Good morning, everybody, and thank you for joining us today for Trustpilot's Half Year 2026 Results Call. The presentation will commence shortly. After the presentation, we will conduct a question and answer session. If you wish to ask a question, you will be able to submit these via the ask a question button on the Spark Live page. I would now like to hand the call over to Adrian Blair, Chief Executive Officer, to open the presentation. Please go ahead.
Good morning, everyone. Thank you for joining us on the webcast today for Trustpilot's First Half 2026 Results. Hanno and I look forward to walking you through what was an excellent first half, with strong growth and strategic progress. Before we get into the numbers, I want to put on record my sincere thanks to our outgoing CFO, Hanno Damm, for his extraordinary contribution to Trustpilot over the last 11 years. Hanno has been instrumental in taking a startup into the era of profitable growth we're now in. I'm sure I speak for everyone on this call in wishing Hanno well for the future. Following on from that, I'm delighted to introduce our new CFO, Marcus Roy, who joined us yesterday following an extremely successful five years as CFO of The Economist Group. Welcome, Marcus.
Thanks, Adrian, and good morning, everyone. I'm delighted to be joining Trustpilot as CFO and look forward to meeting many of you in the coming days. I'm honored to play a part in the business's future success, and will now pass back to Adrian and Hanno to take you through today's results.
Thanks, Marcus, and welcome on board. Given Marcus's short tenure, Hanno, of course, will be supporting me on today's call. As usual, I'll run through operational highlights before handing you over to Hanno for the financials. I'll then finish up with a deeper look at strategic progress. Three headlines for today's results. First, the business is delivering financially. We grew bookings 18% in constant currency, with particularly outstanding performance in North America, up 27%. This drove adjusted EBITDA to increase by 46% to $26.3 million, and a margin of 17.4%. Second, we're winning in enterprise. Annual recurring revenue from customers paying over $20,000 a year grew 35%, with a host of new product features increasing our appeal to large businesses across sectors. Third, AI is deepening our influence. Trustpilot is the number one cited review platform globally, according to research by Seer Interactive.
We saw a more than 400% year-on-year increase in ChatGPT citations. Let's take a deeper look now at operational performance. As you know, a strong growth flywheel underpins our business. In essence, Trustpilot products collect customer feedback and turn it into impact on the actions of people and businesses. Because the feedback has impact, people are motivated to write it, and businesses want to engage with Trustpilot. Increasingly, that impact extends to AI as well as people. In H1, our users submitted 20% more reviews year-on-year, and our data set expanded to 394 million active reviews. Just after the period closed in July, we celebrated crossing the major milestone of 400 million reviews on the platform. Impressions of our brand on other websites grew 15% year-on-year, and Trustpilot reviews received 29 billion Google impressions in the first half.
This flywheel of more consumers writing and reading reviews, more businesses deeply engaging with the platform, continues to reinforce itself around the world and underpins all the growth you'll hear about today. Bookings grew 18% on a constant currency basis, with healthy growth across focus markets, including an outstanding performance in North America. Key verticals driving growth in the period included lending, insurance, health and wellbeing, and utilities. Gross retention remained at the record 87% level we delivered in 2025. On the new sales side, enterprise momentum is excellent. We're increasingly seeing large multinational brands prioritize AI search optimization. We recently won Expedia as a customer, and spending time with businesses like that has made clear to me that they understand LLMs need relevant, authoritative, and recent review data to recommend their offerings.
Customer logos paying more than $20,000 per year grew 29% year-on-year, and we also saw significant growth in customers signing contracts over $100,000 in the half. As this enterprise base expands, it becomes much more relevant to view our growth through a global lens. With large multinational customers, we're seeing a natural shift in how some of these accounts manage their billing relationships. A dynamic I'll come back to on the next slide. Because of the inherent operating leverage in our model, this top-line growth translated into a 46% increase in adjusted EBITDA to $26.3 million, delivering a margin of 17.4%, up 2.8 percentage points year-on-year. Later, Hanno will take you through where we're seeing the operating leverage come through in more detail. But let me now touch on each region individually. Starting with the U.K., where we delivered bookings of $64 million, up 11% on a constant currency basis.
The net retention rate here remains above the group average, and with a 60% contribution margin, the U.K. continues to demonstrate the highly cash generative profile the wider group can grow into at scale. Sales growth was particularly strong among customers paying more than $20,000 per year, up 27% year-on-year. Notable customer wins included Halfords, Space NK, and EE. Growth was particularly strong in the telecoms, automotive, and electronics verticals, which all grew well in excess of 20%. Our business in the U.K. is incredibly diversified across the economy, with revenue above $1 million in 41 separate industry verticals, and the top 10 accounting for just 39% of total ARR. With overall market penetration still low at around 5%, the runway for growth remains significant in both the number of customers and the revenue per customer.
A growing number of our largest enterprise customers are moving their billing relationships overseas as they expand internationally. Expedia is a good example, a global deal won by a U.K. team, but billed out of the U.S., and therefore captured in our North America bookings figure. Overall, this had a 1- to 2-percentage point impact on reported U.K. growth. It is a reminder that particularly on the enterprise side, customer billing location is not a perfect proxy for underlying market penetration, given how these multinational accounts actually operate. Turning now to Europe and the rest of the world. We delivered strong bookings growth of 19% at constant currency to $68 million, with growth across the region in both number of customers and average contract value. DACH, that is Germany, Austria, and Switzerland, delivered a particularly strong performance, up 30% in constant currency year-on-year.
