Stephan Shakespeare: Good morning, everybody. Thank you for coming and thank you for coming at first -- my first one back. And I hope that it won't be too long before we see the rate of growth we've had before. We're at 388 -- GBP 389 million at the moment with a 16% margin, with 8% increase in reported -- in adjusted EPS, and -- sorry, in reported EPS. And the key thing here is that we're showing stable growth. Now stable growth sounds a bit of a contradiction, but actually, it represents the 2 things we're trying to do this year. One is to return to stability. And that means fixing things. And the other is to invest in growth. So we're not seeing yet the kind of leaps that we have had in the past, but we are investing for that very, very thing. And just to remind you of that story of YouGov's growth, there's a huge graph before this, which shows something like 10 years of growth, and we've just come off that. And we need to be reminded that this is fundamentally a growth company. And it's a growth company because we have always been led by innovation. And when we stop innovating, we go flat. But we are back on the road to innovation, something that I will show you in the second half of this presentation when I talk about the new methodology that we have produced. So in the last year, we've had some good successes, which is the stabilizing part that I've been talking about. We have continued rollout of ID verification on panelists. We spent quite a bit of effort focused on -- we're moving forward from panel. This is something that has been bedeviling the industry as a whole. I think we're a long way ahead of everybody else, partly because of our historical asset of a well-embedded panel and partly because we have been using the latest techniques, and I think we pretty much lead the pack on the reliability of our data. We've invested in our Cube powered products, especially on the data science side. A couple of days we'll be announcing a new addition to the team, a very important addition, someone who has led an important team for 10 years at Nielsen, really representing the seriousness with which we're taking the data science side and growing that to create the richness and the reliability of the entire Cube data. We've also established on the client services side a team that specializes in selling and educating clients about the value of our connected data, our data products. And we've continued our program of updating our dashboards, including putting AI into those to help create -- to help with discovery. And finally, for this section, we have done a pretty good job, I think, of integrating Shopper, it's a major job that was and it has been successful and Shopper is actually doing a little better than our expectations were. So that's a pleasant change. So with that, I hand over to Alex.
Alex McIntosh: Thank you, Stephan. I just want to do a quick overview of our lines of business. I just want to point you to the stack charts on the top of the screen. We've gone from GBP 335 million to GBP 388 million -- sorry, GBP 389 million revenue for the year. You'll see the biggest contributor to that. We've got a full year impact of Shopper coming through. On an underlying basis, you'll see the 2 divisions, core YouGov growing at 1%. I think I want to make a -- specifically point out data products. We've turned that from a decline in the last period to growth. It's a lot of investment and a lot of focus that's going into getting us back on track. It was a key driver of our performance for the previous reporting periods that Stephan referenced in terms of those double-digit growth years. And I think you'll see the beginnings of an evolution of things that we're doing in that space. I think we're pleased to see we've had renewal rates normalizing, back up to 82%. And the couple of wins in the media agency space, that may be a little bit of surprise, people saying we are seeing a little bit of weakness as well, but it does go to show when we have high-quality data, there's still a demand for that, and we had a significant win in the retail space as well. In our Research division, a bit of a mixed bag in performance. We've seen some headwinds coming from our government sector and our gaming sector. Gaming has been a long-term decline for us, but we saw some real strength and demand in our academic technology and financial services sectors. Shopper just referencing what Stephan said. On an underlying basis, I mean we don't have this in our numbers. But if we were to look at it on a trailing 12-month basis, it's growing by about 4%. So we're pleased with the way that has performed. I want to make the point is a period of coming off the TSAs under the ownership -- under the sale from NIQ. We're now off the majority of those, a lot of heavy lifting, getting control of the finance systems, et cetera. And so it's been a period of lots of, in a way, disruption moving systems, et cetera, but we're really pleased with the way the teams have continued to perform. You'll see our profit on the bottom of the chart has increased from GBP 49.6 million to GBP 60.7 million. A big driver of that is contribution of Shopper, obviously, but also the amount of cost that we took out at the beginning of the year -- referenced that at the beginning of the financial year, we announced that we made -- we had pressed the button on GBP 20 million of annualized savings. Because of timing, we realized about 70% of that in the year. Just moving to a geographic analysis, a bit of a mixed bag in terms of performance. Europe, you'll see year-on-year growth is 0%. Part of that has been some headwinds that we've had within Switzerland and Germany. We're starting to see some improvement coming into the second half on that. In the U.K., which has historically been a strong driver for us, a lot of disruption going through the redundancy programs. We started that on the 1st of August 2024, lasted about 3 months. And so it was inevitable that we would see a slowdown in performance there as we went through the consultation process. But we've ended the financial year really strong, good trajectory going into the financial year. Areas of growth for us has been Americas. It's always been our big focus. We'd like to see that growing at a much faster pace, but 3% on an underlying basis is broadly in line with how fast the market has grown. And just a small point, Asia Pac continues to grow by 2% -- this chart, which is looking at our sector. We take out Shopper in this because it's so skewed to FMCG and retail, but we continue to be very well diversified. Technology remains our largest segment, and that's a combination of technology clients using our data, but also using more traditional market research type services. Good contribution from banking and insurance. Travel and tourism has picked up again. Retail, I mentioned academic coming through in research. I want to make that point again about shopper. Just moving on to higher -- our cash conversion and our cash capital expenditure. For the year, we remain about the same cash conversion ratio as the previous period. We've had a bit of working capital outflow to do with -- we've had a bit of accrued income increasing. We've also seen panelists redeeming more points this year, and that's in part, we're running a lot of surveys, particularly in America off the back of the U.S. election. CapEx is down slightly. You'll see we spent a little bit less on panel development. That's not necessarily we haven't been getting more panelists. We've been a bit more efficient in how we -- in our conversion, and we'd like to see that improving over time. And we've kept our investment in technology expenditure roughly flat. That's not to say we haven't increased the amount of people in our technology teams. We have been spending a little bit more time on maintenance. And I think when we get to the latter parts of this presentation, you'll see some of the things that they have been working on, which will drive some more performance into FY '26 and beyond. We end the year in a robust balance sheet position. We started the year with a EUR 240 million loan facility. We paid EUR 36 million of that down in the year. We have a EUR 40 million RCF, which of currently EUR 24 million is drawn. We made an adjustment to our amortization schedule in terms of payments. And we had a -- we negotiated particularly aggressive for us. We wanted to delever as fast as possible when we took the loan. We're not trading at the same levels we were before. So we've reduced our payments to EUR 20 million for the next 2 payments, EUR 20 million in FY '26, EUR 20 million in FY '27, just that we have the headroom to continue investing in the group. And again, we really want to get ourselves back into this growth trajectory. But I want to make the point, we remain well within our loan covenants throughout the year. Just moving to current trading and outlook. You'll hear us talk about investing. It's particularly important for the group that we are taking on this market. We used to be the challenger brand, and we certainly see we have a right to win in a number of spaces. So we've got a clear set of execution priorities that we have around panel and product innovation. We've got some investments that we'll be really focused on data science and product development people, which are moving us toward our SP3 strategy of being more of a platform business in the way that we go to market, the way that panelists and our clients consume data. And we're starting to invest in Shopper. And the idea around Shopper is to get their capability expanded in the markets they're currently in, filling out more of the European map. And over time, we'll be looking at how do we invest into getting Shopper into the U.S. For our trading currently has started in line with expectations. I think for FY '26, we expect to see modest improvement in revenue and margin, and that's after making some key investments, particularly in data scientists and technologists. But I think we start the year, I think there's a lot of compelling opportunities for us. I know it's a slightly challenging macro environment, but we're certainly seeing some good opportunities from clients coming into the financial year. And with that, I'd like to hand back to Stephan.
