Morning to everybody in America, good afternoon to everybody joining us from Europe. Thank you for joining us today, whether you're here with us in New York City or joining us online. My name is Ingo Middelmenne, as Head of Investor Relations of Verve Group, I would like to cordially welcome you to this year's Verve Capital Markets Day. Once again, we have had intense weeks behind us, the whole team and I are really thrilled to finally have you with us. As usual, we have prepared a focused and, I hope, very useful program for you. A business update from our management team, a closer look at Verve's commercial and financial development, a set of expert sessions on some of the topics that are shaping our industry right now. We will also make sure we leave enough time for your questions, both here in the room and from our audience online. Before we start, to make our legal department feel a little better, please enjoy the usual disclaimer on forward-looking statements. Now I think we all feel a lot safer. The internet is a powerful alternative to the walled gardens, it lives in your pocket. Your mobile phone, ladies and gentlemen, is probably the most powerful marketing tool in the world. Consumers around the world look at its display three, four, sometimes five hours a day, the industry is only just beginning to use this potential for targeted advertising with outcomes superior to those of any other advertising channel. Today, you will receive deep insights into how Verve transforms under-leveraged mobile signals into premium advertising outcomes. Ultimately, aiming to be the first company to unlock the full potential of the open internet by creating closed loops, that so far have only been seen within the walled gardens. First, let me briefly introduce today's speakers to you. We will start part one with Remco Westermann, our Chief Executive Officer, followed by Christian Duus, our Chief Financial Officer. After the coffee break, we will then move into the expert sessions of the day. I'm very pleased that Dr. Usama Fayyad will join us for today's keynote. Usama, thanks for making that possible. After that, it's Raimund, President of Verve Retail Media, who has built a truly phenomenal thing together with his team over the past 11 months. He will talk about the closed loop we created in retail media, a key driver of our future revenue growth. Finally, Mishel, our Chief Business Officer, who's always fascinating to listen to. Mishel, great to have you back on stage this year as well. Mishel will talk about how predicting consumer intent will fundamentally change how consumer targeting works in the future. How will this all unfold today? We're now starting with part one, the Verve business update. In a moment, Remco will welcome you and introduce you to Verve's general business model, followed by his commercial update. At 10:00, Christian will take over for the financial session, followed by our first Q&A at around 10:30. At 11:00, we will then enjoy a quick coffee break, let me highlight already now, please be back at the room at 11:30 sharp. Then we're going to continue with part two of our day. Those are going to be the expert sessions. First, the keynote by Usama, then Raimund on retail media, then Mishel on predictive targeting. We will close the expert sessions with a second Q&A session at 13:00, then enjoy some lunch at 1:30 P.M. With that, let's get started. Remco, the stage is all yours. Thank you. Thank you, Ingo. Yeah, welcome, everybody. I also see some people from Europe that made a far travel to visit us here. We had some discussions, where do we do our capital markets day? The last years we did it in Sweden, now we decided to do it in New York, in U.S., to really emphasize how U.S.-focused we have become and how much opportunity we see in this market here. I'm going to talk about, or actually take you into our equity story, our commercial update. We've been building or I started this company 13 years ago, and if I look where we are today, I'm proud, but I'm also seeing a lot of opportunity to even get much bigger, much stronger. How we want to do that, where we are focusing, that's what I would like to take you through. Ingo mentioned this already. The mobile open internet market. Open internet, not being walled gardens, Facebook, Amazon, et cetera, Google. There's a huge opportunity here. Mobile internet has been underserved and underestimated for years. When you talk to an agency, they said, "Mobile, that's kids playing games, sitting in the basement." That's changed. That's changing. There's a huge window that's open now, and I would like to first take you through the market. Why mobile? Mobile is the only advertising channel with a one-to-one consumer relationship. It's always on. People use on average five hours a day, 90% of their time inside apps. No other channel comes close to that. It's more persistent than any other medium. It travels with you. It goes everywhere where you are. Unmatched signal depth. Because it's always with you, we have a one-to-one. It's not one-to-many, like with CTV or many other channels. It's very good in capturing intent signals because it knows what you want. With an SDK integration, software development kit, that we technically integrate into the apps, and with that we get the signals, we can play ads, et cetera. We are really very deep into that. There's