As a result of this sustained momentum, our DACH ARR has now surpassed $20 million, and Germany is our third-biggest country after the U.K. and the U.S. We recorded our best ever H1 sales from customers paying more than $20,000 per year, and the share of bookings from that cohort in DACH was up 5 percentage points to 59%. We continue to win enterprise accounts across the region, including Bending Spoons and Isybank in Italy. Moving to North America, which comprises the U.S. and Canada. We delivered outstanding bookings growth of 27% on a constant currency basis. As I touched on earlier, this top-line figure is also supported by the expansion of our multinational base with billing relationships in North America. We saw excellent new business momentum and new business sales up 76% year-on-year.
Both sides of the flywheel strengthened in the half, with organic reviews up 50% year-on-year, and the number of paying customers 11% higher year-on-year. We welcome notable enterprise clients in the half, including Bed Bath & Beyond, Intact Insurance, and Sunrun. I will now hand you over to Hanno to take you through the financials in more detail.
Thank you, Adrian, and good morning, everyone. We delivered another strong half for bookings, EBITDA growth, margin, and continued cash generation. Let us go through the numbers in more detail. Bookings grew 18% in constant currency to $171 million in H1, continuing a clear and consistent track record of top-line growth. Our LTM net dollar retention rate was 101%. As you can see on the right, we continue to scale profitably. Adjusted EBITDA grew 46% to a record $26.3 million, a record margin of 17.4%, which is up 2.8 percentage points year-on-year at the top- end of our guidance, and we delivered $16 million of adjusted free cash flow in the period. Our LTM gross dollar retention rate remains strong at 87% through H1. Since 2023, we have placed significant focus on improving gross retention and delivered a 3-percentage point improvement over the past three years.
Within that, it is worth highlighting the U.K.'s particularly strong gross retention rate of 88%. Net expansion of 14% brought net dollar retention to 101% from 103% in the LTM ending H1 2025, as we completed the annualization of previously flagged package migrations. Retention rates of 100%+ reflect the strength of our growth flywheel, product innovation, and our enterprise strategy. This chart shows the compounding power of our subscription model with each new cohort of bookings stacking on top of a highly retentive, expanding existing base. The mix shift towards our highest margin enterprise segment continues to accelerate. Customers paying us over $20,000 a year now represent 47% of total bookings, up from 27% back in H1 2023. The underlying unit economics of these accounts also remain excellent, with gross dollar and net dollar retention for the segment well ahead of the group.
Momentum in the U.S. has been particularly notable here. The proportion of ARR coming from customers paying over $20,000 a year in the U.S. is now approaching similar levels to the U.K., which stands at 49% of ARR, albeit the U.S. is coming from a much lower base. That is a real marker of how quickly North America is scaling and how our value proposition there is strengthening. Let us look at the income statement from a management view down to adjusted EBITDA, excluding stock-based compensation, D&A, and non-recurring items. As always, an IFRS statement and reconciliations are in the appendix. We delivered revenue of $151.4 million and a gross margin of 83%, up 0.8 percentage points year-on-year, reflecting infrastructure and hosting optimizations as volume scales partially offset by continued investment in our customer success teams following the sales reorganization.
Sales and marketing was $41 million, or 27% of revenue, in line with H1 2025. We continue to make investments into new customer acquisition and are improving efficiency every year. Technology and content improved to 23% of revenue, down 0.5 percentage point, as revenue growth outpaced our continued investment in technology capabilities, including our nearshore engineering partnership and new AI adoption capabilities. We rolled out Claude across our tech product and finance team towards the end of the period, and we see this as a foundational investment with potential to expand to additional teams in H2, which we expect to build efficiency gains over time. G&A was 14% of revenue, down 1.2 percentage points, reflecting underlying cost discipline. Whilst this slide presents the income statement as management views it is worth spending a few minutes explaining the non-recurring items on the face of the income statement.
We excluded the previously announced Italian regulatory fine, which we are robustly appealing. We also excluded a provision relating to a historical U.S. sales tax exposure. We identified this following revised external advice on state indirect tax nexus rules, where sales tax hadn't been applied to certain customer invoices. This has resulted in a $2.6 million restatement of prior period and $0.8 million in the current period. The adjustment has no impact on current period trading performance, adjusted EBITDA or our FY 2026 guidance. We delivered record adjusted EBITDA of $26.3 million, up 46%, and a margin of 17.4%, up 2.8 percentage points year- on- year, which keeps us firmly on the path to our 25% adjusted EBITDA margin target in 2028 and 30% in 2030. The clearest measure of underlying performance is cash generation. We delivered $16 million of adjusted free cash flow, up 5% year- on- year.
Our cash flow is very seasonal, with H1 burdened by the company bonus payment and H2 typically much stronger, so it makes sense to look at it for the entire fiscal year. Last year, we also benefited from a rent-free period in the London office and higher interest income on our cash balances, while this year we paid more cash taxes as the business becomes more profitable. Adjusted diluted free cash flow per share was up 13% to $0.039. For the full- year, we expect our free cash flow margin to be slightly below our adjusted EBITDA margin. We closed the period with $21.9 million in cash after returning $25.8 million to shareholders through buybacks in H1, plus a further $16.5 million purchased by our employee benefit trust. Moving down the P&L.
