YouGov plc (YOU) Earnings Call Transcript

Good morning, everybody, thank you for coming, and it's good to see you. This is my second time of having the last presentation. I think this really will be the last one. So it has been a good year. We've strengthened pretty much on all fronts. We've got a 2% increase in our data products, which a while back, people were thinking, well, that's all going to be taken over by AI. That will be a theme for me today to say actually, AI is a big benefit to us and a big opportunity. It isn't something that's been holding us back. And you can see that in a very difficult market, and it's difficult because apart from all the other anxieties, there are particular anxieties around our industry because of AI and will all be replaced. In that market, we're growing and we're growing where we most want to be growing. We've been strong on controlling costs. You'll hear a lot more about the VDP from James, and there are some other interesting things that he will talk about. I will just focus on my role, which has been -- when I came in 2 years ago or 1.5 years ago, I said I was going to do 3 things in that interim period that I had specific objectives. One is to stabilize the business. I think you can see our numbers have done that, and I'll give you some more details about that in a moment. The second is to return to innovation. And the third, of course, is to prepare and deliver the leadership transition. And on the first of those, we've seen accelerating growth in our key areas with 3% up versus last year in data products as well, we've done particularly well. We've got that cost discipline I referred to and also we have made that switch back to focusing on data products from custom. It's not the custom isn't important. Custom has also become very much more data product than it used to be. It's less about the customization and more about the richness and fullness of the data. So even our custom trackers are really very similar to data products. And we're focused -- we had a unit that focuses on developing that data set is something that can be partnered with and a bit more about that later as well. Innovation, we have accelerated. We've enhanced the product. There is now an AI agent that can tell you about what's in the data and help you find more, and that will be growing. And I mean, everybody has that, some sort of an agent that looks at your data. We have a heck of a lot more data than anybody else and by an order of magnitude. And to get the most out of that is what we're doing. We're adding MCPs. Of course, all of that above all as you'll see later on, we're making it a platform that an agent can use directly. We have -- those are our AI-powered solutions and the platform we've long talked about, now it is becoming really a full-on data platform. Again, I'll show you later on at the end where that's going. And the third priority was that leadership transition. And I said that we wouldn't make the same error twice that we were looking for somebody who has proven leadership in adjacent area, where we can see what they've delivered. In Wayne Levings, we found that somebody who has led teams grown P&Ls knows the industry inside out and also understands the challenges of AI. We have somebody that we know is going to do this job well. I also said that instead of just handing over, simply, we would have a proper transition period in which the new person learns the ropes a bit and not just expected to immediately fit right in because there's a lot. We are a complicated company. We're different to many other -- most -- all other companies in our industry, and it does need a proper transition. So I'm looking forward to that period. I have a very strong relationship with Wayne and I'm very confident that we're going to be working well together, especially since I have a new role in the post-January period when I go back to being a Nonexecutive Director, but overseeing running, chairing the innovations committee, which will be very active in helping to direct and guide the things happening that the executives are doing with the system. So handing over for the financial review to James.