a direct trusted connection between platform, publisher, and consumer that we create. Real intent signals, we'll cover that later in more detail, and especially Mishel will go deeper into that, are really great to see what people want to do and to target advertising in that. Closed loop outcomes. One of the big issues of the advertising market overall is measuring. When I was at university, I learned from $1 spent, 50% is wasted. I would say that it's much more than that. It has to do with attribution, who organized the sale, or who organized the action. It has to do with where do you spend your media, for what price do you buy them, how many intermediaries are there. There's a lot of stuff there. In the end, it's about measuring outcomes. Did I really accomplish with my dollars what I wanted to accomplish? A bit about market size. We're talking about a global mobile ad spend of $670 billion. If we take search out, which is a big part, then we still have a $450 billion addressable market. Search is interesting because of LLM, search markets are changing. There's a lot less search happening. There's budgets that are coming available from there. If you then take this $450 billion addressable market and take the open internet share, then we talk about TAM of roughly $68 billion, which is big compared to our revenue. There's ample possibility to grow. Why is it only 20% or less than 20% of the total market? It's cheaper. Cost per reach is better. Budget just hasn't followed the action. I come to that in a minute. If this market would grow faster, if, let's say, it would take a bigger share of the total, so the open market, and that's what the parties in this open market have as their reach, then it could be a much bigger TAM, actually. That's one of the issues that I'll also cover here, that the open Sorry, I'm going one slide too far. That we see that, let's say, the walled gardens take an enormous amount of spend compared to the reach of consumers, but I'll come to that in a minute. A bit about LLMs. LLMs, important. LLMs have really taken over a big part of search. They are disrupting consumer behavior. How many of you are using LLMs daily? I guess everybody. It's changing the market. It's changing where ad dollars have to go. What we see happening in this market is that basically, consumers are moving in two directions, away from search into an AI answer layer, where they get dialogues, questions being answered. I'm interested in a car, what kind of car, how many people, these kind of things. You get an interactive kind of intent behavior. There is a lot of traffic also going into the high intent app layer. Apps, deep utility, real-time data, trust, and compounding personalization. 60% of searches now end inside LLMs. Traditional search is projected to fall by 25% by 2026. Special interest apps show 15%-20% increases in session depth and in retention. AI is not our enemy, it's our tailwind. Comparing the open internet with the walled gardens. There's a huge mismatch between spend and ad dollars. Between, sorry, time spent and ad dollars spent. The internet, 96% of online time, but only 20% of ad spend. Walled gardens, 31% of online time, but 80% of the ad spend. If you look at the cost side, ad spend per user hour, $0.07 for the open internet and $0.14 in walled gardens. There's a huge opportunity here. That's inefficient. Arbitrages as large as this one don't persist for long. A rebalancing should be coming, and we are well positioned to take part of that. Why is there such large discrepancy? Open internet is not perfect. There are several challenges. The first one is scale. I had a discussion four or five years ago, that was really, let's say, opening my eyes with one of the Holdco bosses, Holdco big advertiser agency in the U.S., he said, "I have $6 billion that I need to deploy every year. I go to Google, they take $ 2.5 billion. I go to Facebook or Meta, they take $1 billion. I have to work with numerous small parties like yourself, where everybody can take $ 20 million, $30 million, whatever million dollars." I think that's one of the problems of the open internet. We are over 500 companies, people are not really able to run big budgets. That should change, we are working on that. Cost efficiency and transparency. We see ads that go over 14 platforms, everybody taking his margin, totally inefficient, not transparent. Quality and depth of data. If you want to target, you need good data, great data. The big advantage of a Google or a Meta, they have such deep data. We have been concentrating on this in the last years to also build a data set that's comparable. We'll cover that later. Targeting and signal quality. If I have the data, I need to use it. I need to target. I need to make sure that I don't waste advertising money. Measurement and outcomes. I need to really prove afterwards that I reached the right people and that the money was well spent. The players who solve these challenges will define the next era of the open internet advertising. Coming to the next part. What are we doing? We have a huge market that has a lot of opportunities, a lot of disruptions. What are we doing? I'll go deeper in where we focus at and how we build Verve to win. First, we brought walled gardens performance to the open internet. What are we addressing? We are addressing the scale. You have all had, as