Trustpilot Group plc ($TRST)

Good morning, everyone. Thank you for joining us today for Trustpilot's Full Year 2025 results, and a warm welcome to everyone joining on the webcast. Hanno and I look forward to walking you through what was an excellent year. Our performance was driven in no small part by the rapid evolution of AI and the increasing value of the human trust signals that Trustpilot collects. I'm looking forward to taking you through some operational highlights before handing over to Hanno for the financials. I'll then come back with a deeper look at strategic progress. Three key takeaways from today's results. First, great execution. We grew bookings 18% in constant currency and scaled highly profitably, driving a 69% increase in adjusted EBITDA and returning $72 million to shareholders through buybacks. Second, the expanding strategic advantage. The growth flywheel is thriving, fueling a 20% increase in active reviews to 361 million. By opening our proprietary data set to large language models, Trustpilot is now an essential part of the rapidly evolving area of answer engine optimization. Click-throughs from AI search grew nearly 15-fold year-on-year. Large businesses care deeply about their exposure in AI search. So this was one of the key drivers of success in the enterprise segment. Finally, upgraded guidance. 2025 has given concrete evidence that AI is increasing the value of Trustpilot feedback and therefore, widening our competitive moat. This, coupled with the operational efficiencies we can already see AI delivering gives us immense confidence in the future. So today, we're upgrading our medium-term outlook to 25% adjusted EBITDA margin in 2028 and 30% in 2030. Let's take a look now at operational performance. In essence, Trustpilot products collect customer feedback and turn it into influence over the actions of people and businesses. Because the feedback has influence, people are motivated to write more of it and businesses want to engage with it. In 2025, our users submitted more new reviews to Trustpilot than in the first 12 years of the company combined. Our unique data set expanded to 361 million active reviews, representing a 20% year-on-year increase. The scale of this unreplicable data set and the ability of AI to drive more value from it is a key defensive moat for the business. The healthy growth flywheel drove excellent financial and operational results. Bookings grew 18% on a constant currency basis with momentum accelerating in the second half and businesses paying us over $20,000 per year, growing 35%. Growth was particularly strong in North America at 21%. Because of the inherent operating leverage in our model, 20% constant currency top line growth resulted in a 69% increase in adjusted EBITDA to $40.7 million, delivering a margin of 15.6%, comfortably ahead already upgraded expectations. Finally, on the AI front, authority continues to compound. In January this year, PromptWatch ranked Trustpilot as the fifth most cited domain globally on ChatGPT. This means ChatGPT is directly referenced or linked to Trustpilot as a source of information when answering a user's prompt. This is driving heightened interest from enterprise clients. Increasingly, our sales teams are leading with AEO as businesses look to secure their reputation in AI models. Let me touch now on the regions individually. In the U.K., our largest market, we delivered 16% constant currency bookings growth to $116 million. The network effect here is particularly well established, resulting in the highest regional contribution margin at 65%. This shows what the wider group can look like at scale. Because the brand awareness and network effects are so deeply embedded, we were able to drive excellent profitability. As other regions continue to mature, the U.K. proves the long-term margin potential inherent in the model. The strategy to focus on larger enterprise accounts continues to deliver. We welcomed a wide variety of flagship brands in the U.K. in the second half of the year, including Sky, Samsung and United Utilities. The share of ARR from businesses paying over $20,000 rose to 47% by the end of the year. With market penetration of just 5%, the runway for growth remains significant. Turning next to Europe and the Rest of the World. We delivered strong bookings growth of 20% at constant currency to $113 million. A few notable wins in H2 included Enco, TotalEnergies, Costa Cruises and Canada. The region demonstrates what happens when we focus go-to-market efforts into specific geographies. Growth in both Germany and Italy, our focus markets was well ahead of the regional average. The enterprise segment is particularly strong in Germany, where by the end of the year, businesses paying more than $20,000 accounted for 58% of ARR. This is supporting a robust regional contribution margin of 55%. Moving to North America, which comprised of the U.S. and Canada, it was the fastest-growing region, delivering 21% constant currency bookings growth to $62 million, building on an already strong comparator last year. We're seeing accelerated enterprise velocity here, helping to drive a 4 percentage point improvement in contribution margin to 38%. Both sides of the flywheel are growing well with reviews submitted in the year rising 27% to 13 million. Brand awareness is also growing rapidly, aided by initiatives like the Inaugural Writer Review Week in October, which delivered our highest ever U.S. traffic week. On the business side, TrustBox impressions rose to 32 billion, up 38% on the year, naturally embedding the brand across the North American digital landscape. We welcomed notable enterprise clients in the second half, including Liberty Mutual and Squarespace. The U.S. remains a substantial opportunity for Trustpilot, and we'll touch on this again later. I'll now hand over to Hanno to take you through the financials in detail.