The non-cash IFRS stock-based compensation charge was $9.8 million, up from $6.3 million a year ago. This consists of an $8.2 million non-cash charge reflecting the annualization of high performance RSUs, our new Share Save scheme, new senior leadership LTIP awards, and accelerated executive severance agreements. Additionally, the charge includes an associated Social Security charge of $1.5 million. Despite this higher charge, total diluted share count continued to fall, down 5% year-on-year to 424 million, as grants were more than offset by forfeitures, employee benefit trust utilization, and our ongoing buyback program. We anticipate a similar SBC charge in H2 given new execs joining the team. This brings me to the capital allocation framework, which remains consistent and unchanged. First, we invest in organic growth. $13 million were invested in the business in H1, deliberately deploying capital into go-to-market, product innovation, and trust and content integrity.
Second, we retain flexibility for strategic M&A that can accelerate our roadmap. Finally, we return excess capital to shareholders, having executed $25.8 million of buybacks during the first half. We are halfway through executing the GBP 30 million program we announced in March, which consists of GBP 7.5 million EBT purchases and GBP 22.5 million of outright buybacks. As of September 11th, we have completed GBP 15 million. We will complete this program prior to announcing another share buyback. As noted in the statement, we also recently identified a technical point relating to demonstrated distributable reserves impacting certain share purchases on the prior buyback programs, which we will remediate in Q4. More details can be found in the statement in note 10. Since we started our buyback program in 2024, we have returned $141 million to shareholders and repurchased 11% of the shares outstanding at the program's start.
Turning to the outlook. Following a strong first half, we maintain our guidance for high teens constant currency revenue growth for the full year, alongside a 2 to 3 percentage point increase in adjusted EBITDA margin. We remain confident in delivering sustainable growth and operating leverage over the long- term, given the significant market opportunity ahead of us. Finally, as this is my last results presentation, I want to take this opportunity to reflect on my tenure at Trustpilot. It has been about 11 years since Peter, the founder, and I met for the first time in New York to talk about Trustpilot and its massive potential. During this time, Trustpilot has grown elevenfold from $30 million ARR at the end of 2015 to circa $330 million today, and we have taken it from loss-making to approaching 20% EBITDA margin and created an amazing business.
The opportunity ahead remains vast, and with the AI tailwind, we are just getting started. I feel privileged to have helped lead such a fantastic organization and remain incredibly bullish on Trustpilot's future. With that, I will hand back to Adrian.
Thanks. Thanks, Hanno. In 2026, as we said in March, we have three clear priorities, scaling trust, accelerating with AI, and growing enterprise, all in service of one goal, to build trust in the age of AI. Let me take each of these in turn, starting with trust. We think about this in three parts. First, review integrity. Our focus remains on an AI-led moderation approach complemented by human expertise. We are continuously strengthening top-of-funnel detection to ensure the content on Trustpilot remains authentic and reliable. We removed 3.9 million fake reviews in the first half, a 28% increase year-on-year, with the vast majority of those removed entirely automatically. Second, business misuse. We are continuing to take steps to remove ineligible businesses and warn consumers about bad actors. By bringing together fraud, technology, and engineering teams to build a dedicated rules engine spam domain model, we are seeing exceptional results.
Our automated systems remove domains and detect policy breaches at scale, and we have reduced threat response times down to a matter of hours. As a result, we have seen 83% year-on-year increase in business profile warnings, reflecting the growing strength of our detection and enforcement approach. These warnings do more than command consumer attention, as they actively signal to large language models that a brand may be untrustworthy. The third strand of our approach to trust is regulatory and policy engagement. We continue to see the regulatory landscape around reviews evolving as bodies like the CMA and FTC move into enforcement of rules they introduced in recent years. We welcome this and are well equipped to navigate this changing environment.
For a deeper dive into the entire trust and safeguarding ecosystem, I encourage you to take a look back at the replay of the May 2026 Trust Capital Markets event that is available now on our investor site. Moving to the second priority of accelerating with AI. Large language models prioritize the relevance, ranking, and recency of reviews. A recent research study from Seer Interactive showed that Trustpilot is the number one cited review platform globally. Research from Forrester in July highlighted that Trustpilot often outranks owned and paid media, heavily influencing large language models' opinions of brands and how they show up in results. As you can see on the slide, we recently reached a Semrush Domain Authority score of 100, the maximum possible, up from 95 last year, and a score that belongs exclusively to the world's most authoritative sites.
This is calculated on a logarithmic scale, making a 5-point jump at the top end extremely substantial. Where a business has no Trustpilot profile at all, AI is increasingly highlighting the absence of independent validation as a real trust gap. By contrast, having an active profile drives up to 75 times greater citation visibility than having none at all, making Trustpilot essential for any business looking to enhance their AI visibility. Turning to our third priority of accelerating with enterprise. We are driving an unprecedented level of product innovation velocity with two major B2B release windows planned for 2026, alongside continuous feature improvements in between. in April, we released AI Search Analytics, an in-app review collector, Invitation Optimizer, and custom dashboards. We also launched a new lower cost digital-only starter pricing plan. While still early, this is already showing good momentum.