Hello, everyone, and thank you, Stephan. This is my second results presentation since I've been at YouGov. You may recall my first one was in the half year after just 5 weeks of joining. And the last 6 months have definitely been fulfilling, we made deep progress on the planning and execution of the value delivery plan. we're investing in areas that matter. We are growing up now, yet we are keeping the entrepreneurial discipline alongside all of the decisions we make. And alongside all of that as well, we are keeping innovative ideas flowing. And they're not just flowing, they're flowing and they're converting as well. And Second, we'll talk about that in more detail later. So before we get into numbers, I wanted to highlight the key priorities of H2 from my perspective and a wider board perspective as well. [indiscernible] 7 of these priorities. Priority one, from a market and investment related perspective, we want to do what we say we're going to do. I've been in the public market for a long time and I know how important that is for employees, for external analysts and investors. So we commit that, that is what we're going to do. And this obviously included being within the guidance range that we announced in March, plus live on the other areas as well. Priority two, rebuild trust and in food transparency and clarity where possible. like this presentation, I will refer to things that we committed to at the H1 results and even confirm the action is completed, we'll update on how we're getting on in solving that problem. And going forward, there may be times where we decide an adjustment to previously communicated commitments is required. It's important that our trajectory may change, and it's important that we're agile and we chip in the changing environment. If our direction does change, we commit to explain clearly what that change is and why that change has happened. Today, there are no such changes to commitments to disclose. We are bang on track for everything that we have committed to in recent times. Priority three, balance sheet stability. In March, we had a fast approaching maturity date for the debt facility. An extension is therefore critical. This has happened. Priority four, obtain approval for a share buyback program and start the program. I'm nice to confirm that the buyback starts today, I believe actually already has started. And over the next 12 months, up to a total amount of GBP 15 million will be returned by this REIT. Priority 5, evidence continued solid execution of the value delivery plan. Wave 1 was executed at pace before the H1 results. Wave 2 and 3 are well underway. Planning is advanced and execution is happening. Priority 6, initiate the wide-ranging strategic review of our software division. Again, that value is ongoing. And priority 7, provides clear guidance going forward, no vagueness or confusion. And in the release, you'll see very clear guidance in pounds millions of numbers with no confusion venous or can despite approach to what we think will happen. And I'd like to see that our results presentation at least a few hours ago, and our presentation were given today provides high conviction that we have achieved 7 of Tier 1 priorities. And please do challenge me or Stephan, either today or later if you feel we've not achieved those objectives. And I've circled 4 of these points, 1, 2, 5 and 7. These priorities aren't just the last 6 months. These priorities will last indefinitely. That understanding we're youngest tech-focused listed business, and we don't have an option and we make sure that we will embrace that in a very growing up and clear way. Before I move on to the segmental operational results and with priorities 1 and 2 in mind, i.e., do what we say we do and then transparency and clarity, I wanted to cover off how we reconcile between adjusted and statutory group numbers. We have rightly been criticized for lack of clarity in this area before. and I committed in the H1 results in March, but the thresholds for items be classified as separately identified will be higher. And also, we'll be much more transparent on the makeup too. I see that understanding the makeup of the separately identifiable items is as important as the total quantum itself. Hence, our commitment to priority 2, in particular around transparency and clarity. Due to our history of M&A and restructuring initiatives, we have required the use of separately identifiable items. So how are these reported items be built up in FY '26. I set them into 3 buckets. Bucket 1. I want to ensure that our balance sheet holding value for assets was thoroughly reviewed as part of my first year-end at YouGov. A large chunk of this impairment balance is the write-down in the balance sheet value of our DAP business. As a reminder, that includes our non-Shopper German business, but also includes the Swiss business, including CASI Research. This is a business that runs our computer-assisted telephone interviews. The balance sheet valuation needed to be adjusted downwards, which we did through the numbers at the year-end. The remainder of this category includes lease impairments as we assess simplicity and efficiency of our global office footprint as part of value delivery plan, along with some impairments of our software assets, which have been capitalized in recent years. Going forward, a continual firm view will always be applied on balance sheet holding values as our business evolves. Bucket 2, amortization. An accounting rules led amount, which has fallen off from our shopper acquisition. This relates to 2 items: Order backlog and customer list. The order backlog was fully amortized prior to the start of this financial year. And the amount of GBP 5 million from last year from the prior year will not continue. That leaves GBP 10 million this year, which covers the customer list amortizations. In FY '27, we expect this aspect to continue with another GBP 10 million amount for further amortization of the customer list. This customer list amortization is expected to continue for the foreseeable future at broadly the same amount of this year. However, the order backlog will not come back. So circa GBP 10 million is going to be the number that fills buckets to you going forward in this category. Moving on to bucket 3. This is the only cash item within this is grouping and arguably the more relevant line item when it comes to assessing underlying value. Buckets 1 and 2 are noncash. In high level terms, Bucket 3 covers restructuring costs. These fell from GBP 13 million last year to GBP 5 million this year as the threshold for inclusion increased and the vast majority of transitional services agreements for the Shopper acquisition have now been completed. Therefore, the cash amount of SRIs as more than half year-on-year with a reduction of GBP 7.7 million. Going forward in FY '27 and beyond, Bucket 3 will continue to be material as we deliver the value delivery plan. Bucket 2 will be GBP 10 million, as I discussed, and bucket 1 will be assessed throughout the year to ensure all balance sheet values are appropriate. The GBP 33.1 million total explains the difference between adjusted and statutory measures on the face of the results announcement. Adjusted EPS at 30p is down 11%, which is broadly in line with the reduction in adjusted operating profit. As a reminder though, as you'll have seen from the numbers, H2 adjusted operating profit was marginally up year-on-year. So the EPS downfall related purely to performance in H1. Now moving on to the segmental operational results. As Stephan alluded to, each of our 3 divisions delivered headline revenue growth. Yes, growth levels were low single-digit percentages. Although compared to the majority of our peer group, actual year-on-year growth in itself positions us towards the top of the pack. On an underlying basis, Data products and research grew 3%, and there was marginal in shrinkage in shopper. FY '26 is our 16th unbroken run of group web revenue growth, an achievement that our teams around the world should be very proud of. We expect FY 2027 to be