Stephan Shakespeare: So yes, I mean, I said the YouGov story is growth through innovation. That was the promise that we made at Capital Markets Day in May 23. And the strategy that we put out there is the one that we are following. We are back on that road to growth, I believe, certainly on that strategic road. And that involves these 5 things: the renewed commitment to increasing visibility and quantity of public data. As you will see in a moment when I demonstrate the importance of public data to us, that is something that drives our reputation, our trustworthiness, our panels and a lot of other stuff. And it's something that we are highly committed to and we have increased our spending on. Innovation around panel recruitment and management of panel. We are changing the panelist experience. Again, you will find always this emphasis that we have on public data on panel, creating data that creates products that are good for clients. This is all part of a flow. And the way that we treat our panelists and the way that we get data from them is being enhanced. We're accelerating the execution of becoming a data platform. More and more of our dashboards are now containing AI and better ways of utilizing our data. Custom research is a huge part of what we do. The degree to which custom research is aligned with our platform is the degree to which our success strategy is working. Those two things being aligned is absolutely critical to us, and we've been putting energy into that. There are aspects of the custom research offer that are not so aligned. We need to bring everything in line. And AI is helping us to do that. And of course, something that we'll be focusing on in a moment, in fact, in the next slide is the innovation in AI that is massive for us leveraging the value of the assets that we've built. So that is the broad strategic view. You've heard it all before. There's nothing new in there other than that we are updating all of that with AI and we're coming at it with renewed enthusiasm. Now YouGov in the Age of AI is the big question that anybody would ask. And we, I believe, and I hope I'm going to show our company that is ideally suited to use this moment, this historic revolutionary moment for our growth because we are all about talking to real people. And the essence of the use of AI is real people. It is building extra value out of the real people in order to get even better data products, even better value to clients. And that is something I'm going to come to several times because we think that our industry has maybe gone a bit wrong in some areas. There are so many wonderful things that AI can do. Replacing humans isn't really the job of a market research company. There's lots of great things that synthetic data can do to get more value to make it easier to do things like ad testing, and there are a lot of places where that really works well. But remember, the vast majority of spend from our clients is in measuring change. That's what people are interested in. And change cannot be extrapolated. Change -- extrapolation is the assumption that things will be the same. Every time you use synthetic data, the underlying assumption must be that things are the same way you're extending into. That is the definition of synthetic data, it's extrapolation. And extrapolation tells you what you already knew. It extends it, but it doesn't tell you where the surprise is coming. And that's why people buy tracking data because they want to know what's changing, that they don't already know. If everything is going the same next month as last month, then it's all nice, but that data isn't very interesting. It's when the change is not what you expected, and that's not going to come from synthetic data. So we are actually very interested in synthetic data. As you know, MRP has been a very big part of our success in accuracy and getting more value out of our data. So we are actually pioneers of synthetic data. And it has fantastic value. There's nothing against that. But the vast majority of the market research spend is in tracking change. And change, you need real people. And real people are the basis of everything we do. So a data company focuses on the flow of data across 4 things: people, data, process and output. And for this to work, we have developed over the course of 25 years 3 major assets. First of all, YouGov has the best panel. We don't have the best panel in every single country, I wouldn't want to pretend that. But in our major markets where we have our strong panels, we have the best panels. Everybody knows we have the highest contact rates. We have the highest levels of representativity. We have the engagement that keeps them there for a long time so that you can build layers and layers of data. It's by building layers of data from engaged panels that you get the second big asset, which is that we have the best data. And it's the best data because it's a single source, it is connected, and it is always recent. It's always being updated. Every single day, it's updated. Recency, representativity and genuineness, which should be a given, but isn't these days, are the things that make great data. To have that, you need good panel that gives you the best data. And I'm putting there that it's also across different areas of demographic data, things about the people, what they're thinking, attitudinal, behavioral, passive data, that's part of that. Qualitative is the bit that's going to be a big new piece, which I'm talking about in a moment, added to quantitative. The third area is our incredibly strong brand. And it's always good to have a strong brand. But for us, it is key. It is a key function of what we do. Because strong brand yields better panels and builds trust amongst clients. And I have 2 examples here that I want to just run. [Presentation]
Stephan Shakespeare: The importance of that for us is Trump says that assuming you know what YouGov is. And if you don't know what YouGov, the name still implies it's an authority. That phrase "according to YouGov" is incredibly prevalent in the media. It is what we want -- I used to say Google is to Google. According to YouGov, is our talisman as it were for this. And Meltwater tells us that we are the most quoted company in the world's press. There are over 1,000 mentions about us -- of us, of our brand every single day. The total number of -- is in the small 385,000 mentions in the last year for the YouGov brand. And you can imagine, this is a value in itself. We're also ranked #2, and this comes from the next -- this one and the last one are coming from independent research. We're ranked #2 for aided brand awareness globally among research buyers. And amongst those, switching to the last one, we are the most trusted market research provider. So even when we're not the most famous, we're just #2, we are the most trusted. And anybody I think in the industry would say, who is the most likely to get a result, like it's YouGov. And these are really fantastically important assets. A strong brand is strong reach. What isn't in here is -- sorry, I skipped it, is we have 4,000 active clients. Now all of those active clients, obviously, people that we can talk to and people we can show our new product to. And you could say this slide is a massive strength. And it's also, in some ways, an indicator of we've got a hell of a lot more assets than we've managed to convert to value. And so we know what to do. This is a massive asset. This is stuff that you can't -- you can't create this quickly. This trust, this reach, this visibility. And all of it will feed into the new products or the new methodology that I'm going to show to you. So what changes about all these assets in the age of AI? And it is a revolution that's happening. But for us, it's very much an evolution because everything that AI allows us to do is an enhancement of assets we've already built. Our mantra from Andrew Ng, who was the founder -- co-founder of Google Brain, really fantastic quote for us, "It's not who has the best algorithm that wins, it's who has the most data." And the other people say, oh, most data is, what's the best data or the best insight, all of that. All of that stuff just emerges from the quantity of data. Genuine human data at very high scale creates good data, it creates good insights, all flows from that. There's no shortcut to the value of really large-scale data. And that is what the whole world is turning on. While other people are trying to cut out the real sources of this, trying to say, hey, we can make more money by not bothering all these humans. We are saying no, it's all about the humans. It's all about the number of people talking to you, how much do they talk to you, how much do they give you? And that's what AI lets us