We have a second major release window this month focused on helping our customers get AI recommended. I want to highlight two specific updates here that are particularly important for enterprise customers. First, multi-domain user management. Previously, user management was focused on individual domains. To help large enterprises operating across large domain portfolios, we are now launching a centralized space to operationalize feedback. This allows tracking of sentiment and performance side by side across every brand in one view. Second, review collection via QR codes. This allows businesses to capture feedback in the moment, creating a faster feedback loop and building review volume. By tagging these codes with specific location IDs, customers will be able to track performance across physical touchpoints. In between, we are continuing to ship updates, including our refreshed product review pages, which I will cover next, and our expanded Shopify partnership.
Back in June, we announced Trustpilot is now a Shopify key product partner, and through our upgraded Shopify 2.0 app, merchants can seamlessly embed Trustpilot directly within their native ecosystem. We have had a basic product review offering for a long time and have just released a couple of major upgrades. First, a universal product catalog pools all product variants, such as size or color, into a single aggregated score, and product review pages give each product its own Trustpilot profile page. These product review pages are currently available across all English-speaking markets, and we are rolling them out in all other languages from the 1st of October. It is important to note that product reviews are not just for retail. Most businesses divide what they do up into different offerings. For example, a bank with different accounts, or TUI, who are using product reviews for specific holiday offerings.
We tested these across a wider cohort of over 400 businesses during a beta trial and saw an over 15x average increase in AI citation growth, generating 2.4 million ChatGPT citation requests. Product reviews are an opportunity for many different customer types to become more visible in AI search, particularly in comparison inquiries. They are available as an add-on module for those already on paid plans. Our focus on large enterprise customers continues to deliver financially. Customers paying over $20,000 a year have grown at a compound 34% annual rate since H1 2023. ARR from this segment is up 174% over the same period. Ultimately, the tangible financial impact of the product innovation we have just been discussing is showing up directly in these numbers, given the immense value customers get from the product.
We're also seeing strong regional outperformance, particularly in the U.S., where 20,000+ customer growth is up 43% year-on-year. As product value and AI visibility grow, our ability to attract and expand these large, high-value enterprise accounts accelerates alongside it. To ensure we continue accelerating this enterprise momentum and scale efficiently, we recently reorganized our sales operations. On the customer acquisition side, we introduced a new outsourced business development model. By shifting top-of-funnel lead generation to external specialists, internal sales teams can now focus on converting qualified pipeline and closing new relationships. On the customer value side, we've split our existing roles to better serve the growing customer base. We introduced a dedicated customer success role focused entirely on onboarding and product adoption. Alongside them, we now have dedicated account managers with a deep commercial focus on cross-selling, upgrading, and driving net revenue retention growth.
It's early days, and changes of this sort always take time to bed in, but we believe this model sets us up best to scale into the opportunity ahead. To bring all this to life, the enterprise momentum, the product value, and the importance of independent trust signals, I want to highlight a fantastic recent example from Navy Federal Credit Union near Washington, D.C. What stands out most to me is how their Chief Marketing Officer, Pam Piligian, describes our partnership. She notes that when it comes to brand metrics, trust is at the absolute top of the pyramid, but a brand can no longer just tell people, "Trust me"; they have to provide independent proof. That's exactly what Trustpilot delivers. We've a short video now to play here.
We are a credit union. We're 92 years young, founded to serve active duty veterans and their families, and we offer a full suite of retail banking products. We talk about brand metrics. Trust is at the top of the pyramid. Trust is the metric that matters the most, and we talk a lot about it has to be earned, not given. I can't say, "Trust me." We have to earn it. So putting our own data on the website that says, "We say 95% of our members are happy," we felt like wouldn't have a lot of credibility. So we wanted to make sure that we went to a third party outside, which is how we got to Trustpilot, to say, "Help us.
We want the data that we get from our members to be visible to people." Without those metrics that really showed we had done what we promised to from a third party, Trustpilot. It was hard for somebody to believe.
Large language models are becoming a really important way that people decide who they can trust. How have you seen the feedback that you are getting on Trustpilot play into those large language models and help you there?
Well, it shows up because it is a third-party endorsement. It is not just us saying what the service was like, but it is exactly our members saying what it was like. So it will show up in those different models. And what we have found is a lot of our content still matters, third-party content matters, influencers matter, as well as these third-party endorsements. Which, for us, Trustpilot is a very high source of credibility.
If you had to summarize the impact of Trustpilot on Navy Federal in just one sentence, what would it be?
It would be earned trust. It is proof that we did what we said we would.
To wrap- up before we open the line for your questions, financially, we achieved 18% constant currency bookings growth while increasing adjusted EBITDA by 46% and margin by 2.8 percentage points. Strategically, we firmly established Trustpilot as the leading AI trust signal. We are the number one cited review platform globally, backed by a perfect Domain Authority score of 100. We are moving faster than ever to capitalize on this accelerating product innovation. The enterprise focus is proving highly effective, with ARR from customers paying more than $ 20,000 up 35% year-on-year. We are excited about the trajectory we are on, the value we are delivering to businesses and consumers, and the huge market opportunity ahead as AI continues to grow. We remain well-positioned for the second half of the year. With that, thank you all. We are now ready to take your questions.
We will now begin the question and answer session. If you wish to ask a question, please use the ask a question button on the Spark Live page. We will pause for a moment to allow the queue to assemble. Our first question comes from Olivia Venancio at Barclays. Can you walk us through why no upgrade, apologies, team. Can you walk us through why no upgrade in the full- year guidance and how we should think about shape in H2? Can you walk us through dynamics region in EBITDA margin? Think U.K. is accretive, but is U.S. still dilutive? When should this change?
