, please go ahead.
Good morning. Thank you for the presentation. Just. The first one for me is probably the obvious in terms of GBR consultation document this morning. At face value, it appears to tick a lot of boxes in terms of sort of level playing field and operating in a fair and transparent way on an ongoing basis. We'd just be really interested to get your sort of headline thoughts on the key points that are in there and possibly anything you felt may should have been covered and isn't.
The second one is on the Digital PayAsYouGo trial. It sounds like that's going well. Just wondering, is there any scope for that trial to be expanded and you to be allowed to use your own app. And then go beyond the sort of 1,000 customers that are currently doing it with East Midlands? And then final one, just on the Google, was obviously a headwind, particularly salient in the international business. Just wondered, now that we're lapping comps, presumably that is becoming less of an explicit headwind. Can you just give us a little update around how that's evolved? I mean, taking on board that GEO is going very well, but clearly Google is still pretty important for international inbound. Thank you.
Thanks very much, Gareth. Yeah, let me start at the top there with the GBR question. Look, there's been lots of news flow over the last sort of few months and an important day today. Let me go kind of right back up to the top and just sort of state that, look, I do think we're actually many years away from a GBR app launching and the 14 different sort of top existing apps ultimately closing down.
Those customers having to make a choice potentially to go to the GBR app. I think what's driving this is clearly the government's desire to consolidate from a customer point of view to take 14 apps into one and also to improve from a kind of cost-based point of view. Just to say right up front for everyone, we look forward to that moment when it happens. We back ourselves to compete with whatever comes. I think, if you remember where we were three years ago with Uber launching, there were concerns and questions, a company that had huge backing that threw a lot of money at the sort of discounting tickets and massive kind of marketing campaigns. I think three years later, you can see kind of how customers have reacted in terms of the strength of the Trainline offering.
To set that out right at the front. Look, as you say, what we're seeing, and as I mentioned earlier, principally, today's about the launch of GBR primary legislation, which is about the creation of GBR. This has to happen for the government to stand up GBR because of the legal, financial, operational, health and safety, and indeed organizational questions that need to be answering. That is a lot of the priority for the government and GBR. As you also referenced, the consultation document that was published this morning. Look, let me share a few thoughts on that. I think overall, we do actually see that, as you suggest, as a significant step forward, both in the context and direction of travel for the government.
Let me sort of pick a few parts out where it's, I think, elements that are encouraging. I just say up front, for those who have not had a chance to read that yet, it talks about the creation of a GBR retail unit within the overall GBR. I think that's important to kind of note. The elements within that that I think are helpful. First is it talks about, within that GBR retail unit, a separation of the industry management functions from the sort of commercial functions, where the former would be a lot of the RDG existing kind of organization and managing the relationship with, amongst others, third-party retailers like Trainline, whereas the latter, the commercial arm, would clearly be the arm responsible for sort of standing up the app and the website. Ultimately, we would compete with them.
I think that separation is really important and very helpful. Secondly, it talks about the establishment of a code of practice of how GBR interacts with all participants, obviously third-party retailers being really important there. That code of practice, and indeed, we've been pushing for a codification of how this would operate. I think that will be a really important step. The fact that that will exist is really important. Ultimately, enforceable under GBR's future license. Finally, what we get today is a sort of laying out that there's the ability for, if we felt that GBR were not living up to or living that code of practice, for us to challenge GBR, working through the ORR and really to effectively ensure GBR do indeed operate within an open and fair market. I think those elements are really helpful.
It is just worth noting, these set of, if you like, safeguards do not exist today. That feels like very helpful. I think in terms of what we still need to see, and it is somewhat self-evident, this code of practice is not yet defined. I would anticipate, but do not have a clear timeline, that work would begin on that in the early next year. Our understanding is it would be led by the ORR working with the DfT. I would anticipate that the CMA would have significant input into that process, along with, obviously, ourselves and other third-party retailers. So I think all of that together provides reassurance. I think in the way, the spirit you asked the question, I am encouraged by kind of all of the above with important questions still to answer.
What I would just note is this all relates to future design. As I've said, I think this is sort of multiple years from fruition. What's also important in the interim is, as I just laid out in the presentation, that we see kind of resolution on areas like Delay Repay in the interim so that Trainline customers have access to that kind of one-click service. And so that's gone pretty deep on the retail. I just spent a minute then talking about the reality of what that means for the app, which is where a lot of the questions end up going. Why do I think it's potentially multiple years away? I think the next phase, and I don't know the details of this, will be some form of exercise that most likely is a procurement exercise from.