Thank you, Adrian, and good morning, everyone, and happy St. Patrick's Day. 2025 was another record year of bookings, margin expansion and exceptional cash conversion. Let's dive straight into the numbers. I'm happy to report that FY '25 bookings grew 18% in constant currency, reaching $291 million. As you can see on the top chart, we've established a clear consistent track record of top line growth and have by now surpassed $300 million in ARR. As shown on the bottom chart, we're scaling profitably, and we have transformed the margin profile of the business. In 2025, we delivered a 4.2 percentage point improvement in adjusted EBITDA margin to a record 15.6%, generating $40.7 million in adjusted EBITDA and $46.6 million in adjusted free cash flow. Looking at the retention rate in more detail, net dollar retention was 102%. The slight moderation from 103% last year was entirely expected as we annualize the one-off package migration benefit from 2024. We landed major new product features in the year like visitor insights and review follow-ups, while simultaneously driving an improvement in the gross dollar retention rate to a record 87%, up from 85% in the prior year. There are a number of factors driving this improvement. Starting at the top with the growth flywheel and product innovation. As awareness of the Trustpilot brand grows, the value of the businesses being active on the platform also increases. You'll hear this firsthand from a customer later. Net expansion remains strong at 15% and product innovation fuels this. We're increasingly weaving AI and LLM capabilities directly into our tools. Below that, you see how we operationalize those products through our go-to-market execution and enterprise strategy. We have embedded a globally consistent sales methodology and shifted to a proactive data-driven customer success model, which has reduced churn with early warnings for those customers that are at risk. Moving on. The chart on the left perfectly visualizes the compounding power of our subscription model. The revenue base is incredibly stable with each new cohort stacking on top of the last. Our big growth drivers are customers paying more than $20,000 per year. As we push further into this higher-value segment, we're bringing on larger customers. These customers stay longer and provide resilient compounding value. As you can see on the pie chart, the share of total bookings coming from this segment has grown from 25% of bookings in 2022 to 43% in 2025. The underlying unit economics of these accounts are excellent. Gross dollar retention for the segment is 93% and net dollar retention is 111% in 2025. Let's take a look at the income statement from a management view down to adjusted EBITDA, excluding stock-based comp and D&A. As always, an IFRS statement and reconciliations can be found in the appendix of this presentation. We delivered revenue of $261.1 million and a gross margin of 82.7%. The improvement in gross margin reflects the normalized sales commission we discussed in the first half as well as operating leverage in our cost structure as the business scales. Network and infrastructure costs grew at a fraction of revenue growth in 2025. This is also the result of active optimization work carried out through 2024 that we fully realized in 2025. Support costs also fell year-on-year despite higher volume. Sales and marketing remained relatively flat as a percentage of revenue at 27.4%. While we faced an accounting headwind here due to the amortization of capitalized sales commissions, underlying cash efficiency is actually improving. We continue to make conscious deliberate reinvestments into new customer acquisition and improving efficiency every year. Our main focus is funding the enterprise sales motion where the underlying economics are fantastic with customers exhibiting higher retention rates. The group LTV to CAC ratio expanded slightly to 3.6x, up from 3.4x last year. Where we saw the most significant leverage was in technology and content, which dropped from 25.3% to 23.6% of revenue. This was driven by AI-enabled efficiencies across technology, where the teams are using Copilot and now also cloud code in our content integrity teams. G&A also dropped to 15% of revenue as we maintained disciplined cost control. Finally, I want to highlight the improvement in impairment losses. As a proportion of revenue, they accounted for 0.6%, down from 1.2% in the same period last year. As remarked last year, 2024 was unusually high as we wrote off aged receivables from the COVID period, which had previously been considered recoverable. Overall, we demonstrated strong margin expansion across the entire P&L. This bridge clearly illustrates the mechanics behind the 4.2 percentage point margin expansion. Thanks to our subscription business model, the long-term margin potential of the business is significant, and we delivered leverage across the P&L. This includes CAC, if you normalize for the impact of capitalized sales commissions, which were EUR 2 million in 2025, down from 3 million in 2024. Product innovation remains a crucial part of the business because it directly drives retention and expansion. As we develop these features for a rapidly growing customer base, it naturally generates structural leverage over time. Trust spend combines elements of all these buckets, except CAC and grew in 2025 as a proportion of OpEx as we double down on our commitment to trust. Specifically, total spend on trust increased substantially as a share of OpEx. For this year, we expect further operating leverage to flow primarily through G&A with some leverage in sales and marketing as these newly acquired enterprise cohorts mature and renew. Moving down the P&L, let me touch briefly on the noncash IFRS stock-based compensation charge of $12.5 million, up from $9.5 million last year. This increase reflects new necessary share awards to attract and retain executive talent and high performers. However, the P&L charge does not tell the full story of shareholder value. I want to draw your attention to the actual impact on dilution. Grants will be on average 1% dilutive per annum in each 10-year period, as you would expect. Our active share buyback program more than offset shares granted and total diluted share count fell 4% from 450 million to 431 million by year-end. The clearest measure of underlying performance is cash generation. We generated $59.2 million in adjusted operating cash flow, driven by business growth, improved profitability and a benefit from more customers signing up to annual upfront payment. In particular, as we're signing up more enterprise customers, annual prepayments are a standard term, and we estimate the improvement in average prepayment across the portfolio pulled forward almost $10 million in cash flow. This is a structural recurring improvement in the working capital dynamics of the business. After capitalized development costs and leases, we delivered adjusted free cash flow of $46.6 million or an adjusted free cash flow margin of 17.8%. The increase in adjusted free cash flow reflects a 173% increase year-on-year. On a per share basis, adjusted free cash flow was up 174% to $0.107. This proves the exceptional cash-generative nature of our model. We closed the year with $47.6 million in cash even after returning $71.6 million to shareholders in the period. Since we started the buyback program 2 years ago, we have returned $115 million and 9% of the diluted shares outstanding at the beginning of the share program. This brings me to the capital allocation framework, which remains unchanged. First, we invest in organic growth, deploying an incremental $24 million last year whilst keeping product investment at a steady 3% of revenue. Second, we retain flexibility for strategic M&A that can accelerate our road map. Finally, we return excess capital to shareholders. Therefore, given our strong cash position and confidence in future cash generation, we intend to purchase a further GBP 30 million of shares, of which GBP 7.5 million will be via our employee benefit trust to satisfy future share awards. Turning to the outlook. Based on the strong bookings momentum in 2025, which is a leading indicator of revenue, we expect to deliver high teens constant currency revenue growth for 2026. Alongside this top line growth, we expect operating leverage to continue flowing through the business, delivering a further 2 to 3 percentage point increase in adjusted EBITDA margin. The ongoing share buyback then compounds this value and drives further improvement in the free cash flow per share. And this brings me to our upgraded outer year margin guidance. We have always talked about 30% margins being achievable. Today, we're putting a clear time frame on this. Over the past few years, we have transformed the margin profile of the business. We have strong top line growth and structural operating efficiencies, which combined with the benefits that we're beginning to see from AI, gives us confidence that we will reach a 25% adjusted EBITDA margin in 2028 and 30% in 2030. In addition to our strong track record of growth and margin improvement, AI gives us the confidence to put this time frame on the margin progression for 3 key reasons. First, AI is driving the value proposition within the enterprise segment where we see the highest margin. Second, it allow us to deliver higher-value products quicker. And third, it contributes directly to internal operating efficiency. With a high gross margin, we have a choice about where to deploy capital, and we have demonstrated over the past few years that we can deliver operating leverage across the P&L. And with that, I will hand it back to Adrian.