an investor, the pain of that. We are integrating what we have. We have SSP, DSP, the whole tech stack, vertically integrated, very important. The closer you get the advertiser to the publisher and to the consumer, the more efficient the whole thing is and the more transparent. Depth of data and curation. The last years we have been talking a lot about ID-less targeting, but the world is more complex than that. It's ID-less and ID-based targeting and a lot of other stuff. That's what I would like to talk about. The targeting and measuring of the outcomes. How do we do that? How do we get better? How do we make sure that the advertiser gets more efficient? The last point is corporate development and balanced growth. Of course, very important as a company. We're still a small company in this space, even though we've been growing substantially, but we need to balance our investments versus cash generation, et cetera. Christian will talk much more about it, but I'll just touch the top of it. Skill levels. We have, I'm pretty proud about that, built a top 15 ad tech platform in under six years. We started six years ago when we were still a publisher. We started as a publisher 30 years ago. We said we are totally frustrated about advertising, we see a huge opportunity, we started building our ad tech platform by acquisitions. We did 15 acquisitions, then we didn't do acquisitions for two years or a bit over two years. We did 3 more acquisitions. That really built our platform. We're directly connecting advertisers and publishers. We did EUR 579 million of revenue in the LTM to end Q1. We have over 4,000 software customers. Over 1,100 of those are large substantial customers doing over EUR 100,000 per year. We integrated over 65,000 apps. We reached 2.5 billion consumers. We served 1.2 trillion ads, that's a huge number, in the LTM. We have a bit over 1,000 employees at the moment. We have reached scale, substantial scale. We can run large budgets. We can also get stronger, and we can further build. That's what we're aiming at. What are we doing? Let me start with the advertising agency side. We focus on mid-size agencies, selective direct customers, and wholesale agencies. With over 4,000 agencies alone in the U.S., that means that you need to have a substantial sales force. We're building our sales force. We were 54 sellers Q1 2025, 117 by end of Q1 2026. We're targeting over 200 by mid-2027. Otherwise, we cannot talk to all those agencies, and it's lost potential. Very important, we also need focus. We're not covering the whole world at once. We're focusing on our core markets, which are U.S. by far, U.K., Australia, Germany, and Scandinavia. Yes, we could start other markets, but we have so much potential in the markets where we're focusing now that it doesn't make sense to go further at the moment. Also the quality of relation is important. That ensures growing the share of wallet. Further optimizing our quality of service, unique product offerings, and Mishel will talk about it with you later, Raimund as well, by the way. It's about curation. It's about ad management. It's a lot of things that we offer to our customers that make them happy with our service. Then it's the other side, the publisher side. We have a pool of publishers. We have one of the largest in-app publisher bases in the ad tech world. We still can grow. We can grow with some larger apps that we are still missing, but we can grow also with specialty apps. For example, going down into sectors, we're focusing on sectors. Retail and CPG is one. We need cooking apps because that's a great place to catch people that later go shopping. Those kind of things. That's what we're working on. We're also working on other channels. Yes, in-app is core, and Ingo started with it, mobile phone first, because that's where you have the data, that's where you start your journey. Still, it's very strong to enforce consumer behavior by connected TV, by Digital Out-of-Home, also a screen in the supermarket, for example. Digital audio, even web. What we want to do is to really be able to run larger budgets, scale again, but also to reach the consumer, starting from the mobile phone, also on other channels. What's very important is, again, to mention here on the publisher side, SPO, supply path optimization. We build this by acquisitions, so quite a bit of our traffic is still also indirect, where we don't have direct to the publisher, and that's what we are at the moment with high energy work. Dave is also here, who's heading our marketplace. It's one of the big tasks here to really make sure that we become the prime quality exchange in this part. That means we have cleaning that we're doing, lower margin, not so great, let's say, paths, indirect paths, we want to get them exchanged for direct publisher relations. That's the scale. We get to vertically integrated. We built 18 acquisitions, or we did 18 acquisitions that made us build this platform. Last year, you had the pain of it in Q2, Q3. We had our last two SSP integrations. We acquired five SSPs. We integrated them into two, and last year, we integrated those two into one. We had problems with load balancers. It really hurt our revenues last year. The result is a single unified platform, and that's a good basis to work from. We have a DSP, we have an SSP, we have the data part, and if you open an