Well, I think we've been given pretty clear guidance of 2%-3% percentage point margin improvement for the full- year. We're well on track on this. In the half year, we had 2.8 percentage point improvement on the margin, and we're well on plan to deliver this, as well as reaffirming the guidance for the full- year and the longer-term outlook, of course, of 25% EBITDA margin in 2028 and 30% by 2030. As you think about seasonality between H1 and H2, obviously we talked about the sales reorganization happening in H1. The H1 margin tends to be lower than the H2. We typically have sort of a steady margin progression in the business. If you followed it for a while, you'll sort of see that trend basically repeating every year. I don't think the progression should be a surprise to anyone.
We talked about regional profitability and contribution margin, highlighting the strength in the U.K. in particular, but we also talked about the U.S. being very profitable on a contribution margin basis. If you reallocate your product and tech and G&A expenses, I think you'll also find that then the implied EBITDA margin for each of these regions is breakeven or positive.
Thank you. Our next question comes from Hai Huynh. There's a couple of questions here. Number one, adjusted free cash flow fell from 85% to 61%. What level do you regard as normalized once billing benefits exceptional bonus U.K. tax credits and lease payments settle?
Yeah, like I said in the prepared remarks, I think it makes sense to look at cash flow on a trailing 12 months or a 12-month basis, really, or rolling 12-month basis, and it's always around a point lower than the adjusted EBITDA margin. Last year, obviously, we had some H1 benefits with the free rent. The interest income, as we've utilized balance sheet cash to buy back shares and brought it down to more appropriate levels for the scale of business, has come down. That has impacted the free cash flow. As the business continues to grow, that will be a less and less relevant factor in overall cash generation, of course. The working capital benefits of the prepayments are now at a normalized level, which will sort of offset CapEx.
For your modeling, I think using adjusted EBITDA margin and then taking a point off of that as a proxy for free cash flow for the full- year is a good starting point, taking into account the seasonality. Obviously the implied H2 margin will be materially higher than the H1 margin given those seasonalities.
The second question, what was the parent company's distributable reserve balance in the newly filed interim accounts? How much headroom will remain after completing the current GBP 22.5 million buyback?
I don't have the headroom off the top of my head. I think we're in the process of doing another interim dividend from A/S, our Danish operating entity, to the PLC in September, then we'll continue to file updated PLC financials. Those will then cover the remainder of the existing buyback and provide additional headroom. As we progress through the year and the time, we'll continue to do interim dividends to generate distributable reserves up at PLC and then utilize those for buybacks.
Third question, Trustpilot spent $42.5 million on conventional buybacks and EBT purchases against $16 million of adjusted free cash flow. What minimum cash and liquidity buffer will the board maintain, and what level of EBT purchases should we model going forward?
We don't look at EBT purchases and buybacks as materially different. I think we look at it combined. That's why I referenced the GBP 30 million program we announced in March and the split between those two. We're certainly agnostic to that. The EBT obviously provides a great way to offset dilution from satisfying share grants that have been issued in the past and that are vesting, and to mop up excess liquidity upon vesting when employee vesting happens and there's shares sold to cover the taxes. I wouldn't break those out separately.
In the beginning of the buyback program in 2024, we had a ton of cash on the balance sheet as a result of the IPO, and the business became profitable and cash flow positive, so we generated more and more cash, and we've used excess cash from free cash flow, but also from the balance sheet to bring down those balances. Obviously, I'm not going to get pinned down on a number of what's the right amount of cash. We're managing this. The business, as you pointed out earlier, has seasonality in cash flow. So in H1, we probably need a higher cash balance than in H2. We're continuing to generate more and more cash flow. I think we're actively considering this as the business progresses and grows and continues to become more and more profitable and cash generative.
Thank you. Our next question is from Joe George at JPMorgan. Could you please talk about the disconnect between the very strong adjusted EBITDA growth, +46% year-on-year, and the slower FCF growth, 5% year-on-year? How confident are you for the free cash flow conversion through full- year 2026 as a whole?
Thank you. I'm not sure what's more to say beyond the question I just answered in the prepared remarks, but I can continue to emphasize that obviously we have good forecastability of our free cash flow. In the back half of the year, we are accruing expenses, for example, for the company bonus, which is a burden on EBITDA, but it's not a cash outflow. In the first half of the year, all that cash flows out of the business. We also have, obviously, visibility around the timing of tax payments, et cetera, and we have good forecastability on our lease payments. I'm very confident with the guidance I gave you for the full-year free cash flow of about point below adjusted EBITDA, and then the implied meaningful uptick in free cash flow in the second half.
Thank you. Our next question is from Gareth Davies at DB. Sometimes when a shifting sales model, you get a little short-term disruption. Has there been any detrimental impact in the more established businesses as a result of the sales team shift?
So look, I think our teams have done an exceptional job of performing strongly through the period. You can see that in our results, 18% constant currency bookings growth. That means, I joined the business almost exactly three years ago. I remember sitting here, reporting a couple of days after joining first half of 2023. We had grown by 16% constant currency bookings. So we are actually growing faster now than we did back then. Now, of course, with the sales reorganization, every moment of time and energy that is spent on that is time that is not spent talking to customers. So there is always some opportunity cost, and as I say, these things are always something that an organization has to go through and get out the other side of. So we have been doing that.