The government around working out what they want to procure. There would then be that process complete, and there would then be a build phase to develop what would be a relatively important app that would have to sort of handle a lot of different scenarios. It can't just do what LNER does on long- distance or what Northern does on regional or some of the use cases of Southeastern coming in on commute. It has to serve all of those. What I anticipate, and I'm somewhat speculating here, is that there would be a period of dual running. Even once the app was created, there would be a GBR app, which would be an important political win. But ultimately, that would dual run alongside the existing top apps.
At various points, they would begin to fold those top apps into GBR, and the customers would then have to kind of restart on the new app. To my very first point, we're excited for that opportunity. If you kind of think forward a few years and sort of the innovation we're laying out today and think forward another three years, I get pretty excited about where we'll be. GBR will kind of be just starting, if you like, there. That's a bit of a state of the nation, right, in how we see all of the GBR points there. To come to your second question. Digital PayAsYouGo trial. Look, of course, I hope there is scope to expand and to ultimately be able to put it into our B2C app. I think at this phase, we don't have visibility of that.
This trial runs through to the summer. I think it's kind of early days. The government is learning. I am really encouraged just to underline how well we have brought this to life. I think there have been apps that have kind of done the fare capping and the tap in and out. No one's really brought this all together with the route planning and the pricing and the kind of UX. And so ultimately, if the government kind of wants to, or any government across Europe wants to sort of bring this to life at scale, having it in the U.K., 18 million customers to help bring this to life, I do see that ultimately could be in the government's interest to see these third-party retailers offering this type of functionality. That is obviously the aspiration.
It's probably too early to speak with any precision on that yet. And then look on kind of Google's headwinds and where we're at there. I think if we sort of stand back on the international results, it's a portfolio of businesses. As we reported over the last couple of years, we saw initially kind of Google Trends, and then we saw this expansion of the SERP page as kind of the AI features coming in and so forth. We did talk about that as a headwind. I think the way to think about that is that actually impacted different GEOs at different points. I would say in the very earliest GEOs, we are, as you suggest, beginning to see that headwind turn into a tailwind.
I begin to see some green shoots in some of our GEOs coming through as we've kind of entered the half, which is helpful. We don't actually begin to fully lap all of those GEOs and some of the domestic European GEOs until later in this half. We don't yet see that tailwind. I am encouraged in the spirit of the question that I do think that headwind turns into a little bit of a tailwind. We'll keep close to that. Thanks for the questions. Operator, do we have another question?
Thank you. Our next question comes from Alastair Reid with Investec. Alastair, please go ahead.
Good morning, guys. Yeah, a couple from me as well sort of following up on some of those things. I guess, firstly, with the consultation today. With sort of GBR ticketing not being sort of structurally and commercially separated out, do you have any sort of concerns it could lead to potentially commission rates being reduced or the ticketing having sort of less need to cover its own costs, not being sort of overtly self-funding? Secondly, I think you've touched on this already, but can you give us any examples you've seen in the past where talks have changed from one provider to another and kind of what share gains you have made in those switching moments? And then lastly, just on the topic of your sort of moats that protect the business from theoretically being disintermediated by sort of agentic AI, can you talk about, is the underlying sort of market data in the industry sort of freely and easily available to all tech firms? I'll leave it there. Thank you.
Thanks very much. Let me pick up on the first one, and then you can perhaps take the second one and come back on the third. In terms of the broader consultation and how we see the structure of that, I think kind of the direction of the question is going to the very nature of the open and fair retail. Look, we absolutely anticipate that we will be retailing on a fair basis with GBR. I think where you're going really is the code of practice and how that is defined. We're going to have to engage and wait to see that to come through. I think what I take kind of comfort from is this statement of the value of third-party retailers and the value they've driven in terms of innovation and driving up standards for passengers.
So I think that's where we now need to see that kind of actually codified out. Look, the CMA is still involved, and they will be involved going forward, and they've really committed to a level playing field. Those are areas that we'll be engaged with over the coming months. Pete, do you want to pick up on the second question of the provider change?
Yes. We've had a couple of examples in the past. South West Trains becoming South Western Railway, Virgin becoming Avanti were both kind of moments where there were shifts. And as Jody outlined, there is this moment where customers have to make a choice. From a traffic perspective, which doesn't fully represent sales necessarily, there's been quite significant shifts in the initial period of time.
Look, maybe around 30-40% in the first six weeks or so is something that we have seen in the past. Yes, it really does represent an opportunity for us as these things unfold.
Great. To come back to the point on U.K. moats and I think the role of agentic AI, what I'll do is I'll go up a level. If I haven't got to answer the question, please come back because there's quite a lot potentially in there. We think about AI within the business in sort of three different areas. The first is how we productize AI into the Trainline app. That really speaks to things like the AI system and going forward. We talk about the ecosystem where it's a way to get traffic, kind of another surface. Then we talk about kind of productivity.