app, in less than 100 ms, the ad spot that's on there needs to be filled. We do that. We technically integrate it in the app. We get the signal. Here's an ad for sale. We get some data with it. We enrich it with further data that we have. We send it to the demand side. The demand side has put in who we want to reach, what's their target, their KPIs, their budget. The bidding process happens. Highest bidder wins. If we are the highest bidder, we serve the ad. It's like a stock market. Important to be the winner, of course. Yeah, only when we serve the ad, we also make the money. What's next? After we integrated everything, I need to make one comment here. After we integrated everything, we're now really working on how can we improve it? How can we get much better? You see some of our peers in the markets who have really been good at that, and I would say AppLovin is really extremely good in, let's say, optimizing their platform, their bidding algorithms, things like that. You need a well-running platform for that. You can really work on those things. That's what we're doing. Let me first start with final integration step. We're not fully done. CTP, as was mentioned also in earlier presentations, is still not fully migrated, but we're almost done. In the slide, we are finishing it by end of Q2. That's only two weeks to go, and that will happen. We will have it fully integrated for no CTV running on the old platform anymore. Yeah, routing more demands through direct supply. With one integrated platform, we can also start working on optimizing the platform for the demand side. Our DSP has performance customers. They would like to run on certain apps. There are certain features, certain SKAN features. It has to do with Apple, I think, that needs to be met. That demand is not yet running on our platform fully. It's running a bit, but we have a lot of potential there. Of course, that's not really giving us extra revenues. If we run our demands on our own supply, it gives us extra margin. That's one of the big focus points. The third, as Jeff already mentioned, it's super important that we further improve the, how do you say it? The quality, the technology of our platform. We call it an ad intelligence platform, and a lot of it is about AI. We have Usama also here to talk about AI later, and we didn't want to go too deep into AI to split the topics a bit, but this is one of the core, core things to further drive revenues in this company. Just to give you an idea, and please don't calculate this as the revenue multiplier, but we roughly get 1 trillion ad requests per day on our platform. We serve 1.2 trillion ads per year. That means that 1 out of 300 ad requests, roughly, we only built into a served ad. You have to deduct from those 1 trillion ad requests that we get, ad requests that come from countries where you cannot make money. Comes from, let's say, places in the app on the part of certain apps that you don't want to sell. It might be in the invisible area. Still, even if it's not 1 out of 300, it might be 1 out of 100 or something like that, where we are still at the moment. There's a huge opportunity there. That means that we are investing and focusing a lot on improving our AI capabilities because that is a very, very strong opportunity for this company. Apart from that, we also work on agentic. Agentic is big. There will be more automated buying. You can do a lot of things with agentic. You can also prepare your platform for agentic buying. Those things we are working on. Yeah, there's a lot of opportunity. When it comes to data and curation, we've been working the last years on really becoming top-notch in data. Data, in the end, is what you need to do your targeting and to measure your outcomes and all these things. Data has become more and more complex over the years. Advertising started with just doing some ads, not measuring. The cookie came, and you use the cookie to really build a profile, an ID graph. Can I match this cookie with this mobile ID? Is this person interested in football or buying a car or things like that? A lot of the industry is still working with cookies. Cookies can be very polluted. Your daughter borrows the phone and suddenly you have a different kind of profile on it. Or, that's even more waste, how often do you buy a car? Once every three years? Once every four years? There is, he has a car, is in the profile that she has a car. That means that you get car ads all the time. As an advertiser, as a car manufacturer, you only want to advertise if the person is interested in the car at that moment. You might do a brand campaign around it, but the real time to be there with your ad is when the intention is there to buy a car. That's super important. We have split our data into 5 tiers. The first one is cookies and opt-in. That's the top one. Contextual intelligence is the 2nd level. Are you in context? Are you on a car site? Are you in a weather situation? Whatever. Contextual intelligence, geolocation, app category, those kind of things is that. Level 3 is that we have built SKAN on-device intelligence. We are with the SDK, we have put an extra piece of software on the phone that can really build segments on the phone. That information doesn't leave the phone, so it's privacy compliant. If an ad request comes, we can attach the category, the interest category, and that works