Thanks, Hanno. So in 2025, we successfully executed against 3 core priorities: trust in the age of AI, enterprise growth and product innovation. Let me walk you through exactly what we did in each of these areas. Let's start with trust. Trustpilot's relevance in the age of AI is becoming increasingly clear. Whatever capabilities AI developed in the future, there will still be people and there will still be businesses. People will always be interested in and want to share their experiences with those businesses. As you can see from the chart, the rise of AI has coincided with the surge in the number of reviews people are writing on Trustpilot and the number of businesses they cover. Since the launch of ChatGPT in late 2022, cumulative review volume has grown by 47% in just 3 years. The 62 million reviews written on Trustpilot in 2025 was more than in the first 12 years of the platform combined. A key reason for this is that the rise of AI has motivated more businesses to ask their customers for Trustpilot feedback because that feedback influences how they show up in AI search. The growth in reviews provides the verified human data set required to train safe AI and deliver the trusted citations that power modern answer engines. But volume without trust is meaningless. And in 2025, we removed 7.8 million fake reviews, an increase of 74% year-on-year. Let's walk through how we protect the platform using a connected system of technology, people and community, all underpinned by increasing regulation. When a review is submitted, the first line of defense is technology. Every single review undergoes automated multi-signal screening where AI models scan device, network and behavioral metadata such as location indicators and time stamp patterns. Given the volume of incoming data, we use a sophisticated range of techniques from AI, machine learning models and neural network analysis. Out of all the fake reviews we catch in 2025, 91% were removed by our automated detection efforts. The volume of historic reviews on Trustpilot with all the accompanying metadata is what enables our technology to be so effective. Technology is supported by people who conduct specialist assessments and perform testing to continuously improve detection. This is all reinforced by the Trustpilot community. Any consumer or business can flag a review they view as suspicious. These reports act as a vital feedback loop triggering further checks that initiate our standard automated and expert reviews. We never claim to be perfect and operating at our scale, it's inevitable that some misuse still occurs. But the evidence is that AI is helping us more than it's helping bad actors, meaning the platform is becoming safer and more useful than ever. Integrity also means applying the same moderation standards to all reviews regardless of whether the business is a paying customer. The chart on the page shows the percentage of reviews we remove across different star ratings. You'll notice this distribution is virtually identical whether a business is paying us or using free tools. The vast majority of fake reviews we remove, around 75% are actually 5-star reviews. The algorithms are blind to a company's commercial status. They only care about the authenticity of the data. Protecting platform integrity is work that never stops. As we improve our existing models and develop new ones, we first apply them to new reviews coming on to the platform. And once we're confident that they are performing as intended, we run them back over historical reviews. This is the absolute core of our business, and we have a lot more to share here. Therefore, I'm pleased to announce we'll host a dedicated trust-focused event on the 6th of May. Here, our Chief Trust Officer, Shazadi Stinton, who joined us recently and was formerly General Counsel at MoneyGroup and her team will take you through what we do in more detail. And I'm delighted that Shazadi could be with us in the room today. All right. Moving on to the second priority of enterprise growth. When we talk to business leaders, the conversation starts with a really simple reality. Private customer experience data that they collect like an internal NPS or CSAT score is incomplete. It's invisible to large language models, and it does nothing to help you win your next customer. To get a complete picture and to grow, businesses need to operationalize public feedback. We give them the tools to do exactly that. Let me set out here exactly how Trustpilot helps large businesses, those enterprise customers succeed. First, build trust. Potential customers, investors and employees can all read your Trustpilot feedback. Engaging with the platform builds trust with all of these groups. Numerous businesses include Trustpilot scores in their Board decks and bonus plans and increasingly, Trustpilot appears in annual reports, including those of many FTSE 100 constituents like Admiral, BT, Centrica and Lloyds Banking Group. Next, Trustpilot feedback directly helps businesses grow faster by boosting their presence in Google and on large language models and by improving conversion and engagement with marketing assets through social proof. A major U.S. cybersecurity business told me they gained market share by using Trustpilot reviews throughout their conversion funnel for new customers to accelerate sign-ups. Third, Trustpilot helps businesses improve operational efficiency. A large logistics business I met found that thanks to their huge volume of Trustpilot feedback, they can identify faulty physical lockers quicker than their internal maintenance team can. Finally, Trustpilot galvanizes entire organizations to care about customer service. And because the reviews are public, it enforces accountability. I gave you an example at the half year of HSBC. Another great example is the CEO of a major European car leasing business who told me he holds the managers of each of over 500 branches accountable for Trustpilot feedback at their location. And part of their job is responding personally to unhappy customers through our platform. The focus on larger customers is delivering tangible results. Because of the immense value they get from our products, we have more customers paying us large amounts of money and getting a fantastic return on their investment. If you look at the middle chart, the number of paying customers between $10,000 and $20,000 a year has grown at a 25% compound annual growth rate over the last 3 years. In fact, we've nearly doubled our customer base in that segment since 2022. The number paying us over $20,000 a year, which make up the largest share of total ARR has grown by around 36% a year over that same period. To put that in perspective, we've grown this critical segment by 150% in just 3 years, scaling to nearly 3,000 customers today. This strategic shift towards larger customers improves unit economics, enhances the quality of earnings and accelerates the growth flywheel faster because big businesses invite millions of people to write feedback and showcase the Trustpilot brand at great scale. One business you'll all know that's embracing every part of the Trustpilot value proposition is the leading U.K. retailer, AO. They are now very close to hitting 1 million Trustpilot reviews. For an online-only business selling high-value items like washing machines, trust and quality of service is their primary differentiator. AO uses Trustpilot as a key operational tool. They mine scores, reviews and trends to actively manage their customer service and inform their core business propositions. They've integrated Trustpilot into national TV campaigns and their core brand positioning. Who better to bring this to life than their Founder and CEO, John Roberts, in conversation here with our new Chief Trust Officer, Shazadi Stinton. [Presentation]