our 17th year of unbroken revenue growth. We're particularly pleased with the margin uplift in data products, where operating leverage continues to kick in. I mentioned in H1 results that our data products margin needs to improve. Everything I was seeing in my first few weeks suggest that this margin should and can be higher. And this is a great step towards achieving the margin, the syndicated subscription-based product fully deserves. The margin in research was slightly down. This will improve and go back to last year's level and beyond as a value delivery plan delivers. Shoppers margin is materially I think were different or lower, either way, it's lower, relative to last year. And this is exclusively to do with the game change investment program we discussed at the H1 results. If you were to add back this specific investments and palatine investments, the margin would be flat year-on-year. Moving on to our geographic performance. Very strong performance in Americas. Ray and his team seam in the U.S. should be very -- had a very encouraging H2. All other regions are showing solid resilience in what we all know is a tough market environment. The vertical sectors in which we have been growing in the last 6 to 12 months, is particularly pleasing. There have, however, been some tough sectors, as we have discussed before, agency is being one, and they are weaker, and they have declined year-on-year. So with sports, our sports branding remains very strong. However, there have been some operational challenges and personnel changes, which have led to a drop off. We see no reason why this can't continue to grow in the future. The academic vertical is also down year-on-year. Although this is not a like-for-like decrease since FY '25, included the uptick due to the November '24 U.S. election work, and we're confident this trend will continue in the future. However, as touched on by Stephan already, and I'm sure he'll talk about it in a lot more detail later, we are all particularly pleased with the notable new client wins from the hottest, most advanced tech businesses. who come to us because they want the best data points and they appreciate and understand the benefits the best data and insights give you. Revenue from the AI Labs themselves are showing nice growth year-on-year. The other vertical, which is gaining keep traction is financial services. We're up nearly 20% year-on-year in this vertical. And we've got some great new blue-chip global financial services clients with a very high and deep tech focus coming to us in the last 6 or so months. At our H1 results, in March, I said that we needed to improve our cash conversion. And I'm delighted to say that we've increased this to 76% compared to 71% in the prior year and a mid-60 percentage outcome at the half. I'm not yet comfortable giving a forward-looking target externally, but I'll give you a commitment that this will continue to be a key area of focus in FY '27 and beyond. And CapEx has increased year-on-year for 2 main reasons -- for 3 main reasons, I guess: Panel investments as part of the Shopper investment program. A big chunk of our investment is in modernizing and improving the depth and strength of their panels. And we have been increasing investment in technology to enable the next steps in automation and product capabilities. The value delivery plan is not just about efficiencies and savings. It's about investments in the right areas, too. We have a strong balance sheet. Are we well capitalized with ample liquidity? The answer is a clear yes. As at the end of September, cash was GBP 48 million, and we had GBP 8 million of undrawn RCF. At the end of June, these numbers are GBP 42 million and GBP 6 million, respectively. We are generating cash and continue to generate cash, hence, our confidence of launching a share buyback program today. During the year, we successfully extended the debt facilities during H2. There is no additional funding, no increase in borrower or garments obligations, no change to pricing, covenants or security. Total commitments remain at EUR 224 million comprising EUR 184 million to term and EUR 40 million of RCF. The one change we did get through with the banking group was to allow us to do a share buyback program. You may recall at the half year, we couldn't do one because we needed to see a formal approval. As Stephan has already mentioned, as I say, the buyback has started today and will be up to GBP 50 million as already disclosed over the next 12 months. As guided at the half year, net debt to EBITDA reduced back to 1.7x with interest cover over 6x. And these covenants will be improving over the following reporting periods. Value delivery plan. We launched it in March, and momentum is building nicely. The transformation plan comprises 3 waves. The waves combine automation of key processes, efficiency improvements, simplification, enhanced water working, along with significant transformation enabled by AI, plus Wave 3 also includes significant new revenue streams as well. As you will recall, Wave 1 completed in March '26. We delivered an annualized GBP 2.5 million reduction in our cost base, which targeted predominantly headcount and other cost reductions in nonrevenue-generating areas. H2 26 includes some benefit from this action already. FY '27 will benefit from the full year impact. Wave 2 is deep in the detailed planning and implementation phase. It includes a broad range of initiatives. Automation and simplicity are 2 key themes that already disclosed. The combined impact of the first 2 ways is expected to deliver an annualized adjusted operating profit margin uplift of 350 bps over the margin achieved in H1 FY '26, which was 12.3%. We have high levels of conviction that this target will be achieved prior to the end of FY '27, another example of delivering what we said we delivered a few months ago. Wave 3 will focus on opening up new revenue streams through development of AI-powered products and our data as a product capabilities. Our pro development, data science and technology teams continue to make strides in these initiatives, including introducing imputation of our data sets and developing newer products such as YouGov Powalax. I know Stephan will enjoy talking about it and adding more detail on these shortly. Wave 2 and 3 comprise 12 individual work streams each with an executive owner to optimize the speed and festive delivery. I was going to name the 12 names, but there are 12 and it will take some time, but the everyone is doing a great job. And I personally -- I know the Board is very grateful for the efforts the team are doing to really push on with this big transformation initiative. So final slide for me. Coming back to the priorities I mentioned initially, I want to give more clarity and granularity on trading and outlook. We've had an encouraging start to the year, bank in line with our expectations and our budgeted numbers. Currently, on FY '27 revenue expectations are contractually secured. That's not a weak pipeline that is contractually secured. That is marginally up on the same period last year. Hence, the confidence earlier about FY '27 being our 17th year of revenue growth despite the market conditions. We're making a loan payment this month. The last day of October, we will pay back EUR 20 million of the loan. We've disclosed our share buyback already. Investments in Shopper will continue in FY '27 at broadly the same levels as FY '26. Although the benefits of the investment program are expected to start flowing through from a revenue perspective towards the end of FY '27. Based on the significant progress achieved in planning and executing the value delivery plan, over the last few months, along with continued investments in Shopper, the group expects to deliver adjusted operating profit of between GBP 58 million and GBP 61 million for FY '27. This is in line with current market expectations. And this represents a year-on-year growth of around double-digit percentage points. I'll stop there and go to pass back to Stephan, who can talk about the exciting stuff now.