well. With active search intent, with the acquisition of Captify, which was a great acquisition, we were able to acquire the search intent capabilities and search intent integrations. For a lot of U.S. publisher properties, we get intent from people. What are they looking for? The last tier 5, is conversational intent, zero-party data, asking questions, are you interested in buying a car? Yes, no, et cetera. That's circumstantial. Also LLM data. To work with them. Still, we need to pick the right ones. We need to save the right ones. We need to build the right profiles, things like that. We further focus on LLM derived intent signals because that's where we have really good results. We're working on this flywheel. More publisher integrations, more signals, better predictions, better outcomes, and in the end, more revenue, more advertisers. In the end, the advertiser is interested what really makes the sale, what really makes the person, yeah, buy a product in the end in the store or online. This measuring has become more and more difficult. Cookies are disappearing, as I just mentioned already. They're not disappearing. There's still a lot of cookies, and there's also still device fingerprinting. The industry has become more complex. There was a time that we thought cookies are disappearing. Google said we also deprecate a cookie. It didn't happen. What do we have now? With cookies, without cookies, and a lot of, how to say, confusion in between. The industry has become more complex. There is some because that gives the best result. Google said we also deprecate a cookie. It didn't happen. What do we have now? Well, with cookies, without cookies, and a lot of, how to say, confusion in between. The industry has become more complex. Just to say, Apple, with 50% market share, roughly in the U.S., ask people, "Do you give consent that your data are being used?" 80% are saying no. 40% of the market for sure doesn't have a reliable cookie. Sometimes they're fingerprinting. There is some fingerprinting, which is not allowed, and it will probably also further disappear. The market has become complex. What are we doing? We're not saying we do only ID-less. We do wherever it fits best, and we use the signals combined, because that gives the best result. We operate across ID and ID-less of environments with ID-less as the baseline and ID data enriching where it exists. LLMs reach contextual intent at the moment of position without assistance in identifiers. Creative signal match pairs the right format and the message to the intent. That's what we didn't talk about. Very important also in advertising is creative. A good and a bad creative can make a factor three of difference in effectiveness of your advertising. Which data we have, male, female, is already very important. The car is very often advertised in a different color for male or female. Also all those things work together. We're talking about a super complex industry where a lot of things have to be in place, and we're now with an integrated platform. We're really working on fitting those things and making something that really working very well. Sector-specific closed loops in the end bring the most value, and that's what Raimund will talk about later. Just targeting and measuring. What are we further doing there now? We did a lot already. We are very far. We want to make media better. We want to support our customers to get better ways to target and to measure the outcomes. Closed loops, sector focus. We have decided you cannot do that for the whole industry. That's just too big. You need to go deep. You need to have sector experts if you want. Retail and CPG is one of our largest sectors. You need the expertise, you need people that understand the market, and you need to build your solution for that. We are hiring sector experts, adding sector-specific data sources also, and for example, point-of-sale data. We want to know if people really bought a product in the store. We are continuously improving our targeting capabilities and benefit from improved measuring capabilities. Corporate development, balanced growth, very important. As you saw on the previous four points, this company is not ready. We are work in progress. There's so much that we need to do. That means also that we're investing. We're investing extra salespeople. We're investing in our technology stack in AI, and we grow our verticals also with expertise. That's money that's, I think, well spent, well invested, but it takes time to get yield out of that. Christian will show the numbers later. I think with over 30% revenue CAGR and so much opportunities, we are really well-positioned for the future. It's not only about growth. It's also about controlling costs. It's also about internal efficiencies. With AI, we have a lot of opportunities there. Also without AI. We have focused our people more on where we think the growth is. That's also that we cut quite some countries out. With AI, we can really build strong efficiencies, and there's some examples in the company now, and there are only first examples. We have a regional sales team that is selling ads from local advertisers, where we increased the efficiency tenfold. One person is doing tenfold the output as before by fully automating their processes with AI. We have another department, where Raimund is working on at the moment. We