do. And AI enables us to do data collection and discovery at scale. Data collection actually, you didn't need AI for until now, but we are talking today about qual data. And qual data is a different type of data. And it's a type of data that we haven't done much with. It's a type of data that is the Cinderella, if you like, of the industry. People do this as a good way of getting insights, of brainstorming and so on, but you can't base big decisions on qual data because it's touchy-feely stuff, right? It's not stuff that you can create a measure out of. Well, that changes. That changes when you have AI to, first of all, use the background data to choose the right people to talk to and to know what to say to them. And then to take all of this unstructured data that's produced by interviews held by AI and turn those into data that the clients can actually use. It's not enough to be interesting, it's not enough to be good for a brainstorm. It has to be things that you can use and base decisions on. And we are now doing thousands of interviews driven by AI on a daily basis. I'm showing you one example of this. Now I'd love to really -- not doing any demos here because you can't really demo this stuff. So I'm going to show you this one thing, which is a snatch of a conversation. And I'll just -- you probably can't read it, so I'll read it out to you. It says, I've noticed you've given top ratings to quite a few music artists recently, everyone from Rick Astley and Hall & Oates, to 50 Cent and Pussycat Dolls. They each got 5 stars out of 5 from you. What shaped your views on these artists given they span such different musical styles? Now this is a question that the bot came up with that was not based on a prompt of us asking them about anything in particular. They have the background data, the panelist, they were definitely told to talk about music, but they got the bot. The bot got the -- found something interesting in the data to turn into a question. And the answer is, I grew up listening to and appreciating music from different genres and eras. Bot comes back and says, what first got you into such a wide range of music? My mom, school friends, going to gigs and music channels. Which one of those -- which one do you think have the biggest impact on shaping your taste in music? Music videos in the '90s and the '00s. It goes on and it can go on as long as you like. But you've taken previous data, turned that into a relevant question that's targeted at this person. They know that you're listening to them. They know that you know something about them. And that's why they're here, by the way. It's not creepy when we do it. It's creepy when Google or Facebook or whatever does it because you didn't ask them for you, you didn't come there for that. You come to YouGov to be listened to. So this is listening to. And it's responding to and it's getting you in a conversation and it's coming up with an insight. And that can be used, built on in lots of different ways. Now we're not doing one of those. We're doing literally thousands of them, thousands that would cost you -- you could never imagine the cost of just this 20,000 conversation study that we're doing previously. And it is a very low cost. I'm not going to tell you what the costs are now because we'll have a Capital Markets Day before too long and we'll go through all the things and our expectations and things. I'm just showing you a new methodology. This is huge scale at low cost. It's automated, customizable, configurable, continuous data collection. Only YouGov can do this. Nobody else has the combination of things that this requires. This requires a large connected data. Imagine that bot going into the 2,000 or 3,000 things we know about a typical panelist and being able to use that and find the interesting things there to maybe open up a discussion, or to look for the particular thing that the client wants. Maybe the client is only interested in their supermarket habits. So the bot goes into there and finds anything you can find about supermarkets and takes that at a starting point. Only YouGov can do that, because nobody else has the range of connected data with live panelists now. Nobody else. And then only YouGov can do the scale of continuous questioning, so not asking 100 or 1,000. We can do 20,000, we can do 100,000 interviews a day. And we can do them at this scale because we have highly engaged panelists and they come back. They're not -- of course, there's a lot of churn, but our stable panel is with us over time, and we can build up a relationship and we can build up all of that data. So this is our right to win. It is the assets built over 25 years, the best panel connected at scale, the strongest brand and now adding the AI. All of that comes into something that is unique to us. This is a slide that attempts to encapsulate just in one picture what we're talking about and really what we're adding. So over here, we have, on the left-hand side, we have the world of things, the entities. YouGov, as you know, covers over 20,000 brands and products in our tracking. When you combine brand index and ratings, it's more than 20,000. But we say that because it's a changing number and ever growing. But it's musicians, it's TV shows, it's media products. It's supermarkets, it's brands, it's consumer goods, everything that you can think of that is in that commercial world that you might want to track is in our database, is in our Cube and it's all being processed through all of these people's heads. That's what's happening here. They're living their lives. And they come into YouGov and they ask questions and they become obviously noughts and zeroes, and that creates a line. And that's brand index. And brand index was the first, is still the only reliable daily measure of brand strength. And it goes up and down. And you really need to know that. You need to know a lot of companies put this stuff into their risk -- I mean, for example, Bank of America, it's embedded in their risk modeling. It's part of their understanding how news flow affects their accounts and new accounts opened and accounts money taken out and so on, is predicted by reaction to news flow measured by YouGov. But what this doesn't do is it doesn't tell you why something has gone down here. You might know why there might have been some incident, and you know why already, in which case, you might want to know, okay, how does it bother people? Who does it bother? And why? Or maybe you have no idea. There's a trend line, you say, I don't know why it's going down. What this new data does, of course, as you've guessed already, is it gives you the why, not the what. I can't remember if I mixed that what and why. But this is the what's happening. This is the why it's happening. It's in here. The way people are talking about your brand, you can be very specific and say, have you heard anything about Tesco lately, to try and pump that, or you can just say, what do you think about Tesco, or which is your favorite supermarket? You can decide how you want the prompt to go and generate these conversations, and then you can find out how are people talking about you. Then you can do several things. You can compare and contrast things within the data and -- within this data, but you will also take all the previous data that you've had because, if you're a Brand Index customer, we'll have a bank of sort of background hum data as to know what the normal conversation is like, and you can compare the normal conversation about you to the conversation happening today and find out what is it, what's driving this change? There's going to be a huge amount of value in this that we have yet to discover. This is like a whole new treasure chest. But just it's not abstract. After this, I can't demonstrate it to you here, but after this presentation, this goes up online, at the end, you'll find 2 links. One is to about 20 transcripts and the other is to the functional output. Now it is an output for one study. So all the buttons don't work the way -- I mean, it's showing you how it would work, but it's specifically around one study. But you can play with it and see. Because if you have all of this vast data coming along and you don't have a way that it turns into something usable, then that's interesting but no good. So obviously, there's a lot going on there. And really, it's very -- it's delivering real value to clients. It gives the why and the what else to the what that we've already done. It's automated, customizable, targetable and actionable. That is to say, it is really a custom thing as well as a product thing because you can turn it onto anything. You can have a single study from it or you can have it on all the time. Its scale reveals the long tail of new information. It