Thanks, James. And it's called strategy update, which sounds like it's sort of a distant prospect. Everything I've ever had in the strategy update becomes reality very quickly. It's actually a road map of what we're doing. It's not some vague thing of this is what we hope to be doing. This is what we're actually doing. And the story of it is that we're benefiting from AI in 3 ways. It is the opposite of what people expected. AI is very much our future. The first, which is very narrow is -- I mean this whole thing is a new vertical, essentially for us, AI in terms of companies, obviously, 2 new companies, Anthropic and OpenAI are the 2 main ones. They become big clients of ours. And just very narrowly, we've done GBP 4 million just in traditional reputational work that they use to communicate -- to understand audiences, but also to communicate. Anthropic has made a big deal on a very large surveys that we've done for them, which connect them to their users and to the public in one of the ways that are central to their strategy. So their spending there is already the biggest growing part of specific of our business. And that number doesn't include lots of other stuff that is adjacent to it, things that they need to know, things that are consequence of news items and so on, and I'm not talking specifically those 2, I'm talking about the whole sector. That comes to about GBP 11 million now in sales over the last 12 months. So that is really a big deal for us. They're not -- they turn to us -- they don't do it themselves. They -- of course, they need an outside agency. They need to talk to humans. They do that mainly to us than other companies as well. But the idea that human research is replaceable by machines is simply wrong. The machines really need the humans to supply that data. So that's the first bit. The second is internal efficiency. Obviously, we're using AI to make our systems more efficient. I'll go to that slide in a second. And the third area, the most important is the AI-enabled products, really, it's a transformation in products and in how we deploy data. So that second box, automating research at scale. We are looking to make an end-to-end automated research flow. That doesn't mean, by the way, that the humans disappear. I actually don't think they will even necessarily reduce, I think, that we will grow in scale and be able to do much more with the people that are there because the handling, the helping clients around these flows, getting the most out of the system is something that remains very important. So I just say that to ease concerns that people might have at the company that the jobs go I don't believe that I think we'll see growth getting scale out of our systems being the way that we most benefit from the advantages. The advantages are also higher quality in some respect in terms of speed, consistency, accuracy, all of those things as humans make mistakes, the more that process that we can automate, the more of it we can streamline before it even gets into the system because at the moment, it's typical that people come with their projects and they do them any way they design them in any way that they want, and then we make that happen. And it's highly customized. But the advantages to them as well as to our own efficiency is to streamline how they think about the work and how that works its way through the system and how that helps to reduce error. So the whole process is one that benefits all sides. Paralax, our new invention, this time last year, I talked about voices. Voices has become a successful part of our of our offer. We now have just launched Paralax, which has also had its first 6-figure client confirmed. Paralax uses synthetic data, it's synthetic data together with human data. You all know what the Paralax effect is, having 2 eyes means you get 2 views of something that together make a 3-dimensional picture. That's what Paralax means. And that's exactly what we mean here. You get 2 views of the world. One is a narrower one, if I can put it that way, narrow but accurate and deep which is the human view. We can -- we talk to 1,000, 2,000, 10,000 people that we know well and that we have lots of data for. And the data that comes out of that is the most reliable human survey data in the world. And that is the foundation. That can yield -- then that can create twins, synthetic wins. We have a partner, original voices that helps us with that, that is experienced in that and has operationalized that in various ways to help get the maximum value out of that data. And that can be whether it's our data, BrandIndex for Cube, all of that, to make the most detailed digital twins you can imagine, or whether it is somebody comes along does a survey with us, and then we turn the participants of that survey into digital twins that can answer lots and lots more questions in the future. But just having those 2 things separate isn't what makes the real 3D view. It isn't just having either or combining the human view and the synthetic view. It is combining them through verification and validation. So at the same time, as you do, you generate questions, answers from the synthetic twins. Paralax allows you to take some of those questions. Let's say you're asking 100 or 200 questions of the twins. It doesn't matter how many you ask. It's very, very efficient. It's very instantaneous. So you got no problem. You can ask as many questions as you like. You take 2 or 3 of those questions and run them against the human version of those same twins. At the same time, you can then use that to calibrate all of the other questions. So you're not -- it's not only 2 different data sets. This is the first data set that you generated synthesis that is calibrated by the human data that comes after. I don't think anybody else offers this or can offer this the way that we can because we can talk to the same people at the same time as getting the twins because you have high response rates, and we have lots of data that we create. So our twins are constantly updated. They're constantly available, and you get this thing that answers the central problem of today, which is how good is synthetic. Do we have to throw away and just have human data? No, you could actually get the benefit of both in a way that strengthens both. So that is a fundamental change to an operating system. That is a fundamental offer that we make to the industry, which I don't believe anybody else can make. And that will be something that goes through somebody previously asked us how much you expect to make from this next year. Very hard. I'm not going to give an answer to that question, as you already know. But I would say in this case, precisely because it affects everything else we sell. Every single client have twins of