were just working on the 25% personnel cost saving and expecting minimally the same output after that. With AI, we really see big progress in AI. Someone will say more about it. It's a no-brainer, it's not that easy. Still, we see really good use cases in the company. Focus on internal efficiency is very important. Organic growth. We had a lot of discussions in the past, and especially me. I like to do M&A to grow, to build critical mass. I think we have really come at a stage now where we have all the tools, all the technology in-house. On the data side, we've really done good things. I don't see a lot that would make sense for us to buy. It would more distract us than really bring us further. One exception, I believe that this market still has two small parties, we need to see bigger mergers, bigger combinations, business combinations in the future. If there's something really accretive and very big, it might make sense to do it. For the rest, we as a company want to and will focus on organic growth. Very important. I showed the opportunities that we have, there's so many. Coming to the last slide of my presentation, I've seen that I'm a little bit too fast actually, Christian can start earlier. We want to bring the same performance that a walled garden has proven that they were able to do. We bring that to the open internet at scale. We have been working on exactly the things that you need to be comparable to a walled garden, to be able to really have the skill, to be vertically integrated, be efficient as in that sense, have the depth of data. Have the targeting solutions, is able to measure the outcomes, also having a balanced financial profile where we invest, but also make sure that we control our costs and further become profitable. That brings me to the end of my presentation, I would hand over to Christian for his. Christian. Thank you very much, Remco. Now to a overview of our finances. My name is Christian Duus, I'm Chief Financial Officer of Verve, I'll try to bring you an overview where we are now also what lies in the future. If we start very overall, here depicted behind me, our revenue growth our growth in adjusted EBITDA over the five past years, you will see that we are, in essence, a growth company. We've grown on average 32% our revenues each year on average 36% our adjusted EBITDA. That's a phenomenal performance, it really talks to the success of the company. We continue to have strong ambitions to grow beyond market rates also going forward. I think here we have also shown the ability, we've gone through various investment cycles through those years, where we have demonstrated that we are able to convert our investments into revenue, top line growth and profit growth. We've also been through a number of different challenges through those years. We've gone through COVID, we've gone through trade tariff wars last year, we've gone through FX headwinds for the same time, and we've also last year gone through the reunification of our platform. I think it talks to the robustness of our business model that we're actually able to, over the years, to combine all that into a profile like this. We remain very ambitious on part of the company. We are looking towards, if we take the midpoint of our guidance for a lift on a reported basis of 28%, on a like-for-like basis, 17%, and we have a lift in our midpoint guidance for our EBITDA of 19%. We continue on a path of growth. You might ask, what is going to drive that growth? What is driving the growth? Well, there's a number of opportunities in the market. Remco already mentioned some of it. Some we'll hear more of. I think the way that we like to present it and we think about it is basically in four layers. These four layers. Number one, tailwind from the markets that win. We are in mobile in-app. It is one of the places where people use most of their time, and that's where also most of the investment dollars goes. Just by being in this sector and being in the right place in the market, we have a natural tailwind of something like 10% growth from the market. There's customer expansion. We have typically been very successful in landing customers and expanding customers and growing the share with customers and also into new verticals as Remco was talking about, and we'll also hear more about that later today. Of course, new products. Be it AI-based or other products that tries to expand what is our share of wallet with each customer and also being relevant for each customer. Fourthly, in this industry, there is a number of platform synergies. The bigger you get, both from scale, but also the more relevance and the more inventory you have, the more relevant you become for buyers. Vice versa, the more buying power you represent from the demand side, the more relevant you become for publishers. There's also other benefits which are significant, which is that you actually can extract all the learnings from all the data that you're treating. The more data you have, essentially the better you are to train your algorithms and benefit from that. Through those four layers of growth paths, that's really what underpins our growth journey. I think what's important to note is it's not just one. It's not a singular catalyst that is driving our growth. We are basically working on all four levels to underpin our growth. If I take a step closer and just to