There we go. Feedback is a gift, folks. So a great example there of how Trustpilot is playing a central role in how a large enterprise business builds trust grows and improves as we're now doing for many thousands around the world. On this slide, you can see how focus on higher-value customers is transforming our footprint in the U.S. By selling premium tools and integrations, we're moving upmarket. U.S. customers paying over $20,000 a year have grown at a 48% compound annual growth rate since 2022. The ARR from this segment has grown at a 51% compound annual rate in the same period. As the U.S. growth flywheel accelerates, go-to-market efficiency is improving. North American contribution margin expanded to a record 38% in 2025. North America continues to represent a huge opportunity for Trustpilot. We remain less than 1% penetrated into the addressable market. Now moving on to the third priority we delivered on product innovation. Our road map is focused on a fundamental shift in how the Internet works from traditional search to AI-driven answers. Large language models can't afford to hallucinate about brand reputation. They need structured, verified third-party data to ground their answers in reality. Trustpilot provides that essential layer of human insight. LLMs care about the quantity and recency of content as well as its trustworthiness. All that means that for businesses which care about being visible in AI search, being a Trustpilot customer is going to become increasingly important. We're seeing that reflected in the conversations we're having with new customers and in the strength of new sales. In each of the last 2 years, we had one major release window for B2B product innovation. These releases help businesses get actionable insights from their customer feedback and turn that feedback into growth, for example, with new visual assets featuring the Trustpilot brand. The outcome was the strong net revenue retention numbers we reported for both years. In 2025, following the arrival of our new Chief Product Officer, Ciaran Dynes, we established a multiyear product road map to build trust in the age of AI. This year, 2026, the road map is focused on greater AI visibility, expanding review collection methods and platform trust and helping large enterprise clients operationalize customer feedback across multiple domains. You'll see many of these specific features launched in just a few weeks in early April. And there'll be plenty more this year. Thanks to AI accelerating our ability to release new products, we'll have a second major release window in early Q4. We see our advantage through 3 structural pillars that together form an unreplicable moat. First is the vast proprietary data set. We have 361 million active reviews on the platform across 1.3 million claim domains, adding around 200,000 new reviews every weekday. Because people can see that Trustpilot feedback has influence over other users and other businesses, they're inclined to leave more of it. This isn't just a basic star rating. It's often deep, rich, descriptive human content that is unique to us. AI increases the value of this data set in important ways. We're becoming an essential reference point across new surfaces. We've mentioned a few times Trustpilot's importance to large language models. But with the rise of agentic commerce, Trustpilot's role will become even more critical. Autonomous agents will need information to determine which merchants to trust. The best agents will be the ones that use Trustpilot's feedback and our other proprietary data points to make their choices. A key proof point of this in 2026 will be showing our ability to secure merchant trust data partnerships with leading commerce platforms. Finally, AI increases the quality of insight businesses and consumers can get from Trustpilot reviews. By using our APIs, businesses get deeper, better and quicker insights by fusing our public feedback with their private CX data. Our AI summaries also allow consumers to get richer insights more quickly. So to wrap up, I started out by saying that at its core, Trustpilot products collect customer feedback and turn it into influence. As we've outlined today, AI will multiply that influence. In 2026, we've got 3 clear priorities: scaling trust, accelerating with AI and growing enterprise. Our strong execution and structural leverage, combined with the benefits we already see from AI, give us the confidence today to upgrade our margin targets to 25% adjusted EBITDA in 2028 and 30% in 2030. And with that, thank you, Hanno and I are ready to take your questions.