the work that we're doing for them in custom or in the syndicated part. So that is a big deal for us. That transforms the whole offering. And just to remind you of our ecosystem. It uses all the assets that we already have. I mean, we did this a year ago, what are our assets? We have a high reputation with the public as well as with clients that allows us to have the best panel, which produces the best connected data, which then allows you to do pretty much anything you'd like because if you start with engaged live panelists, you can do all sorts of things, you can experiment and you can you can verify and validate synthesize everything. So that's the basic offer. You've seen this before. The red stuff is what we've had. The 2 new parts there are voices and Paradox. They are all different entries into our world of data. They are -- you can go from one to the other, you can add them together. It is a connected system. And that leads to -- that connected system leads to a seamless client journey throughout the whole world of data. This system is already -- you are able to query it using agents. We have MCPs. We have on our site, you're able to ask questions of the data. That is at an early stage. I will not say to you that this is -- I do say in development. You can right now, as a client, go on crunch, ask questions of the cube and see what that is. And that's -- but that's almost table stakes. I said before, everybody else is doing that as well. The difference with us is the sheer quantity of data. But where it becomes exciting is that the agent can not only say here's what the data says, but do you need some other data, we can get it for you within 30 minutes, and you can write your own questions or you can have the LLM -- your own LLM or ours or whatever we use in that place to write the surveys for you from all the knowledge that we have to answer your questions. then we can use synthetic and Paralax at the same time. In other words, the agents can arrange the entire journey to give you the most full , the most connected and the most verifiable data that there is in the industry. That is our system. And what you see there, all of that exists, except for the Green Circle. The Green Circle is obviously an advanced AI system. We are not advanced creators of AIs, and we will be looking -- you have in mind someone like Thomson Reuters who's made steps in this direction, creating their own domain-specific foundation model. That's the area that we are very realistically looking at, very specific. We're not saying we're going to be a gigantic AI company. That's not at all a thing. We're talking about in our industry, focused on how clients of market research use data in their businesses. That's our expertise. That's where we are focused on, and that's where we want to create the most advanced agent model that there is which we can because of all the things that we already have here. I do repeat this model of being able to look at the data we have and instantly verify and add to it is from the same people is, I think, unique in the industry, certainly at the scale that we can do that. And we do have partners in this. So I know some people have said some of very good investors who communicate with say, are you doing the Reddit thing? Are you selling lots of -- are you selling your data. That is a very dangerous thing for companies like us to do. If you have vast amounts of unstructured, if I may call it, dirty data like Reddit, then the AIs make value out of that by taking all of it, we can't possibly give all our data to somebody else to turn into something that people can buy through them. It's our data, and we have to jealously guard it. Of course, we can share it through the systems, like MCPs that can be controlled through agents that can be channeled in the right way to use that data. And that is what we're doing. But we are looking at partnerships here. I mean we already have, it's known to you a very strong relationship with Deloitte. Deloitte have upped their level of partnership with us significantly. And for the trust model, this is not a new thing. I think it's out there somewhere, but that's a developing an important developing relationship, and we have other relationships that we're developing in order to get the value out of that and maybe actually some give and take in there as well that we can gain from their expertise. So this is the way forward. This is a client journey that I think nobody else can mimic exactly and that is what already most of that already exists. So I end by saying we got a framework for the next phase of YouGov's transformation. First, the new generation panel ecosystem. I haven't said anything about that. That's actually my big final task in the next 3 months to make a bunch of changes that we have been making we have some new methodologies for panel and I'll tie that up and that be for somebody else to present subsequently. But how we recruit and maintain panel is fundamental, obviously, to everything you've seen here. I think we're the best at it, but we need to be better because people are changing young people are engaging less. We all know that. It's getting harder to engage people into a continuous relationship, and we're doing new things in that area. The second AI-powered platform transformation. I've just shown you that in some detail. The third part I've alluded to before, but it is saying this is a platform company. This has a certain kind, is looking for a certain kind of work, and we have a unit that is focused on those partnerships that has had its first successes. We are -- and this goes to all parts of the business. This isn't data versus custom. The difference between custom data, custom projects and data products isn't the same anymore. They merge together. And I think one piece that I want to mention that I should have mentioned before, when even the data products all survive at this as well because they are data as a service. They continue to do really well. And I've said this before, but it's continued that the very -- the most sophisticated biggest companies are increasing the size of their subscriptions here and probably the most scientific big name company in the world now has a $1.5 million subscription to a BrandIndex. So far from being becoming irrelevant, essentially, that data is becoming more valuable, not less valuable. And then finally, the operational efficiency and simplification we've talked about making all of this as automated as possible so that our experts can also give the best possible advice and service because different customers need different levels of service. That is what it's all about. That's the end of that bit. We are now going into the Q&A. There is a microphone coming around, and I'll take that from...