give you an overview of how are things going now, what's the current financial momentum? Well, you'll see here our revenues and adjusted EBITDA for Q1 and the four preceding quarters. We had a very good Q1, a very solid Q1. We had a strong organic growth of 6.4%. That was, by the way, up from the previous quarter from 5.3%. We are accelerating. We did EUR 137.2 million in Q1 versus EUR 132.3 million in the preceding year, which is a like-for-like growth of 3.7%. We did EUR 28.3 million of adjusted EBITDA. That's slightly down from the prior year, where we had EUR 30.2 million, and I will come to speak more about that. That is basically because we are in an investment phase of the company. We are expanding our sales capacity, and that means there is a drag on our EBITDA for Q1. We will see similar in Q2. It will flip in second half of the year. We are consciously taking investments to expand our relevance in the market and our coverage in the market. As Remco talked about, there are numerous agencies out there that we simply do not have the firing power to cover. That's a little bit the financial performance. If you look at quarter, and I think that's interesting too, the sources of growth that we have right now, and I'll just take Q1 as an example, because we land at a numerical number which is maybe perhaps a bit smaller, 3.7%, but actually, that's the combination of 6.4% organic growth plus 6.9% in acquisitive growth. We saw in Q1 a very significant headwind from FX. The majority of our revenues are denominated in US dollars, but we report in euros. Because of the depreciation of US dollar, we basically get a headwind of 9.6%, which lands us at 3.7%. A strong entry into Q1, continued somewhat drag into Q2 from our investments, with a good insight into second half of the year, we will see a good traction for the year, which is, by the way, also the big season of the year. Our seasonality of our business is normally that Q2 is bigger than Q1, Q3 bigger than Q2, and Q4 is by far the biggest quarter for the company. Underlying this is basically the growth of our customers, the growth of our business. Here you see the four business KPIs that we normally report on, which is a blend of our impressions and our number of customers, and also our ability to grow our business with customers. Just trying to break it down for a second. Here on the top hand right side, you see our ad impressions. We grew 25% year-on-year, our ad impressions. Doing 25% more impressions, which is a significant growth, just tells you how relevant our platform is with our customers, that it has a uniqueness and a relevance that is appreciated by customers. 25% is quite a big number when you think about where we are in the economic cycle. We've also been very good at growing our number of customers. We are right now at 4,086, if you can be that precise. We're roughly taking on board 1,000 customers if you take it on a full year basis. Granted, that is a mix of organic and inorganic growth in number of customers. We see that as a real opportunity to also grow these customers to become scaled customers and reach the level of software clients, which is defined as customers that use more than EUR 100,000. That number has been stable. We see the possibility and the opportunity to grow all the customers we've now onboarded to add to that number going forward. You can also see client retention has been strong through the period. Some customers might have used us a little bit less, especially in Q2 and Q3 last year, but they remain loyal to us, and with the possibilities to expand further. All in all, very strong customer metrics. Are they all perfect? Not picture perfect, maybe not. I think it shows really the potential of the business, where we can go with the business, and how we've performed over the last year or so. Gross profit margin. We've started to report our gross profit margin. I'm super happy to report that the very strong uptick we saw in Q4 from having our unified platform has carried into Q1. We see a gross profit margin of 41%. Excuse me. That's 2.7 percentage points lift versus Q1 2025. You'll see here in the dotted box the comparison. We do the comparison towards the same quarter previous years, because there are seasonality effects that also impact the gross profit margin. That's the best and most clean way to actually look at gross margin. We've had a significant lift both in Q4 last year and now in Q1. It's really the results of three things. Number one, we have a unified platform. This unified platform has a much richer feature for us to do dynamic pricing. It enables us to, on a much more rigid and nuanced way, to manage and extract margin from the impressions that we serve. We've been much better at managing our hosting costs and bringing down our hosting costs. Remco also talked a little bit about just the possibilities of making sure that you actually address the parts of the request that you want to address and not receiving too many requests. You can work a lot with that. We are also doing the different exercises in looking at, let's say, the more low-margin inventory, the more non-core long tail, to prune out that. We have a lot of things going on that supports our gross profit margin. We really think this is the key indicator for how we see the future, but also to improve our cash flow generation for the