It's Jessica Pok from Peel Hunt. I'll go with the usual 3. In terms of the product road map for the year, how are you thinking about the monetization? Is it a matter of some of the products going into the top packages? Are you going to monetize separately? can we get some idea of that? And also thinking about the future, new pricing packages, was it last year. Is it a matter of a couple of years, you add another top package? Or how are you thinking about those pricing packages and renewal of those? The second one is just on Trustlayer. That was launched at the end of Q3 last year, if I'm right. Just on the progress of that. And then the final one, just on costs. I mean, very clear in terms of your targets for the margin going into 2030. How do we think about the different lines? I mean, clearly, it's going to be sales and marketing, which will be as a percentage of revenue will reduce over time. But the other 2 cost lines, the tech and also G&A, especially as you've mentioned that the number of fake reviews have gone up, but you're managing that quite well. Could we expect those lines to edge up a little bit as a percentage of revenue whilst sales and marketing come down? Just some guidance would be good.

Yes. Let me take the first 2, and then I'll pass to Hanno for the cost question. So road map monetization, I mean, we really think about this in terms of start with value creation, which the features need to deliver and then think about value capture and the different mechanisms for capturing that. When we roll out new features this year, there will be pricing obviously associated with those. And for the most part, customers won't get the new features until they've renewed on to higher price points. But we're sticking with the same essential different packages. So it's not a repeat of 2024 where the whole structure of how we do pricing changed. In terms of data solutions, we don't break that out as a separate line in our financials. A lot of some customers buying the Data Solutions API are also core Trustpilot customers for the business proposition. But the key thing that changed last year is we made the API accessible as a product in its own right. So businesses like consultancies or investment firms can buy the API that gives them access to those 361 million reviews with all of the insights in it without having to buy the rest of the Trustpilot business proposition. We continue to see growth in that, albeit from a low base. But I hope you can see with everything that we're saying about AI and the potential value of this data, for example, in -- as a merchant trust signal across different commerce platforms and for agentic commerce, I hope you can see the sort of strategic importance of it. I don't view data solutions as a kind of ancillary revenue driver purely. It's really about strategically how do we make the most of the data set that we've got. And all of that helps the growth flywheel because as I've been saying, the more impact, the more influence the feedback has, the more motivated businesses are to engage with the platform, the more motivated consumers are to write feedback.

And then on the margins, so 30% by '30, obviously, we've always said 30% is very achievable. This is a high gross margin business. And ultimately, if we don't add cost, the natural progression of the margin flow-through is pretty quick. And so the question is, where do we make the decisions to add costs in the next number of years, and we'll make those decisions, obviously, in these areas where we see the highest return. And so I think what you would expect is G&A to have a natural sort of operating leverage flow through. And then between tech and sales and marketing, we don't want to sort of predefine a shape of the P&L in the next couple of years, but rather sort of look at this because ultimately, whether it's going to be adding humans or tokens, it's going to be somewhere in those buckets.