But I am just impressed that the team has performed strongly through it. As I say, these changes are all about setting ourselves up for scale on the new business side because we have got a vast opportunity to go into, and we need a mixed model to go after that. So I think about the combination of digital acquisition, which we have expanded with the launch of our starter plan, that is going very well. I think about agentic AI, I think about outsourcing, and of course, in-house capability that we still have. So we have got this kind of mixed approach to go after an opportunity that is absolutely vast. Then on the customer success side, we have got an increasingly diverse and complex portfolio of products. As discussed on this call, a lot of new product coming up.
So we really need that segmentation of roles between CS, that is all about helping the customer use the product, and account management that is more about the commercial side and the renewals. So I think we have done a good job of executing that in the period, delivering strong results at the same time, and setting ourselves up for the future, most importantly.
We have a follow-up from Joe George at JPMorgan. What gives you the confidence that requisite controls are in place to ensure no further provisions will be required in relation to historical U.S. sales taxes?
Yeah, thanks, Joe. We have done a lot of work obviously, in the half leading up to the provision that we made. We retained external advice from a Big Four firm, and we looked at this together with our auditors and gone through this on a state-by-state basis, and looked at the historical invoicing that we had done and where we should have applied taxes. We feel that we have a robust provision now on the balance sheet that will cover the potential past liabilities.
Thank you. Our next question comes from Jessica Pok at Peel Hunt. Is there an incremental cost to having outsourced business development capability? Can you give me color on the shape of the $100,000 contracts and what they normally consist of, e.g., multiple international domains, deeply complex integrations? How are the new clients using your platforms at this level?
Yeah, happy to talk to both of those, Jessica. In terms of outsourcing, the model is more efficient in the sense that you can do more activity at lower cost. As I said in my earlier answer, the whole rationale of this is to have a model that is suited to the scale of the opportunity. When we look at the size of our addressable market across certain geographies, we are still very low penetration rates, particularly in the U.S., but also in the U.K., and our other focus markets. We have got a very long way to go. We have got multiple industries with businesses that have a need for Trustpilot, and we need a cost-effective way to go out and capture that. As I say, that is not all about outsourcing, it is also about digital acquisition.
We're increasingly looking to AI and how we can use AI to go after it more efficiently, along with substantial in-house teams as well. But it's all about cost effectively addressing that scale of opportunity. Secondly, you ask about customers paying more than $100,000 a year. So what gets businesses to those sorts of price points are a combination of a few things. First of all, the sheer scale and size of the customer, meaning that our product is going to have very significant impact on them. Then we've got the grade of the product they're buying. So typically, businesses at that scale would be on our enterprise plan, which contains all of the key features and is the most expensive. Then you've got territories, countries, and then you've got domains or brands that the business wants to use us across.
Typically, once a customer gets to that level, they're using us across the full value proposition. So when we speak to businesses, we talk about build trust, grow, and improve. So you collect feedback, you respond to it to build trust with customers, you showcase your feedback in the Trustpilot brand to help your business grow, and then you look at our analytics products to help you improve. And typically, a customer at that level will be using us across all of those things, making us very important both on the marketing side and on the customer experience side for the customer.
Thank you. We have a couple of questions in from Sean Kealy, so I'm going to break these up. Number one: Clearly, you can provide visibility of citation uplifts for brands with an active Trustpilot profile. Where do you get that data from? Does it come in directly from Google or OpenAI, for example?
Yeah. So the answer to that is we do two things. First, we look at external sources like Seer Interactive, which I talked about a couple of times in the prepared remarks. Secondly, and this is where the 75% figure comes from, we look at our own server logs, which tell us every time we get a request from one of the large language models. Now, those internal server logs aren't a perfect representation of citations because the models are training on our data. They're not just doing real-time inference. And what that means is they're very often citing Trustpilot without it actually hitting our server logs. So the data we see in our server logs is a small subset of overall citations. But we combine those two things, the external sources and our internal server logs, to give us a good impression of what's going on.
The consistent theme across all of that is that relevance, recency, ranking are the key things. In other words, having a large enough volume of reviews such that the LLM can surface an answer using reviews that is exactly relevant to the user's query. Then having reviews that are recent. If your reviews are stale and many months old, then it is much less likely to get cited by a large language model. Thirdly, ranking, which is the authority of the source effectively, which is where Trustpilot itself, with that Domain Authority of 100 we are talking about, the authority of us as a source becomes important. All of that means that basically the more feedback and the more frequent feedback companies are getting, the more likely they are to be showing up in large language model search.
And follow-up from Sean: Is there more that you can do to support brands looking to assess their return on investment on their spend? For example, click-throughs from AI overviews, subsequent purchase behavior, or even trying AI search patterns into spend on the other paid media sources?
Y eah. We are doing more with customers in this area. You will have seen an ROI study that we did a while back with Forrester that showed a 401% ROI on average on customers' investment in Trustpilot. But we are doing more, particularly with larger customers, to help them really quantify the ROI. It is not, of course, as straightforward as simply things like traffic metrics, because so much of the value is from citations or from the growth that they are driving through featuring Trustpilot in their own media. So getting a higher ROI on media budgets. So it is not as simple as just looking at, for example, how much traffic they are getting through Trustpilot. But when you add up all these sources, it is really substantial value creation, and we think that is only increasing as AI takes off.