isn't just the thought things that you -- and this is kind of under the next thing, it isn't just the thing that you thought you wanted to know, that's the known unknowns. That's what a survey is. You know what you were trying to find out and you write the questions for that. This is the unknown unknowns. The long tail. The stuff is -- what are people talking about? How are they talking about these things? Does somebody, maybe one person in that conversation come up with something anomalous? The AI will surface that and you can find out things you didn't know. And the last bit, all of those first 4 things, of course, happening already. This isn't just a plan. This is actual delivery. But the last part, the alerts, we have not got to and that's something we're productizing. So the idea is as this flows along and you're getting actually just open-ended questions to sort of pick up the continuous hum of chatter -- by the way, not the same as social listening on social media because the whole point about our panels is they're highly representative. They represent all the subgroups of a population and they represent them fairly. So when you get this hum, you find out what people are actually talking about, not what's on Twitter or whatever it is. That's not -- those are not the same things. So you get -- you look for the -- across this entire horizon and you will get alerts to say, "Hey, here's something you might look at, something that was unexpected." This significantly enhances the values to our data products, I should have actually said, to all of our outputs because you can do this to a single survey, if you want. This is not simply a new product, although it will exist as a product, you'll be able to do just a study of this based on conversations. But it enhances every single data output we have. Everything that we do that was a what becomes a what and why. And that's why this for us is a major revolution. It is going to, I think, have as much importance to us, the qual side as the quant side. And that's what's new. So we've got these 3 dimensions: the sheer quantity of data, the recency, the daily collection, and we've got this massive range. Nobody else has the range. Nobody else has the quantity. Nobody else does the daily collection. And you can say, well, you add up all these assets and you add up the stuff that I talked about before about our reach and our trustworthiness and so on, and then you may ask yourself, why the hell do you only make GBP 388 million? And I do think there is a massive gap, and that's something that we really have to address, how we do better at teaching people the value of our data. So that remains something that we are -- that we have ideas about but that we are working on. The last thing about this slide, this data is an ideal for processing and analysis by AI. I've said that AI, I think, is great for some things and not so great for others. This is right in its area of strength, taking large amounts of unstructured data and turning it into something meaningful is what it does like magic. It's like when we first saw ChatGPT talking. You can't really work out how it is. It isn't really the algorithm. It is the sheer quantity of connected data. Right. Well, we already have our first paying client. I have to say it's a tiny, tiny alpha version of this, but very good. So we already have engineered into the system into Brand Index, that if you want a daily collection or an occasional collection of open-ended data, you can trigger that, and you get that every day. And the version we have now is simply one question, why did you say that? So you've given Tesco or whoever it is a good rating or a bad rating, and it comes up and says, "Why did you give us that rating?" And that adds up to a really useful continuous a little bit of insight that we add to the Brand Index subscription. And that's just had its first subscriber. It's only been out -- we've only been talking for a couple of weeks. We have a lot of clients lined up for further discussions there. But what we're obviously really talking to them about is that question becomes a conversation. And that will be engineered in. It will be ready by Christmas. Just a last couple of points. We've talked about one very important use of AI, but we're using it across everything we're doing. It's helping us with fraud detection. And I think that we will be the -- we are the leaders in genuineness of data. It allows us to do new types of data collection at scale. As we've seen, it does data analysis for us. It does discovery and interactivity on our dashboards. And finally, it also is being used, we're working with a couple of LLMs to turn our data into usable things in search and so on so that the public side of our data is inserted in its best possible form inside the infrastructure of search, because we are a trusted source of data. We want to maximize the value of that data. Everything that we do, everything that we do for ourselves, our proprietary data, is available for free in top line form. That doesn't, in any way, hurt our products, I believe, because you always want the detail. No marketer just wants the top line. But the public is interested in the top line, like the President. It's valuable, it's used almost always in its top line form. And the more that's available, the more it teaches about the data that we have. So all of these things add up to YouGov becoming an AI-driven data company built on real people for all society. As I said, at the end, when you look at the end of this presentation later on, you'll find 2 links, one is transcripts, one is to the interface. It is obviously in a curtailed form. There's also a video to watch that's being added. And we are ready for your questions.
William Larwood: Will Larwood from Berenberg. Firstly, just if you could provide some color on sort of the visibility for the top line in FY '26. Obviously, we've got the key renewal period for data products in November and December. That would be great if you could share a little bit more detail on that. And then secondly, in regards to sort of pricing more generally, how you're thinking about that in FY '26 and potentially beyond? And then finally, just -- is -- do you feel there's anything further that you need to do in terms of, from a commercial point of view, there's obviously been some change over the last, say, 18 months or so, particularly on sort of both the CPS side and the data product side?
Alex McIntosh: I'll take first 2. I'll start on visibility. So I want to point to a couple of things. I'll pick on the U.K. We've got Will, who's the U.K. CEO here. We ended the year quite strong, in particular, building momentum into the second half in some of the markets that, in the first half, it underperformed. And so we go into the year and we have talked in the past about our backlog, the committed revenue that we have coming into the year. We came in 3% higher than we were last year. We're just a shade under 45% this year. We were just -- we were up 41% last year. So we're seeing that backlog increasing. So we've got fairly confidence on a sort of good performance in the first half. We're not talking -- I'll make this point again, expect modest growth. We're really looking at how do we continue doing a lot of work that's happening under the hood. But I think we're moderately pleased with that. Obviously, it shows some strength coming into the year, particularly with the macro environment. And so I think looking forward to the renewal season. For us, that's typically clients take a data product renewal from the 1st of January, and so November and December for us are key months to make sure that we're getting on top of that. We've amended the way the team structure works. We have a dedicated data product team back to the old model of a team that's really incentivized and focused on those renewals, getting those renewal discussions early. I think having some interesting things to talk about, new developments, in particular around this capability that Stephan's pointed to, shows we should be garnering more interest in that. So we're quietly confident on that. A little bit on pricing. I mean FY '25, we didn't touch it, there was a lot of change going on in the teams, and in particular, getting ourselves set up for changing some of the incentive structures in the 1st of August, which it's very hard to change the incentive structures midyear. We are now putting through. It's a relatively small thing, but we are pushing through inflationary price increases. That's something that we hadn't pushed in the last sort of 18 months. That's when we just started. And again, coming back to our peak renewal season, we should see some of the benefit coming from that. I'll pass to you on for the...