It's Tim Ramskill from Bank of America. I have 3 questions also. I guess you've been very consistent about your expectations on medium-term top line growth. I just wondered if you could share with your thoughts on how that shape might evolve in terms of does the gross retention you expect sort of to nudge higher. Obviously, you've shared a stat today around what that looks like for enterprise and everything else that follows. So just the sort of shape of things going forward. And then from a working capital perspective, there were clearly some sort of pretty meaningful movements in contract liabilities, which again, I recognize reflects sort of an evolution of the customer base and the billing dynamics. Can you just give us a sense as do you expect more of that going into 2026? Or have we kind of seen the bulk of it, just some sense of scale? And then I guess, just interested in the North American growth opportunity and how you go about tackling specific industry subsegments. It feels like in a market like the U.K., there's almost been, if you pardon the phrase, a little bit of FOMO among some of your customers. So x number of businesses adopt. And then I guess, if you're not adopting, there's perhaps a bit more pressure to think about it. So just how do you go about kind of building that momentum and where are you starting to see success? I know you highlighted Liberty Mutual as a recent win in the insurance space.

Let me maybe take the first and the third, and then I'll pass to Hanno for the working capital question. So top line growth, you can see we're guiding to high teens revenue growth for this year based on last year's bookings. We're not changing our medium-term guidance of mid-teens growth. And you can see that the shape of it, as you say, evolved this year with gross retention improving to -- in 2025, improving to 87%. That's actually a really great stat because it shows that the customer base is becoming healthier. We trended up. I think it was 84% when I joined the business 2.5 years ago. So we trended up quite meaningfully to 87%, which is the result of a lot of effort by the teams, like it's been a real focus in 2025 to improve that gross retention number. And that's obviously a big component of the 102% net revenue retention that we delivered. In terms of North America and the approach, I mean, I think what the evidence shows that we put in front of you today is that the approach is working. So we're not changing it. So we're focusing on large businesses. We're finding strong traction in particular verticals like financial services, health care, education, software, and we're going hard at those verticals as you would expect. We have a playbook for how we do this that has served us well in the U.K., is serving us well in Germany and Italy as well. And we're going about it in the U.S. in that way. We can see as we grow stronger in a vertical, it becomes easier to win additional customers in those verticals because people can see their competitors benefiting from all the great stuff that we do.

Yes. And then on working capital. So I think there's a couple of things that happened. Firstly, we talked about this earlier. We introduced an incentive to the sales team to focus on annual prepayments rather than sort of quarterly or monthly. And especially as we're shifting more towards enterprise customers, they're very much used to paying annually upfront. And so the share of enterprise is growing, the share of customers paying us annually upfront and a new business sales last year was growing and increasing. And that's basically just pulling cash flow forward that is now going to be a recurring benefit because ultimately, these customers renew, especially the enterprise customers tend to renew at higher rates. So over time, you should see a mix shift in the entire portfolio. And we said last year, the benefit was about $10 million, and so you would expect to see that sort of continue and compound over time. So structurally, the working capital dynamics of the business have improved.

Should we go to the back over there and then we'll come to this side.

It's Hai here from UBS. Congratulations on the strong results. I have a couple, please. So I guess, first on free cash flow and buybacks and capital allocation, right? So you returned north of $70 million last year. You have $30 million for this half year. How should I think about capital allocation going forward, given your free cash flow is growing strongly. And you mentioned M&A is the second priority of your capital allocation. What kind of targets -- what kind of M&A targets are you thinking about here? And why is the buyback not the same run rate as previously?

Let me take the M&A part of that and then Hanno for the buyback. So nothing's changed as far as M&A is concerned. There's nothing on the agenda right now. And I've always said in these forums that were we to do it, it would be product add-on, small-scale product add-on kind of M&A. It wouldn't be -- there isn't any sort of large-scale target that we've got in mind.

Yes. So basically, the allocation -- the capital allocation framework remains unchanged, like we said, right? And so if you think about, we've been very consistent in not only sort of returning cash to shareholders from the free cash flow that we're generating, but also looking at the balance sheet and looking at how much cash do we actually need on the balance sheet. And certainly, as the business becomes more and more profitable, that sort of margin goes down, and we're able to return some excess cash also from the balance sheet. We've done that successfully last year. I think in terms of quantum, we've been very consistently sort of doing GBP 30 million buyback. The nuance this year is not all these shares getting canceled, some of them getting put into the EBT to offset sort of future share issuance, but the net impact to shareholders is basically the same. It's just the technicality. And so I think we had a -- we took advantage of a dislocation in the share price last fall and concluded the buyback sooner than anticipated and then sort of did a small top-up earlier this year. But the run rate, and if you look at the sort of the announcement, it's been very consistent GBP 25 million, GBP 30 million every 6 months.

My second question is on the AI.

You had 2? that's [indiscernible].

So just bigger picture, right, in the age of AI, it's easier for mass -- well, AI-generated reviews [indiscernible], right? So how do you see the cost of combating that using your own AI increases over time? How do you see that dynamic? And is that factored in your midterm guidance already?

Yes, it's absolutely factored in. And you can see in our 2025 results, how that's played out. We've delivered rapid margin expansion while doing more than ever to keep the platform safe. And a lot of that is thanks to AI. As I said in my comments, AI is helping us more than it's helping bad actors to do what we do to get better at it and to do it at greater scale with cost efficiency. Let's go over to Gareth, please.