Thank you. And the final question: Can you run us through in some detail what has happened with distributable reserves and what the likely impacts are, and if you expect to be able to plan or launch a further buyback in the future?
Yeah. The impact on the financial performance is nil. It is a technical issue that we need to cure, but it does not change any of the numbers or any of the cash balances or whatever. Also, I think what is really important to note and state for the record is that the total amount of buybacks do not exceed the total amount of distributable reserves that were available in the period. I think what happened is largely an issue around when and what we demonstrated in the PLC filings in terms of available distributable reserves, and then having done subsequent interim dividends that were not reflected in the filings while the buybacks were ongoing and utilizing those reserves, which happened to other companies as well.
As you will see in the Note 10, in the context of the first buyback, we actually executed the buyback prior to having the reserves, which were then subsequently generated through the share capital reduction in May of 2024. All that was due to a technical accounting mistake where we had erroneously included share-based compensation reserves in the amount of retained earnings that we deemed distributable, which, upon further review, we identified as wrong and they should not have been included. We thought at the time that we had demonstrated sufficient reserves when in fact we had not, and we should have filed interim updates on those PLC financials to also demonstrate them. Obviously we will now remediate this.
There is a well-trodden path on remediation in the U.K., and once it is all remediated, we will continue to. Obviously the business is generating profit and retained earnings, and we can then demonstrate such with PLC interim financials and full year financials that we continue to file, and that will then allow us to continue to do buybacks.
Well, I guess the key point there is this has no impact on our trading, and on our actual cash or anything to do with future performance.
A couple of questions from David Nolan now from Morgan Stanley. Number one, North America new business sales were up 76% year-on-year. Is that level of acceleration broad-based and sustainable, or did H1 benefit from unusually strong large deal timing?
We were thrilled by the pace of growth. I think a few things to unpick. First of all, the market opportunity for us is vast, and the growth flywheel, as we call it, is clearly accelerating. When we look at all of our flywheel metrics, whether it's review volume or TrustBox impressions, they're absolutely flying in the U.S. So we can really see the proposition catching on there. We saw very significant growth in the enterprise segment. There was a tailwind from some of those billing relationships that we mentioned shifting from the U.K. to the U.S. But overall, we expect to continue to perform very strongly in the U.S., which is a huge opportunity for us.
The second question from David Nolan. Group NDR is 101%, while larger customers historically have stronger retention and expansion. As enterprise mix increases, should we expect group NDR to trend back up from here in financial year 2027?
Yeah, there's a lot going on under the hood with NDR. As you correctly point out, the enterprise segment has the most attractive NDR and that's part of the reason why it's strategically our priority and our fastest-growing segment. We also, as we mentioned, launched in the period digital acquisition of our starter plan. That's having a very beneficial increase, at the low end on logo growth. But those low-end logos obviously have lower net revenue retention rates. The way I look at it is we always target NRR of 100% +. What's great about the business today versus a few years ago is it's off a base of gross retention that is far healthier. So we're at 87% GRR, where a few years ago that number was 84%, and that is a much healthier basis on which to deliver long-term sustainable NRR.
What gives me confidence in it is the level of product innovation in the business. We talk, for example, about product reviews, which is an add-on module that customers need to buy with huge value associated with it. That is the sort of thing we are going to continue to do to bridge the 87% GRR with the 100% + NRR that we aim to deliver going forward.
Thank you. We are going to go back to Gareth Davies at DB. Accepting that the U.K. is selling globally and would be healthy teens including that benefit, the 11% felt a little sluggish. Penetration only 5%. What gives you the confidence that this is a business that can still grow teens, or do we now see this as a single- digit 10% growth profile looking forward? How much does the uncertain macro backdrop impact?
Well, think about some of the stats we have been sharing today, Gareth. We saw enterprise growth of what? 27%. We can see our GRR, gross retention in the U.K. is exceptionally strong, above the group average, 88%. NRR is also ahead of the group average. The contribution margin, as Hanno pointed out, is really exceptional at [67%]. What I have also pointed out today is how incredibly broad-based that growth is in the U.K. So we have got over 40 different industry verticals where Trustpilot makes at least $1 million of ARR in the U.K. If you think about that for a second, I used to run a business where we built the entire company off one of those industry verticals, which was the accounting vertical.
We have got paying customers more than $1 million ARR in more than 40 different industries, and that is why we are very clear that the U.K. penetration is still extremely low, around 5%. The business has a very long way to go here. The way we look at growth across regions, is always to optimize group performance, as we say. The question I think to ask is, how are we going to deliver long- term mid-teens bookings growth globally for Trustpilot, and that is ultimately what we are aiming for and optimizing for, rather than the growth of any particular market. As we have seen in the U.S. a few years ago, it was growing more slowly, now it is accelerated. So individual markets will fluctuate, but we always aim to deliver and optimize for group performance at the end of the day.
I should also add, we've got next week our biggest ever customer event in the U.K. coming up, Trustpilot Horizon, where we'll be launching some of our new B2B products that I mentioned on this call. That's had an absolutely fantastic take-up, so we're still very bullish about this market and pushing hard to acquire more new customers.
Thank you. And a final follow-up from Gareth. "In the U.S., can you talk a little around sub-sector momentum? Any sub-sector stand out as particularly strong? Any newer areas of focus that have opened up?