Stephan Shakespeare: Just one thing I wanted to say about pricing, because you may have some more to say. I just wanted to say that there's something in the -- in one of those slides that I could have expanded on, and I thought we've already spent a fair amount of time on it, that one of the outputs of a data company is a Data Lake with APIs or an API. And we haven't done a lot of that. We have, in fact, got a number of clients who just take a feed of the entire Cube. But if we're a platform company, we won't be always just thinking of selling this product and this project and you have to come in at this high level or you don't get anything. In fact, it should be the opposite. You should be able to buy exactly the bit you want in any slice or any form that you want. If I just want one question and one -- that's always been impossible on Omnibus, but it should be possible for all of our data. And I think this is quite a big project. This is not something we can deliver. The API and the Data Lake bit is available now, and it's -- but it always involves some extra work on something. But a real front end to that, that says, I want just this particular data should be the way that we allow clients in. And part of maybe how -- when I said there's a big gap between all of our assets and what they're buying is make it easy for people to buy any bit that they like. There's no reason for us to say you have to have a very big subscription to Brand Index. You can ease your way into that. And when we've sometimes tried to do little data slices, it's been very successful. It's just we haven't wanted to do that. And that is a bigger project. That isn't an overnight thing. But that is definitely what a platform company would do to sell data at lots of different levels.
Alex McIntosh: Further changes to the commercial team?
Stephan Shakespeare: Yes. I mean we have put a very large -- not bounty. We've put incentives in place to make data products get more prominence, and hurdles that you have to hit before you make money on selling custom to sell products. And we have a dedicated team that does nothing but products. It's a small team, but it will grow. And this is really the change in our commercial -- in our sales approach. And there haven't been -- it hasn't been a massive -- it hasn't been as some -- somebody said we're having an overhaul or whatever. It's not an overhaul. It is an evolutionary change to our system. It's -- we've done well. We wanted to be better and we've made, as I say, some significant changes, including putting product as the #1 thing we're trying to sell and making it impossible not to sell a product if you want to get -- sell custom as well. The 2 things are so aligned that it is a matter of how you incentivize, when you're selling one, you can sell the other. But it has to be that you have to sell subscriptions first. Otherwise, you're not going to be a data company.
Lara Simpson: It's Lara Simpson from JPMorgan. My first question was just to come back to the P&L. So you did GBP 61 million operating profit, which was really in line with expectations, but you clearly have benefited from sort of lower central costs and then some delayed spending in Shopper. Can you just talk a bit around the margin pressure you saw in Data Products and Research? Clearly, profitability was a bit weaker there. So where are you investing? Or is it sort of slow realization on the cost optimization side? And then you've obviously outlined increased investments into technology and data science. Can you just outline, sort of quantify those investments? And then maybe just give us some line of sight on exactly where they'll be going?
Alex McIntosh: Yes. On DP, it's a very simple answer. We acquired Yabble at the beginning of the year. It was a loss-making entity when we bought it. This is about just a shade under GBP 3 million. That's been completely allocated to the Data Products division. So yes, the margin pressure is purely as we're ramping up the activity, wrapping up the integration, ramping up the Data Products, the capability behind this is in part driven by Yabble's applications. So we expect to obviously see some of the revenue growth coming from that to help absorb some of that cost that's going into the business. In terms of the investments, we're budgeting around GBP 4 million. There's a question mark on how fast we can bring that, about having head count. And so Stephan pointed to we have a new hire coming in as our head data scientist. And so it will be primarily focused on platform technology, which will support product, depending on the types of activities because Stephan is correct, this could be applicable to customers as well. So once we make a bit of progress, I'm just going to repeat what Stephan said, we'll come up with the Capital Markets Day to really flesh out what that looks like in terms of where do we see the growth rates coming and where do we see that landing. But for now, it will probably be even spread between the 2 because we're going to see some applications that are applicable to both of the lines of business.
Lara Simpson: And then just another question for me was around the balance sheet. So you've obviously closed 1.7 net debt EBITDA, maybe slightly higher than what I think some were expecting. You've obviously pushed back some of the payment terms. It feels like there is more sense of urgency to deleverage post CPS. Obviously, now you're investing a bit. Can you just talk about sort of balance sheet expectations over the next 12 to 24 months and how we should think about that new deleveraging as a priority going forward?
Alex McIntosh: Sorry, Lara, to kick off. It's still a priority. I mean for us, it's really -- we have to be -- not careful. Careful is the wrong word. We do need to make sure that we have capacity to invest. We do see some clear opportunities for us. As we start to go out of the market talking about some of this capability, if we can see some revenue potential there, then, of course, you'll see us being much more aggressive in terms of being able to go for a market. In terms of deleveraging, we're taking that down to EUR 20 million for the next 2 years. So we expect that to come down, albeit at a slower pace, but we do expect to have delevering happening. And of course the other side of that, we're trying to significantly increase our profits. So we're trying to achieve both, where we'd like to see some significant movement over the next 2 years in that deleveraging and, at the same time, making sure we're getting into that growth trajectory.
Jessica Pok: Jessica Pok from Peel Hunt. I've got 3, please. The first is, can you comment a little bit on the custom -- the sentiment for custom research amongst your client base? I mean Data Products slowed down, but also custom research. And the second is on Shopper -- the Shopper segment and the investment going in. What is the key focus for Shopper over the next 12 months? I mean you've talked about broadening geographies and you've talked about product, but which is the main focus? And the final one is just on the new innovations that you've showcased. Does that change the way that panelists are monetized -- are paid by going into this form of interaction?