Gareth Davies from Deutsche Numis. A couple around the U.S. for me. The first, a healthy step-up in terms of the U.S. contribution margin, 32% in '23, up to 38%. How should we think about that progress going forward? And I mean, 65% in the U.K. already, is there any structural reason why you think the U.S. won't get there over the next few years? And then sort of the related point is really around the top line opportunity in the U.S. and you're only 1% penetrated into that customer base. Can you just talk a little bit about that balancing act from your perspective in terms of running harder at driving the top line in the U.S., building up penetration versus managing that operating leverage?

Absolutely. Yes. So this is a high-quality problem to have, having a vast addressable market with such a strong value proposition that is unrivaled by any other business in the U.S. economy. So -- but you're absolutely right. It does mean that we have to make a decision as we go through every budget cycle about capital allocation, how much -- particularly how much we've put into the U.S. compared to other geographies like Germany, Italy, U.K., where we're also seeing strong returns. So the way we think about this is we're trying to deliver the best possible performance for the group. So we look at the return on those dollars. We look at the efficiency of our sales and marketing spend, and we take a view about taking everything into account, what will deliver the best possible performance for the group. Now as we've said today, the growth flywheel is improving. It's accelerating in the U.S. as we've seen from other markets, that's one of the key drivers of the efficiency of our sales and marketing activities. So all of that all is in favor of the U.S. But the other high-quality problem we have is that we're also doing extremely well in Germany, for example. So we have a lot of these important choices to make. But that's how we think about it. We're ultimately optimizing for the group.

And just to be very specific, there is no reason -- no structural reason why the U.S. shouldn't over time, achieve the same contribution margins as the U.K. It's merely a function of the retention base becoming larger and larger at above 100% retention rates and then you're adding new business on top of that, but the retention revenue obviously drops through at high margins. And with increasing brand presence and awareness, the retention rates improve, the gross margin improves, the CAC improves. And so everything drives higher contribution margin.

It's Mark Hyatt from Morgan Stanley. I've got 2, please. If I could just follow up on the midterm guidance. Hanno, I know you've talked a lot about the long-term ambition around 30%, your confidence in getting there. But could you just talk a little bit about the confidence that you have to pin to a specific date now? What's really changed to put a date on it? And could you talk a little bit about if you have any AI cost savings, productivity savings baked into that guidance, that would be helpful. That's the first one. And then maybe for you, Adrian, you talked about how Answer engine optimization was partly a driver of that 2H bookings acceleration. So can you just give us a little bit more color on that? What types of customers are you seeing traction in? And I presume this is more of a new customer-led thing at the moment. So how do you plan on driving attach and benefit within the existing base in AEO?

Yes. So I mean if you think back to the -- probably the first slide in my presentation, you can really see the margin progression we've already delivered. And so just by extrapolating this, you will obviously naturally get to these higher margins. I think in the past, we've been more cautious in the way we've been giving guidance, but the sort of 2% to 3% margin guidance that we've given for this year, extrapolating gets you to 25% in 2028 and then 30% by '30 Obviously, we do want to have a little bit of room to make investments in each of these years in our budget cycle. And so this may not be a linear path, but we're very confident we can achieve those targets. And we were confident to put exact dates on this now specifically because in addition to the very consistent track record that we've delivered on both top line and margin expansion, we're also starting to see the benefits of AI that gives us a lot of confidence that we're going to be able, for example, to deliver more product more quickly, more efficiently. We're going to have a more efficient sales organization. We're starting to see it in the finance function, for example, and in other areas of G&A where these tools are incredibly powerful. And so if you think about the natural sort of progression of the top line and then having AI allowing us to not add more cost, we will drive more margin through the business.

And then on AEO as a driver of the business, you wanted a bit more color. A couple of things I'll say. We ran a series of webinars. We are, at the moment, running a series of webinars on this topic of answer engine optimization and, of course, illustrating how Trustpilot plays into that. The first of that series of 3 webinars attracted 10x the number of attendees of any webinar that we've done in the company's history. And you ask what kind of business that is. It's mostly larger businesses, some medium-sized businesses as well because this is a topic in -- I'm sure it's a topic in all of your organizations. It's talking in every organization now how we're showing up in these new surfaces. So we can see from evidence like that, just huge interest. And the great thing is what we are saying about this, gather more feedback, authentic human voices, recency, frequency of the content will help you to show up. That's not a controversial position to be taking in this market, right? If you go out and read any sort of best practice guides, they all talk about this stuff. So we are very much going with the flow here and saying it's really, really important to engage with this platform. I think the final thing I'd say is it's really throwing into stark relief the contrast between what we do, which is open, transparent public feedback and what large enterprises have often relied on in the past, which is this secret NPS sort of systems where you answer an NPS question as a user and you never hear anything about it again. It goes into a black box. All of that is totally inaccessible to large language models, right? So the value of what we do and the way we integrate with those more private systems has just shot up as a result of what's happening with LLM.

Yes, Please.