I think you're really picking up on that middle one around the broader ecosystem. Just to say, we're excited for the first point on productizing. I really feel like that AI is giving us kind of ability to solve new customer problems as it relates to disruption, not just around an AI system, but the datasets and predicting travel patterns and potential delays. That's kind of cool stuff. I think speaking to the kind of moat point here, I just don't believe that there's many other players who can have the data and the datasets, the kind of data smarts and AI capabilities in their organization, and then the ability to invest behind it. We're really pretty excited for where that's going. Then to the ecosystem point.
Look, I'll give you my perspective on this, which is I've been doing this for well over a couple of decades now and watched various players enter the market. Over that period of time and worked with them to develop as they develop sort of as traffic sources. I think we're at that phase. I think the announcements we have seen from ChatGPT, for example, increasingly give me that sense that they really want to ultimately send traffic to us and find a way over time to monetize it, which we think is a good thing, right? Because it's much better to have three players, kind of Google, Meta, and let's say ChatGPT, that we're effectively buying qualified, high-quality traffic from rather than having a single player or two players. I do think that's going to take some time.
These things, whilst we see and we spoke to the growth, it still represents less than 1% of our total traffic. I think we're talking years for these ad products to develop. I don't think this is happening in the next few months. That's kind of how we view it evolving. I think you specifically asked about moats. I kind of break our moats into two sets. We've got what I call our consumer moat and our platform moat. You were pushing on the platform moat part. Just as a sort of reminder, we have a very strong consumer moat. I think in many things that are transversing technology change, the consumer moat really stands and is hugely important. That's about the quality of our brand. It's about the sentiment that exists towards it.
It's about our really deep vertical UX. It's about the app install base. The fact we have 18 million users in the U.K. or 27 million across Europe means we have this really strong engaged base that will want to keep using us. I think when you extract out and say OpenAI would have to develop a great vertical UX, not just in trains, but in hotels, in planes, in cars, but also in black dresses and selling drones, it gets very complicated. They will definitely work with the great brands and the great UXs that exist out there. I think that's part of the moat. Finally, to finish, in terms of that platform moat, yes, we have a huge number of data feeds that go back to the industry all over Europe.
These are kind of complex and difficult to develop, and that's part of the moat. We have huge commercial agreements because, as you know, we're basically selling billions of GBP worth of tickets every year. That requires bonding and obviously lots of due diligence. You can't just initiate these things. It's not like we're selling a few thousand GBP worth. That is also a moat, which makes it incredibly difficult to just start a business. Even if you were to, to make it work at scale becomes even more challenging given the sort of size of the numbers we're talking about. Thank you for the question. I hope that picked it off. Should we go to the next question, Operator?
Thank you. Our next question comes from Ed Young with Morgan Stanley. Please go ahead.
Thank you. Good morning. I've got two on numbers, please, and then one on strategy. On the EBITDA guidance upgrade, you've laid out some of the drivers of the stronger EBITDA outlook. I wonder if you can help us understand what changed specifically since you gave the guidance in September. Is it fair to say that stronger top line has endured into H2, or is it other moving parts that led to your change in posture there? Second of all, free cash flow growth was suppressed by working capital movement a bit in H1. Should that reverse in H2? Should free cash flow growth mirror profit growth for the full year?
And then finally, on Spain, I know you touched on it in the presentation, but I'd love to hear a bit more color about why now is the right time to move to this more balanced posture for growth versus profitability in Spain. Thank you.
Great. I'll let Pete pick the first two up and come back around on Spain at the end.
Yeah, thanks, Ed. Yeah, let me start with the top line and then get to EBITDA. And if I think about the U.K. business, we've had a robust first half performance. There were a few benefits, as I've kind of highlighted, lapping strikes and finally getting those behind us kind of supported that figure somewhat. We do have the headwind of overall expanding as the year further unfolds. We've obviously had the first 47 stations, which dropped in February. There are another.
50 or so which are expected either in December or in January, and there are more later in the year. There is a kind of headwind that is building. I think net of that, I would expect H2 to be growing at a slower rate overall versus H1. Nonetheless, EBITDA has performed well in H1, and that confidence we are kind of taking forward into H2. Of course, there will be drop-through from that performance at the top line. I think the other thing of note really is how we are thinking about marketing spend in international. In H1, as Jody outlined, we have put more into supporting the expansion of the France and the Southeast corridors there. That will somewhat persist into H2, where there are more services being run by Trenitalia. We will keep pushing there. In Spain, we are kind of balancing growth and profitability.