Sean Thapar from Canaccord. Three questions for me. I guess -- to start off with a big picture question is sometimes hard to get kind of a client-led view on spending. So across the industry is pretty low growth rates. -- they're pointing out industry competitors pointing out AI-led growth, but there's a lot of AI washing that's going on there. Could you just kind of explain to how clients are approaching their research budgets spending areas where you're seeing demand and maybe where there has been a little bit of disruption from AI.

Yes. And there certainly has been disruption. I think I said at the beginning that you lose in some areas, but we're gaining in others. And when I look at the growth, it actually represents that loss on the 1 side and the growth on the other, it's a much better growth for the bits that we are obviously most focused on. And 2 things really. Clients are expecting to get more or less. I mean, that's the basic assumption with AI is that we are already benefiting from the magic of it. Obviously, we're not -- we are not at the point yet where we have the savings that are coming. I think this is true around the whole world in all industries. It's for most suppliers they're not yet getting the advantages of the efficiencies, which the clients are wanting to bake in. So that's 1 pressure. -- it hasn't actually hit us badly yet because we have the other say yet. We are moving very quickly to get those advantages. That's what the automation program is all about. -- and we hope actually to get uplift from it because we will be the best suppliers of that automation and can, in fact, respond to those demands. But the other area is that the need for human data is growing. So you're seeing -- that's why you're seeing a lot of companies wanting more of that wanting to make more out of it, and that's where we're seeing the growth. So where we're seeing loss is, for example, media agencies, they're spending less You'd expect that brands, I think, are reasonably stable and the AI companies are -- or anybody doing an innovation in that area is growing.