Stephan Shakespeare: I'll start with the last one because I remember it, the -- and the second one, I remember too. It fundamentally changes our relationship with panelists and we're changing the structure of panel. We've already talked in the past about having a core panel that can do a lot more, called YouGov Plus. And we have -- we know that, by the way, people who are doing a lot more are not giving us worse data, they're giving us better data. It isn't like there's a professional survey taker that somehow gives you worse data. They give you better data that's more aligned with the reality, in fact. And so there's a core panel that we would talk to more and that we'll -- we can rely on more. And we are also now recruiting people not on the basis of any cash reward whatsoever, only on the basis of participation. This is a good way of actually making sure they're not frauds in the first place and as they come through the system. But more importantly, lots of people want to take part just for the sake of participation. If you give them large, boring surveys, that's not going to help you very much. But if you give them these conversations, they will -- we know that they enjoy them. And I mean not everybody wants to talk forever, but lots of people do. And so we have not only interesting surveys that are contributing to public data, but we can have these conversations. And actually, they will also do market research surveys and they would do -- and in any case, a lot of the things they notice and talk about is a form of unprompted market research. So these are sort of 2 ends and there's things in the middle, which is like our regular panel, which we don't interfere with because it's worked so well. So we're doing lots of things in panel and changing the relationship at different ends of that range. The second bit was Shopper. And there are, yes, 2 things: more countries and changing the product. So we've invested in the receipt stuff, which is a form of -- well, it's not entirely automated, but it's less onerous than scanning your shopping. Always remember that the old style here of actually scanning or shopping gives you a level of detail that no other methodology does. So that's why that -- even that old-style methodology of Shopper is incredibly valuable and retains its clients, and grows its clients actually because it goes down to the SKU level. But also we're doing passive data collecting and we are looking at other forms of doing that. And that is also something that will drive our entry into America with this behavioral data. I think that's the aim, as we add more types of behavioral data in there. So there's a mix of ways we're looking forward. I don't remember the first question.
Alex McIntosh: Yes, just a change in appetite. Yes, we're seeing a little bit of a mixed bag. I think in some of our clients, we're seeing and having seen some good wins in Data Products in the financial year coming into this year, we are seeing some pressure from media agencies. And so we should anticipate that's going to be a bit of a struggle for us. There is an element of doing a fair amount of custom research for that sector. But on the flip side of that, we're starting to see more opportunity to pitch for larger things as well in the U.S. So I think it's, to one degree, it depends on what country you're in. It also depends on what sector you're in, it's a pretty obvious statement. But I think we should still see some progress within the custom team coming into FY '26 besides macro. And to come back to some of the points that Stephan is making, it is around the measurement. It's people looking for more tracking opportunities. We like that. There's lot of visibility in it. And I think the U.S. team has been working pretty hard to uncover some significant opportunities here. There's a couple that we're working on in the U.K. as well. It's difficult to -- when you're in the summer months, not very much happens from a client decision making. So I think when we come into our Q2, we'll start to see some of that, potentially unlocking, we'll see some decisions made from clients.
Stephan Shakespeare: And I think that there's a change happening as well in expectations of clients. So they're expecting something new from AI. And they've been holding back, I think, because they're saying, well, what's this amazing stuff going to deliver? And so far, it's delivered essentially toys. The things that you get, people are not paying for those things. They think they should be there because it talks back at you and stuff like that. But it doesn't give you data that you're going to make decisions on, not for the majority of the market. So I think, obviously, I would say this, that our use of AI goes to the heart of what they are looking for. And so I think this is what they've been waiting for. I think they've been waiting for something that is new and yet that they can rely on them, that tells them something they really need to run their businesses. And that hasn't happened from AI yet. And I believe this is the start of that.
Hai Huynh: It's Hai from UBS. I have a couple on Data Products and then one bigger picture, please. So on Data Products, you haven't mentioned category view this time. I know you mentioned that you want that to be the way going forward. But is there a bit more of a tangible time line on when you're expecting it to add into the 95% customers you haven't monetized from? My second question on Data Products is just a bit deeper on the margin perspective. So Yabble brought the margins down. But without Yabble from the numbers, it will be 35% margins, right? So what were the drivers in there? Was that the cost savings? And is that going to be continued? Where do you see the margins going forward essentially with Yabble, and deep into that, when do you expect Yabble to break even? And then the third question, a bigger picture, is you mentioned the LLM potential monetization. How big of an opportunity do you see that is? And how aggressive are you pursuing it given the data quality that you have. Do you think monetization opportunities are there?
Stephan Shakespeare: I'll do the 2 outside ones. I think the monetization of our data is potentially high but it may be 0. I mean I can't give you a better answer than that because should they want it, of course, they should, because the thing that those models need is recency and trustworthy sources, and that's what we do. But are we of sufficient scale for that? I don't know. So we are going to scale it up, but I couldn't possibly make a -- say something about that. On the first question, category view. It didn't go well. We launched it, and the feedback was, well, we like what you have, but you're missing things that we need. And I'm afraid it was dropped at that point. There was no going back to fixing those things, which are highly fixable, and we are doing that. That is now with Joe Razza, our Head of Product, who is working on that as one of the things he's doing. So it very much ought to work. And it's actually, we have a good way of bringing it back, which is -- well, really, I'm not supposed to talk about it, but I mean, we want to apply it to a new category that doesn't have very good tracking. And that category is AI. And so we'll be seeing for long a variation of category view for that sector. Yes. And on the side, I would say that AI companies are definitely buying from us now. It's -- I have Investor Relations that I shouldn't mention that it's our fastest-growing sector because it goes from so small to something quite large, but we have made our first 7-figure sale to one of the LLMs. And we think that there is a need for our data by them.
Alex McIntosh: I'll pick up on the Data Products margin. So a couple of things moving the margin around. Yabble is one of those and another is the cost reduction program. We've also had changes in the level of capitalization that we have and a lot of our developers are focused towards the Data Products. I think coming back to when do we expect Yabble to break even, a lot of that depends on the pace that we can get these particular products out. I do want to make the point -- repeat the point that Stephan has made, you've got a difference in the way that clients are approaching AI products. And some people are finding they don't want to pay for them, they've seen this as a hygiene factor of having summarization, et cetera, built into your tools. We're really looking for ways that we can monetize that. And again, we'll update more concretely when we come up with the Capital Markets Day. But we do think we should be able to get that being a positive contributor fairly quickly. Where do we see margins growing? I think there's a couple of things we'd like to see evolving. So one of those is just referencing back to what Stephan said, around data slices and having clients be able to come and self-service their own delivery of data. Obviously, that would be delivered at a high margin. You're just taking -- it's a repackaging of existing data. But we're also focused on data partnerships. It's evolving the way that we -- the market that we point to, the set of users that we point to. Primarily, we are still talking to market research buyers. I think, clearly, you can see there are opportunities for us to go beyond market research buyers, particularly the LLMs. A lot of people are doing data deals with LLMs. And we already have a relatively small, we call it data activation, but it's data that goes into marketing campaigns, as part of Shopper's investment area, we've also been -- we make a few million pounds in the core YouGov business around that. We can see that also accelerating. The more clients are using AI tools for their own campaigns, it's a clear space for us to be putting data into that. So as the use of that data evolves and the sophistication of clients using data for their own AI models, yes, we see that certainly moving up into their [ forte ] and beyond. But the pace of that is still to be determined.