If I take a real step back here, we started, what, three, four years ago when aggregation in Spain was clearly going to be this kind of big all-in-one-go moment. We had a very small footprint there. So we really pushed hard to build brand awareness, to make the most of this kind of very dynamic moment for the rail- traveling public in Spain. And you know we pushed hard on the marketing spend, and we've been really pleased with the strong growth that we've seen there. What we've now reached, though, is a point of kind of evolution where we're developing the next bit of the playbook. We're thinking about how we balance the profitability and the growth. And so you should expect marketing in international as a result of that to step down a bit in H2.
Net-net will be also additive to delivering on the EBITDA guidance that we have shared today. And then your question on free cash flow growth and the working capital movement, yeah, this is always a slightly tricky one because the day of the week and the slightly odd 13 periods of four weeks that the settlement process in the U.K. rail industry means that we get some oddities on the two points in the year that you see it. Fundamentally, there is a good guy here for cash from a working capital perspective. It is typically a bit better at August than it is in February just because of the cycle of these things. When it is a Monday or a Tuesday, that can impact it all as well. So, Net-net, I do not think you should pay too much attention to.
The kind of puts and takes that we see at different points in time. I just believe that there is a goodness that lies underneath. It is normally around the kind of 15-20 million across a year that we would see on average, even if that is not at that particular point in the year.
Thanks, Pete. I think Pete's mostly answered the third question around Spain, but let me just give you a little bit of a strategic overlay there. As Pete said, this was a moment in time when we saw what was happening in Spain. We were almost at 0% market share. We were in a position where all the lines and all the competitors were launching in a kind of 12-month window. We knew we had to show up there with Trainline not really having a footprint to date.
Through those three years, we have got to having double-digit market share. And I think mportantly, we are now the number one domestic operator by a distance. We are significantly ahead of Trenes, of Omio, of Uber, and any of the other players. The kind of go-to place for aggregation. And so really, we are moving to what I would call balanced growth. We are still leaning forward into Spain and investing there, but we just do not need this kind of launch-level marketing spend. As you have seen us over the last few years, we sort of play the overall portfolio. We are now moving, if you like, that firepower into France as that launches to ensure that we become and remain, frankly, in France, the number one domestic operator. It is just a moment in time. Look, you will have heard.
As it relates to Italy, we see a huge opportunity there in 2027, and we will think about marketing there again. We have to kind of balance the overall portfolio and have these kind of launch phases and then what I call more balanced growth phases. Thanks for the questions. Ed, should we take another operator?
Thank you. Our next question comes from Andrew Ross with Barclays. Please go ahead.
Great morning, guys. I've got three, if that's okay. The first one is to follow up on Alastair's question on commission rates. Is it completely ruled out that there will be no review of commission rates for the industry as part of the detail of how GBR kind of comes together for next year? Or kind of more broadly, where are we at on that as we kind of transition from RDG into GBR around the industry commission rate?
That's the first question. The second one is on PayAsYouGo and I guess scenarios as to when a kind of nationwide contract may be awarded or how this might look and kind of what it might mean for you if you were or were not to get it. I guess I'm thinking about kind of the incremental take rate you might achieve if you were to kind of get a broader contract and any kind of cannibalization risk to your core business that you think may or may not happen. I think maybe not, given the solution would be portable into your app, as I understand it. Third question is on agentic AI. Kind of taking your view, Jody, but is it more likely for OpenAI to partner with kind of a vertical specialist like yourself, which I agree with.
Do you worry about any risk to kind of time-spent-based monetization, things like cross-selling hotels, ads, that type of thing if we move into a world of kind of agentic transactions and booking trains?
Thanks, Andrew. Look. On that first point, this is what we've discussed kind of many times, I guess. The idea of commission rates. Look, I'll give you the overall kind of answer here. They have been reviewed, I think, 3x in the last 20 years. Every time they've essentially come back and concluded we were at the right commission rate, apart from, as you know, the last time where it was a net 25 basis points impact. We feel that the commission rate is absolutely at the right level. And multiple independent players, every time this has been looked at, everyone has concluded we're at the right level.
And so I think that's what gives us confidence. The way we operate, we have a number of years, kind of 3+ years of confidence in our commission rate. It's never off the table forever, right? There is no discussion of that. I look forward, as I've always done, believing we're set at the right commission rate. Pete, do you want to pick up on the second one? Just by your go.
I think just as a reminder, the trials that we have are kind of the first step forward. And quite deliberately, the government is testing different technologies and really seeking to learn what those technologies offer, how customers engage and relate with the different aspects of what's being tested. So we're kind of really quite early on.