Sure. Yeah, happy to do that. We saw great growth in insurance, in wellbeing, in software technologies. We're starting to see, as in the U.K., that our growth is catching on across different industries. Often the way it works is one or two leading companies in a sector will start using Trustpilot, others will see that, and then it will catch on within that industry. So we're still more concentrated in the U.S. than we are in the U.K., for example. Financial services in the U.S., insurance, credit, banking, absolutely fantastic vertical for us. But that is gradually broadening, and as I say, we saw great growth in software wellbeing, parts of retail, like appliances and electronics, doing extremely well too. But one of the lovely things about this company and running it is we're constantly choosing between attractive options.
We have many different paths to deliver growth, and that includes different verticals within the U.S. market.
Thank you. A couple of questions from Chris Luyckx. Number one, "On revenue, H1 came in at 19% constant currency against a full- year high- teens guide. Given the revenue base for H2 is largely locked in by H1 bookings, is the implied H2 deceleration a statement about booking momentum?
No, I think it's a reflection on this year's bookings performance. As you rightly point out, revenue in the back half is largely a function of bookings last year and year- to- date. I think the guidance remains high teens for revenue, and our bookings growth is mid-teens guidance, and then adjusted the EBITDA margin improvement of 2-3 percentage points. I don't think it speaks to or implies any massive deceleration in any of these metrics.
Thank you. Chris' second question. "ChatGPT citations were up 400%+ year- on- year, 75 times citation visibility for active profiles. Where does that show up commercially? Is it converting into higher win rates or shorter sales cycles or price? Is there evidence AEO is expanding willingness to pay, or is it mainly a retention argument so far?
So in a sense, it's all of the above. When we talk about 35% growth in customers paying more than $ 20,000, almost every conversation we'll have had with those businesses involves AEO at some level, large language models, how they're showing up. Is it supporting willingness to pay? Absolutely. It's helping in those conversations. It's helping with our GRR. I think it's just becoming a core part of our proposition of how we go to market. The great thing is we're not saying anything that is going against the flow. If you look at any kind of best practice around answer engine optimization, what everybody says is you should be engaging with real customer feedback out in the open in a way that the larger language models can actually access. So we're helping businesses to do that in practice, and we're operating the world's largest open customer feedback platform.
Yes, it's a fantastic time to be doing that.
Thank you. And final one from Chris. "On the U.S., you've given a 60% contribution margin for the U.K. Can you give the equivalent for North America or at least the direction of travel?
Yeah, I think we've given a number at the full- year, and we'll probably update it again at the full- year, but it's continuing to improve as the business grows. Keep in mind, though, in particular in periods of very strong new business growth, the margin expansion is not as quick because the new business obviously comes at a higher acquisition cost, which is reflected in the sales and marketing costs, while the retention costs are all reflected in our gross margin. We're extremely pleased with the performance of the U.S., the 75%+ new business growth that we've seen year- to- date, which is a testament to the ability to monetize the AEO tailwinds.
But in periods of strong new business growth, the margin expansion is less pronounced, given that comes with sales and marketing expense and commission expense that are higher in the period, while the revenue from those bookings will then follow subsequently and then drive the margin and retention in the subsequent years.
Thank you. Our next question comes from Roger Phillips at Investec. Why did you do a sales reorganization in H1 in the first instance? Is top of funnel so good now that Trustpilot has critical mass in the U.S. that you can afford to outsource new business generation, or is pipeline conversion something to be improved with internal sales resource? In other words, can you phase this direction completely in different ways? What was the reasoning? Thank you.
Yeah, I think I've already commented on why we did this part that you're referring to on new business, which is all about having a go-to-market that is suitable, that's designed for the scale of the opportunity. As I say, it's not only about outsourcing, it's having a mixed model. The mix is digital, outsourcing, in-house, and we're starting to see good results from adding some AI into that mix as well. So it's a mixed model, but outsourcing was the new component for us in H1 that this reorg entailed. I think as it speaks to the U.S., it just highlights the need for something like this, which is, we're talking about the largest economy on the planet. We believe we've got a proposition that is accessible to ultimately dozens of different industries across that economy.
We're never going to have a large enough group of people in a place like New York, which is where most of our team are based, to go after as much of that opportunity as we can. So it's all about finding the most scalable, cost-effective way to address an opportunity of that size. As I say, it's ultimately a mixed approach, mixing digital, outsourced, and in-house, which is what we've now got.
Thank you. Our final question today is a follow-up from Sean Kealy. In terms of product plans from here, what's next in the business information research or operational insights part of the offering? Are there steps that you could take, e.g., to allow customers to access your audience for market research purposes?
As we said, we have got a lot coming up this year. We have got a major event next week where we are planning to announce some of these enterprise features that we have talked about on this call. Particularly multi-domain in the product is important for some of the largest brands because they will be operating lots of different domains on Trustpilot. At the moment, they have to log in separately for every one. Now they will be able to see an overview with all of their brands. We are improving our analytics products as well. You mentioned how we are planning to deepen those. We are looking at search query insights. So when people are typing something into a large language model, what insights can a business get about that that is going to help them with product development, for example?
That is the sort of thing we are working on at the moment. We have got a lot of ideas for 2027, and no doubt we will have another very busy agenda for next year.
There are no further questions, so that concludes today's results presentation. Thank you all for joining. Have a nice day.