And then maybe coming back on those AI companies is -- so the AI labs out are some of the biggest companies in the world. from a kind of research spending perspective, how do they differ to the kind of MAG 7 type companies that you have already got a very big contribution from in your client base? And I guess the -- so question on top of that is, do you expect to be able to capture the same share of wallet as you have done in previous large tech platforms?

I mean one of the funny things about it is that those clients come back to you and tell you how you should be doing a business. And they say, well, actually, we've looked at how you structure your things could get the sneak look, I don't know how -- and say, why didn't you do it this way? It would be much faster or much better. So there is -- they're very -- they can be very demanding in a good way. It forces us to innovate faster and to listen and to see how we want to deal with that. They are much less interested in your advice in our advice, our consultancy, I think, is going to be assumed to come free because actually, they don't need it, and they don't often even care how you're packaging up the data. In fact, they tell us we just want it more. And so yes, this is changing the demand -- but what we want, I mean, where we make our margin is in quantity of high-quality data. The bit that they can't get from anyone else is the actual relationship with humans that is a trusted personal relationship actually, and you need that, I think, in order to create great data.

And then a final question for me is just kind of tying together the strategy update with the value delivery program. I guess if we were to take research automation and additional AI revenues, do they map to a certain wave in value delivery program or to ask this a different way, is the 350 bps you're getting from Stage 2 dependent on why scale research automation and/or additional revenues?

I think it's the 2 bits together, but...

Typically, the automation is within Wave 2. And if we get the full wave of automation, which is not expected to happen all in a that will be a benefit above and beyond the 350 bps for the full automation, but there is a part of automation within that number. And the new revenue streams are in Wave 3. There will be no new revenue in Wave 2.

It's Lara Simpson from JPMorgan. My first question was just to come back to ag products. Obviously, we saw quite strong acceleration growth in the second half of the year. I know you spoke about the U.S. But could you talk a bit more around the nature of this client and contract wins, and I suppose the sustainability of that, if they were more one-off in nature or a bit more recurring? And I suppose the follow-up to that is trying to think about the building box to growth. in FY '17. And if you think that strong exit rate is sustainable on the data product side, in particular? And then my second question was on Shopper. I know you said that the strategic view is ongoing. Just wondering if you could give us any indication on when we should expect to get an update on that or otherwise how far discussions are you? Or if the Board has at least given an internal time line on when they would expect a decision to be made? And with that, just interested again on the rationale for continuing to invest in that business. I think James said the investments will be similar to 2026. I think the one discussion I have is you're obviously investing in a business that might not be with you on a go-forward basis. So just trying to understand the rationale for those investments and how you're thinking about yielding the return on those.

Well, I think on the latter, I'm really not going to say anything about the timing, it's obvious that I can't and won't. The other bit of that, though, is that shopper needs to remain strong and yielding growth and that requires investment. So if you wanted to sell something that had gone flat, that wouldn't work either. So that simply fits the plan. It doesn't say anything about what we think is going to happen. It just has to be. The first question, which I think you're saying how -- are there new clients for data products?

I'm just trying to under data products accelerated, I think you do close to 7% growth in date products in the second half of the year. So just trying to understand the drivers of that acceleration.

Look, it is a really efficient way of understanding something that you can only do with human data, and that has changed. One of the things I meant to say earlier that it's fundamental to understanding our business and how it's affected by AI is to know that most of what we do is to measure change. We obviously do some static poles and studies and so on. But that isn't the core of what we do. We measure change day in, day out. by lots of different products and by big customer factors. All of those things are about understanding how today is different from yesterday. And the 1 thing that AI and synthetics cannot ever do is tell you how things are changing. -- accept retrospectively. If they are to project forward, and I think a lot of synthetic data companies, like tell us they can because they're a brilliant and they know how the weather is going to affect everything and all these other mysterious things. If they could do that, -- they could only tell anything about the future by continuation of patents from the past. They might say if things continue as they look today, it will go this way. That's not why you buy trackers. You buy trackers to see what actually happens. -- and you look at you're interested in surprises. So this product that we're launching the testing well and theoretically should go well, how is it actually going when people take it home and that gets play with it and the kids friends play with it. How does that work through actual usage. Now what happens when there's a scandal or a new advertising campaign or some related economic change -- if it's as expected, it's priced in. It's not valuable. It's only valuable when you don't know which way it's going or you need to know when there's a surprise, and you can only do that with live human data, by definition, because of everything else is the continuation of existing patterns. So that stuff is not going to -- in my view, it's not going to be attacked by -- and that's why I'm so confident about synthetic from our point of view because it's a way of getting extra value. It is hugely valuable. I wasn't knocking synthetic debt, as you understand, just for measuring change. You can't meter change with it, but you can get a lot more value out of the data you've produced. So that does not attack our core business. Indeed, it gets more value out of it. Having said that, I do imagine that some people who are used to having trackers might simply not do them any more or not do them as much as frequently because they think actually, there are some things that we can get elsewhere. So I don't pretend that there will be no effects. And that's why it's important for us to be the best that we can be, so that when you do keep tracking studies going. It's us that do the best job for you. And that's where I think the richness of our data and the smartness of our science is ahead of everybody else. But I can't say to you, but I can make confident predictions about how all data products will go. I think we're well defended. Our moat is our panel. Nobody can touch that. Nobody can equal it. The entire desire of the synthetic company is to squeeze out humans. They go and they buy mass polling and hope that they get this. They make their twins and they don't have us to talk to them again for another year and get lots of value out of it. That doesn't work. They have to have live data to make a great product. So I think we're in a strong position.