Whilst we see some real political support for finding a solution for this, and there's mention of this in the conduct, so that kind of gives you an indication of the direction of travel, there's a lot of detail that still needs to be worked out in terms of how this will be implemented. As we said in the presentation, I think whilst one route might be a kind of nationwide contract, we don't think this technology is particularly helpful or is likely to be implemented for long-distance travel, right? Perhaps a nationwide contract might enable a series of different cities to pick it up, but even that's not a given at the moment as to the way forward. It could be rolled out more regionally. So there's still a lot to see there. Likewise, the commercials that would sit alongside this would need work.
I think it is fair to say that on the basis of kind of fair and open, we would expect, if there is some sort of national implementation, that even if our technology was not chosen, there would be some way for us to partake and offer that to customers. So anyway, a lot of detail still to be worked out. We are really pleased about the progress we are making and the political support for continuing down this path. We will have to see how it goes.
Let me pick up on this kind of agentic AI and the sort of broader transactions. I understood the spirit of the question was, would we lose other services if it was kind of the transaction occurred within, let us say, a text interface, in a ChatGPT app or something?
Whilst I do see that as a use case, and one could imagine certain trains being booked that way, I think we have to assume the core use case is kind of what happens on Google and on Meta. If they fully want to realize the value of that customer, then ultimately, there needs to be a high-quality conversion rate. And the best way to do that is for the customer to go into, in our case, a Trainline user experience, or it could be any other shopping or commerce user experience where if you think about the pages we show and the clarification of which fare you want and the flexibility of it, and frankly, what time train you want, and the ability then to get customer support and to get your ticket and your barcode, that's clearly better done within our app.
Or some experience that is effectively our app. I still envisage a world where the full experience arrives, and they have given us very qualified traffic. Worth knowing that we are absolutely developing and have developed the kind of frameworks and the MCP piece to allow the AI system to engage. You could imagine the early stuff happening, if you like, within ChatGPT, where we work with ChatGPT, if you like, with that, and then popping into our experience to actually make the transaction come to life. We are also not hypothesizing of how this will look. Just to say again, it's less than 1% of traffic right now in our most sort of forward GEOs. I kind of feel pretty good. We've got more time, I think, for just one more question, and then we'll bring it to a close.
Thank you. Our final question comes from James Lockyer with Peel Hunt. Please go ahead.
Yes, good morning, guys. Thank you for taking my questions. I'll just ask to this stage, so based on AI. It's good that you've spoken about one million conversations or over one million handling the customer queries, but less than 10% being handed over. Based on the types of questions that are being asked versus the typical questions that your human agents are getting, once sure, where might that 10% land? What level of cost savings might you be able to realize? Secondly, can you talk about how you're using AI internally? You mentioned that you are, but it'd be good to hear about what you're doing there. Lots of companies thinking they'd be able to grow the top line without growing headcount as fast as they might have done without AI.
Is that something you're finding too? Any ROI, productivity, or hiring stats you can provide there would be useful. Thank you.
Sure. To give them kind of at the top of the hour sort of brief answers, I guess. I'd say overall, the types of questions, what it's allowing is customers to get reassurance. Things that they kind of were like, just didn't know how to. Is this ticket valid on this train? They would then go and try and find station staff or ask a friend or start reading very detailed conditions. AI is doing a really great job of that. It wasn't that they were necessarily going to customer service. Where you're going is right. I do think it allows efficiency.
It allows our customer service teams to work on the higher- value questions, if you like, because it filters out a lot of the ones that can easily be handled by AI. I think, yes, where you're pushing, right, it does lead to kind of greater productivity. As it relates to what we're doing internally, lots of good stuff. In terms of using the kind of Copilot-style tools, in terms of helping engineers code, that is definitely seeing kind of productivity improvements. I think the spirit is to be able to do more with our existing employee base is absolutely how we're looking at it. Then just allowing other things.
We have an experimentation GPT, right, which has got now all of the experiments that we've kind of almost ever done over the last few years in one place, which can be accessed by a UX or a product or an engineering or commercial lead really quickly versus having to, frankly, call nine different people and try and find out which is what happens in most companies up to this point. We are increasingly finding those use cases which are driving a more cohesive workforce and allowing us to kind of better pull that knowledge and be quicker in the development. I think, again, where you're going, delivering more with our existing base is very much how we're thinking about it. Thank you very much for the questions, James. Thank you all for listening today. That's all we've got time for.
To recap, we've had a strong first half, delivering a robust operating performance and improving profitability. In terms, we have today improved our guidance for the full year and see sizable growth opportunities across all three of our business units. I look forward to speaking to you all again soon. Thank you.