Ross Broadford from RBC, and congrats on the results as well. Three, please. The first, could you give any more color on a run rate contribution from Trustlayer? or what sort of proportion of sales you think that could reach? Number two, what's your view of where the pricing of the product currently sits? $25,000 still feels rather good value to me at the enterprise level. And number three, has there been a material step-up in the safeguarding ecosystem in recent months to prevent the kind of reviews flagged in a recent research report from recurring?

So happy to deal with each of those. So Data Solutions, we don't, as I said, break it out as a separate line, and it's really baked into the underlying strategy of the business. It's often bought as a package with other parts of what we do. And the overall gross margin of the group is 82.7%. So we would expect Data Solutions also to be a very high-margin business. You asked about pricing. I fully agree. I think our products are great value. I spend a lot of my time telling that to people. As I said earlier, we think about the balance between value creation versus value capture. I think you heard very beautifully put by John Roberts earlier at AO, the value capture that really exists for a large business when you fully embrace what our platform can do. And I still think we're on a very healthy end of that value creation versus value capture share. And then you asked about safeguarding the platform and stepping up our efforts there. So as you can see, we significantly stepped that up through 2025 in terms of the volume of activity, and that means both volume of fake reviews being removed, also the amount of business enforcement activity we do, which is a whole other piece, putting warnings on profile pages when we think businesses are trying to manipulate the platform or have a regulatory warning against them or something else, a significant increase in the volume of that sort of activity in 2025 as well. But I think the important thing to recognize is that's baked into the numbers that we are reporting today that we've just delivered. And part of that is because it really is extraordinary with AI that so much you can do better, faster and cheaper at the same time. So we've ramped up the amount of spend in this area, but we've also done disproportionately more with the effort that we're making and a lot of that is thanks to technology.

Sean Kealy from Panmure Liberum. Two, if I can. Firstly, just on investment levels. Should we be thinking -- so would you give us any guidance for going forward this year? I'm just conscious that I think, Adrian, you mentioned you've got 2 sets of releases this year. And Hanno, you sort of mentioned that the balance between sales and marketing, tech and content may change a little bit. Just how are you thinking about current investment levels and whether or not either we should be looking at either cap dev or the part that goes to the P&L changing at any point in the next couple of years? And then secondly, would you have a gross sort of retention target you'd be willing to sort of communicate for maybe medium to long term at this point?

So on the cap dev, I mean, I think it's been about 3% in the last couple of years, and we would expect that to not go down. We're definitely looking to invest into product. And at the same time, we're making the team more efficient. So I think it's going to be in a similar level in the next few years and not specifically, no. I mean I think the gross retention rate has been a meaningful focus for us in 2025. We talked about this. We've made it part of the company-wide bonus plan. We've made it part of the commission plans for the customer success teams. And we've seen the results. And obviously, we're going to keep focusing on it, but we're not putting a specific target out there.

Joe George from JPMorgan. Just 2 for me, guys. Firstly, I just wanted to ask on any shift you've seen in the competitive backdrop and the barriers to entry that you see for your product and maybe not like-for-like alternatives, but alternatives that enterprise customers may look at in the era of AI search versus traditional search. And then secondly, maybe just on the long-term margin guidance. Can you just talk a little bit about the extent to which this builds in further monetization opportunities and further product launches? I guess I'm asking the achievability of this based on the current product suite versus to what extent it's baking in further launches as well.

Yes. So just to help with this. So the competitive situation I would say if anything has strengthened in the last year. I mentioned already the sort of disproportionate value of public feedback today versus -- because of LLMs versus where it was a couple of years ago versus private feedback because it's readable by AI. But I think what Trustpilot does being an open cross-platform, cross vertical global system for customer feedback, there isn't really anything else quite like it. And so internally, I've been in other businesses where every day, people are talking about the competition. Internally, we're really focused on customers and doing more for them and relying on the kind of natural competitive moat of growing the volume of feedback, growing the volume of customers with all of the dynamics that you're familiar with. And then you asked about product development and to what extent is that baked into the margins. I mean I think you can see with the product development spend that we've got, we're rolling out improvements with, if anything, increasing velocity. So I'm sure we will roll out. We have -- as I mentioned in my remarks, we have a Chief Product Officer with a multiyear product road map. We've got all kinds of ideas about what we'd like to do going forward. And I'm sure you'll see a lot of those developments in the coming years. So it's not assuming that our product set remains static by any means.

I think it doesn't require that, though, I mean, in terms of -- we have a great product that we're selling. We're continuing to build on it. That's going to drive continued gross retention and net retention. But if you just think about the white space ahead of us in each of these markets, there is massive markets like Germany, Italy, France that are still significantly smaller than our U.K. business, which there's no structural reason why there shouldn't be a similar size. And then obviously, you have the whole U.S. opportunity, which should be orders of magnitude larger. And so we can just continue to do what we're doing and compound this business at sort of mid-teens, and it's very clearly achievable.

All right. I think we're done. Thank you very much, Everyone. Apologies. There's a question online. Let's just take that.

It's from Patrick at Goodbody. So in terms of AI, can you define how much you're spending on external software packages? And do you expect to internalize a lot of this with AI tools?

No. Certainly not off the cuff.

All right. Thanks, everybody.