I wouldn't want everyone to model 7% or 10% growth in data projects next year on the top line. But I think the 2 or 3 encouraging things to add to Stephan is, even though agencies, the overall income we got from agencies this year is lower than last year. but we are still winning some pretty big high-profile data products revenue from the agencies on the sales support Stephan's point, that product is still very truly alive in a vertical sector that's kind of moving against it in some ways. And also, what's encouraging is that there's some new technology clients, 1 of which we alluded to with H1, that's come in relatively materially in H2 as well. So technology clients that often are more custom trackers and more research based, they're now seeing the benefits of data products. So my focus on data products personally is very much the margin side. and to ensure that we still got the new clients coming in, but the margin side is something that the value delivery plan is sitting pretty hard on data products because it's syndicated it's got amazing operating leverage as we food in H2. So that's as important as any top line growth from a value delivery perspective.

Sean Kealy from Panmure Liberum. I've got two for you, James, and then hopefully, 3 for Stephan, if that's all right. So apologies for possibly setting a record for a number of questions. First of all, James, you're very clear at the half year that you thought MI and sort of infrastructure inherited was good enough to run the business. I just wanted to get your updated view on what you think now of quality financial reporting and infrastructure that you're working with, and particularly noting you've had a bit of a restatement on tax this year and so on.

I'm actually reporting no issues at all. We just had -- I did my first audit with and it was a very clean audit very relatively easy audit. The tax issue is I'd say, a minor frustration. Happy to talk to you one-on-one about what the detail is in Phase with the shopper fair value versus cash value when we purchase it, there's an accounting change, which wasn't picked up at the time. we paid for the accounting change, we did not give us a credit for the extra payments. So that was a minor sort of acquisition-based change, which is annoying it's not ideal, but that does not reflect on anything to do with our systems and our MI. And I'm feeling that each week I'm now, I'm getting more comfortable with the commercial MI that actually helps you to make decisions. And the whole thrust of the value delivery plan, there is 1 of the 12 work streams is on systems more generally and focusing on margin more generally and time sheets more generally. So all this sort of wiring and sort of a good discipline is a key theme, and that's all happening. And for example, now previously time sheets were filled in by some cohorts are seen because not anybody so you could get misplaced margin on some projects where now 100% time sheet completion, full stop. Are they overviewed and everything on copy not quite, but we're already seeing the benefits. So to answer your question, systems from a commercial perspective, give me what we need, and I wouldn't read anything into the buy you adjustment that's an annoyance as opposed to a systems issue.

Perfect. Very clear. Second question, given the ongoing route program, are you able to give us a gentle guide at this point for cash restructuring charge for the coming fiscal year?

I don't want to give an actual number, but the areas we are investing in is in data scientists and engineers. And we're only going to be investing in that area if they're going to deliver a particular feature or product -- and by definition, that means it will justify going on the balance sheet. So there will be a more structured, more tech company focused approach to our capitalization of that spend. But I think it's too early for me to give you an indication of a number happy to be asked that question at the half year when I'll have more clarity.

And then Stephan, if I can rattle through three. First of all, in terms of Paralax, can you just run me through -- I appreciate you said in the presentation that it's token-based pricing. But where are you seeing prices that product sort of I appreciate it's really early stage, but maybe it's more a question of where are you targeting pricing for that product relative to other services that you got provides. Secondly, with respect to the MCP connection, I'm just really interested to discover, are you primarily seeing clients pull your data into their own environment? Are they pulling it into an environment credit by a third-party company. What I'm hoping to understand is sort of where clients are interacting with your data and so on. And then finally, on the AI Labs revenue, both the figures of 4 and 11 that you gave. Are you able to just sort of give you a run-through where exactly that sits in divisions and various halves of the year just gone?