Johnathan Barrett: It's Johnathan Barrett from Panmure's. I guess I've got 3 questions. Just first of all, thanks for the interesting presentation on the AI interviews. I wondered if you could just walk us through the model for that, the commercial side of that. So what sort of volume of interviews do you need for this to be useful? What's the cost of that? And then how do you commercialize it? Is it a case of bundling with other products that you -- where you're already selling? Is that an uplift? Is it just a question of clients expecting more value for money and you end up with the same pricing? Obviously, at the moment, you're saying you're getting inflationary pricing increases through, but that sort of implies that volumes are flat. So is that -- what's the driver there? Does this drive growth in actual customer numbers? Or does it simply enhance your -- the value of the sale to existing? If you could just walk us through that. And then second question, and you've said a few things around this, so I'm just going to sort of try to round it up a bit, I guess, about data activation being used with clients. And just a more general issue of predictive work that you can do. Obviously, you've talked a lot about historic data, the what, the why, flagging what's going on. Can you move in that direction? Can you build your own personas for those purposes? Are you getting any commercial interest from clients? Just if you could wander into that side of the equation as well. And then thirdly, a very simple question, I think. Yes, obviously, we're wandering into this AI period. You're back in the hot seat, Stephan. Are you really just the CEO for this AI period? In other words, the company needs someone experienced like you who's been in the business for a long time to see through this and you don't want to take risks? Sort of a difficult question. I think it's just an open question. So I think we're all keen to understand that.
Stephan Shakespeare: Well, on the first one, the business model is, in some ways, it's early -- maybe too early to have been talking about it because there is so much work to do as to how far does this go, what kind of other interviews can you do, how can you use the prompt. We are at an early stage of that. The reason it's legitimate to talk about it today is because we've sold and we will sell a bunch of subscription add-ons from the alpha version. And the methodology is one that is really to be used tomorrow. I mean we have lined up in the U.K., I'm looking at Will, 5 or 10, I don't know how many clients -- 4 clients. Always inflate. We have lined up 4 clients that are -- will be very excited to want to use this. And so it is active and it will be able to answer their questions now. And if you came along afterwards and wanted to -- representing brand or whatever, wanted something, we would do it. We could run it today. It is not a methodology that requires engineering other than what we've already got. It builds on every asset we have. It's putting together things we've already been doing. We've run 20,000 interviews. They are very low cost. It depends on whether you're paying the respondent or not paying the respondent. But if you're not paying the respondent, you can imagine that one of these sections is going to be less than $0.20, right, for -- just I'm talking about that bit of the cost. And this engineered part isn't a high cost thing. Brand Index isn't a high cost to collect. So I can only give you very broad indicators of the numbers involved. I don't know if it's 20,000 interviews per day or 5,000 will do, or it depends how many countries we're in and so on. So it's legitimate for us to talk about it because we are selling it now in some version and more versions over the next weeks. But it isn't in a place where I can give you a business model for. That's what we intend to do and that's what we've just done. And we're looking to see what does that yield. Does that -- is that something we could have got just as well with 5,000 or 3,000? Or did we need 80,000 or whatever? It's -- and it depends, I say, how many countries you do it in. And so we've done it in a way that allows us to come up with conclusions, initial conclusions, about all of those things. I mean the beauty of our system is experimentation is incredibly simple and incredibly interesting from the first -- from the get-go. The second part was predictive, yes. Well, so prediction is hard because prediction is extrapolation really, unless you know -- you can't know what's going to change something. So I think prediction is about extrapolation and tracking. And we've done with MRP. I mean the best thing -- the best measure of prediction is something real, and that takes -- that is totally visible and we are the best predictors of elections. There's no question. We do it in many, many countries. We have just had 3 MRPs in Australia, Germany and Canada, I believe -- may not be Canada. I think it might have been Spain, actually, which were bang on -- huh? Spain, sorry. And as have been -- many previous elections, bang on in market research terms would be within 5% or 10%. For elections, it's like 1% or 2%. And that is our average. So that's a prediction of sorts, but it's assuming that people -- what they say going to do soon. Long-term prediction. I don't see how you do it from what we do. So if there's a predictive model that somebody has, I think they would use us as opposed to us doing that stuff. As an aside, you may remember with the Trump election, one of a hedge known as the Trump Whale, made large amounts of money on betting on Trump, and praised the quality of the data that we had. That data was ours. That was a client of ours. He went on to eventually say. We would not have used it the way he used it, our data. So it's a good example -- we will supply the real data. And if somebody else is better at doing that, that's their job. It's not our job. Our job is to provide the best, most accurate, real description of things now. And that's -- we know how to do that. And my personal stay, I'm here to make sure we're back on track as the growth company we were in the sort of 9 or 10 years that we were growing at double digit year after year. Obviously, I'm not staying for that period, but I'm staying to the point where we feel, hey, we're back on track. Now that could be just the end of this year. I think our expectation was 1.5 years, something like that. I've just gone the half year. If it was 2.5 years, it's too long because I'm -- I should have made bigger success by then. So I can't say. I mean it could be 3 years, but it's more likely to be a year or so. But there's really no point in deciding that by the second. We've got at least 6 months to see what happens before we have to make -- start making a planning decision. I don't know if that's...
Johnathan Barrett: It's really just, are you there for handling the AI thing right now?
Stephan Shakespeare: Yes.
Johnathan Barrett: Everyone's got this headache on the horizon, or it's right there right now hitting you. It's depending on who you are and what industry you're in, but.
Stephan Shakespeare: Yes.
Steven Craig Liechti: Steve Liechti from DB Numis. Just a few. On Data Products, I guess, in the second half, in my head, we had 3 things that you needed to do, which is category view, the AI -- sorry, the user experience tools and AI tools, which I thought was Yabble. Just talk us through -- you kind of alluded to category view didn't happen. Can you just talk us through on the other 2? And then going into fiscal '26 now, we still got -- have we still got those 3 to get the benefit from plus the qual stuff that you're putting into and launching into Brand Index. Is that the way to think about it? Can we do that, first of all?
Stephan Shakespeare: Yes. So on category view, it's exactly what we should have been doing. We should have changed the product once we realized what it was they wanted. And that was -- they wanted more questions. They wanted 2 things. They wanted more questions around specific use in that sector, so more sector questions. And they wanted more -- they wanted historic data that we had. I've tried to avoid going back over what didn't happen and should have happened in my periods. We had a head of product that came to us and did not -- decided not to make category view, not to go back to it, not to fix it, just to move on. And I think it was wrong, and we've picked that up. But it's -- of the 3 things, it's not our -- wasn't our #1 because it already had been dropped. So it's that we're definitely doing category views, as I just mentioned, and reviving it in a new area. The other 2, the -- Yabble, obviously, is a major contributor to the product that we're just looking at. The summarization, they're working out what the data means and so on. And the other part was, I think, yes, the interfaces. And those have been improved and there will be a continuous improvement. So I think that those 3 things, the category view is the one that didn't happen, but is going to happen.
Steven Craig Liechti: And you think that the 4 things that I said, i.e. repeating those 3 original plus the qual stuff are the key drivers for DP into fiscal '26?
Stephan Shakespeare: Yes. And I think you were asking as well, Johnathan, I think you were asking, was this a separate product or an add-on or just making it more attractive? I think it's a major new type of data, which means it will be deployed as an enhancement to existing things, which you can -- which you have to pay for, or it can be used on its own. So I can't go further than that because I don't know how you will engineer some of these things, on what order we would do it. Number one is it's an enhancement to Brand Index that you pay for, which I think would make Brand Index a more exciting product to sell and therefore -- and to buy. And it would mean that existing users who are interested in this data are likely to be early buyers of it. That would be the beginning. And yes, the AI tool.
Steven Craig Liechti: And I suppose we should put in, a fifth one actually, which is a more focused sales team as well on the product side. So that you've got the 5 things.
Stephan Shakespeare: Yes.
Steven Craig Liechti: So if you took those 5 things and I think about fiscal '26 between the first half and second half, last year, you did about 1% growth in DP in the first half and 1% in the second half, give or take-ish. How do you think that should flow through first half, second half? Is it kind of more of the same in the first half and then acceleration in the second half?
Stephan Shakespeare: Well, there are so many moving parts here and, obviously, I'm so optimistic that you shouldn't probably listen to me anyway. But I think that we will have a Capital -- I mean I know that we will have a Capital Markets Day. I know you're going to ask for these updates. And we will have that as soon as we have some -- the next stage of concreteness around this. I would like it to be very soon, but I can't promise it until I have a little bit more customer feedback.
Steven Craig Liechti: And then last question, just in terms of the overall profitability of the business, given you're putting in the investment that you talked about, this year, roughly, give or take, 30-30, won't be between the halves, if consensus is now low to mid-60s, how should we think about the profit flow through first half to second half?
Alex McIntosh: I think these investments will take us a bit of time to come through. So I think we'll start to see those coming in really Q2, Q3 of our year. So as you project before both going into the latter half of the year with a higher cost base. Of course, the thing that we're still sort of working on, and not to rehash this too many times, we could see an acceleration of client adoption around these things. But for now, we're being fairly conservative in terms of we just can't predict what the uptick will be, it will be very client-dependent. So yes, the main factor will be how fast can we find people and employ them within these data science teams. But we've already made some progress in terms of getting a leader.
Jessica Pok: Jessica Pok from Peel Hunt. I just have one follow-up, please, about the new products. Obviously, you're doing a lot of testing for clients in the U.K. who are interested. Do you -- I mean is the process now to deploy it to these initial customers' feedback, reiterate? I mean when do we get to the point, I guess -- or is it more second half of this year? Or are you really envisioning FY '27 of when you could possibly do a full launch of products? I mean, I'm assuming that whilst you're doing this testing, you're holding back a little bit on kind of showing it to all your customers and deploying it to all your customers.
Stephan Shakespeare: So I mean, if you -- which you do know the business, you'll know how easy it is for us to do this, because we do at least 20,000 surveys anyway every day. By surveys, I mean, interviews. And we can add this to the end of every single survey and say, would you like to talk about anything that's on your mind now? And actually, we will. We will add it to every single survey as just a final sign-off question. And they can tell us what a s***** survey it was or they can tell us they'd like to do some -- they'd like to chat a bit. And some of these people talk about their divorce. They talk about their football team. In the first few thousand that we had, the variation in those are incredible. People want to talk. Now what proportion? Every single survey that we run, we will ask them if they want to talk some more. So it's really easy for us to set this up. And we have a prompt that allows them to talk about anything they choose or to choose one of the things that we put in there. So they actually can use it the way that they want. And it's sort of pre -- sorts itself as it's going through. It's unbelievably simple and rich in its product and enhances user experience. It isn't a cost for this because we're not going to pay people except when we wanted to talk about something really boring. That's the bit about being a panelist, sometimes you have to talk about your use of OXO cubes or something, and that's not what most people want to talk about. Nobody ever wants to talk about an insurance plan, right? You have to pay them for that. So you do have to do that sometimes. But you don't have to talk to them about music -- pay them to talk about music or about their lives or whatever. And that gives you a lot of background information that you can then divert when you need to and offer extra incentives. So what I'm saying is that this will be very quickly engineered to the entire running of surveys that we do. And we'll be able to do -- create products of it, I believe, all over the place. But it will not run out for a long time of ideas. There's so many things to try. But it just sits there as something people want to do anyway. I mean I should say, just -- please, just last. We have had a box at the end of surveys for years, which nobody reads. And we realized we had 10,000 comments coming in the day every day that we ignored that nobody ever looked at. We still don't know what they said. I mean it's a very bad thing. And that kind of triggered this.
Unknown Executive: We have a question that's come in online. It's from Jonathan Cohen from Zipper Line Capital. And Jonathan's question is, SP3 called for 500 million of revenue excluding M&A and CPS, and 25% operating profit. Is that still what you're aiming for? And is that what you're guiding to in the medium term?
Stephan Shakespeare: Yes. And you know what I'm going to say to that, I'm going to say that we will have a Capital Markets Day and we will remodel everything for that. It's impossible for me to say that now. But I will say that we are a data company that is incredibly ambitious to be the world's #1 supplier of opinion data everywhere. That's what we can do. It's incredible that we haven't done more in that sense because you see what's there, it's all engineered, it's all there, and we just haven't done enough. So our ambition remains huge. Our execution has not been good enough, and we are doing quite a lot. We haven't talked about this, but we're doing quite a lot. The Board has been very active in helping improve execution at YouGov. We have board members that are actively involved. We have new Board members. Nobody has asked about the Board members, but you'll have noticed, we've had some very high-quality board members. A couple from Silicon Valley, a couple from very good experience in U.K. PLCs, one of them from Kantar. We have high involvement now from the Board in pushing for better execution. And so my answer to that, Mr. Cohen is -- and he's been an interesting contributor in comments, I totally agree that we are pushing -- that we are a data company that should have very high ambitions. And we will give you a more realistic steer on that at the Capital Markets Day that we'll have. We'll have it as soon as we can -- as soon as it's ethical for us to do it. In other words, that we have enough actual information, complete information to base it on.
Unknown Executive: Great. Thank you. Thank you, Mr. Cohen, for the question online. We don't have any more online. So unless we have any more from the room...
Stephan Shakespeare: Thank you very much.
Alex McIntosh: Thank you very much, everybody.