company. Remco mentioned, I've mentioned, we are, right now, in these quarters at least, in a bit of an investment phase. We have a very strong scaled sales machine, sales team. We have 117 people right now in customer-facing sales roles. It's working well. It's been growing quite significantly with 54 in Q1 2025. We still see the possibility to add more salespeople on an already well-working foundation. For that, we have set aside, for this year, EUR 10 million in OpEx investment that we will support, adding more salespeople towards the target of 200. Naturally, or at least in our business, there is a lag from bringing salespeople on board and somewhere between 9 to 15 months before a salesperson really proves out and starts adding back to the company. There are different realities. It takes time to hire these people. You will also have some salespeople that work out, and you might also have some that don't work out. Therefore, it can be a bit difficult for us to precisely exact what is the inflection point. When will all this investment pay off? We started this journey already last year. We are seeing good results, and we are progressing very well. We will see the effects here in second half of the year, perhaps mostly towards Q4. Cash flow. Cash flow is, of course, an important point to talk to as a Chief Financial Officer. I am super happy to report that we now on an LTM basis with Q1 managed to rebalance, so to say, our net working capital and perform well on cash flow. On an LTM basis, March 26th, we report EUR 105 million in operating cash flow before working capital and EUR 94 million operating cash flow after net working capital. Those two bars that you see here are ideally equal height. That means that we are not investing too much in our net working capital. In this, for LTM basis, we have invested EUR 11 million. But you will also see a marked difference from the prior reporting, where there was an imbalance in our investments in net working capital. So we are back on a rebalanced way for our cash flow. You can see here on the right-hand side also our cash conversion. If you look at our cash conversion before working capital, it is actually pretty stable and at a high level, around +80%. We have had certain periodic swings, which is basically a result of our net working capital. I wanted to dig a little bit more into it because I think the cash flow is super important part of this business. We are at a phase where it is important that we generate cash flow also to repay our bonds and generate free cash flow for all stakeholders. You can really see here with this chart a little bit what is going on. So we have depicted our trade receivables in dark blue, our trade payables in light blue, and the green line is the net working capital, the difference between the two. You can see here the imbalance that we saw in Q3 and Q4 end of last year. You can also see that it has been rebalanced, not fully, but rebalanced to a level of EUR 22 million for the Q1 2026. Why does this net working capital need happen? That happens because structurally, if you look at the company, we essentially, on average and per payment terms, we will pay our publishers, our big suppliers on 45 days, on average. But we receive money from our customers on a 90-day basis, and that creates this gap of 45 days. We have historically been very good at managing this. You can see the green line has been close to zero through the different quarters. We have done that through mainly our securitization program, which I will come to in a second. That has really been instrumental in managing our working capital needs. Overall, you can see the profile. When we grow, we will have certain investments in working capital because of this gap of 45 days. Just to come a little bit into this part around our securitization. We have a receivable securitization program of EUR 100 million, and it is in place both to manage our working capital needs, but also in place for us to be able to really offer good payment terms to our publishers, to our customers. We use that as a moat. We use that as a way to be attractive and also build our position with smaller customers, and the more flexible we can be with that, beyond also, of course, having a great product, helps us win new business. The unfortunate thing, or at least from a pure finance angle, is that it creates this cash gap of 45 days. Higher growth means higher working capital tie-in. We have a securitization program in place through very well-seen providers, Finacity and Nord/LB, who are very active in this area and offer similar solutions for other companies. What we have there is a possibility to sell. We sell our receivables on a non-recourse basis, and we receive a funding up to a frame of EUR 100 million on very lucrative terms, Euribor + 2%. That's because it's a Grade A client portfolio. We really use this securitization program to manage our working capital needs, but also to be, or intensify our relevance in the market and use it as a moat. I'm happy to report that we are also expanding the coverage of the program to further two entities that have just been approved, and we are adding to the program. Turning now to CapEx needs for the business, I think, if you look at Verve historically, we've always said our maintenance and investment CapEx is roughly in the order of EUR 40 million -EUR 45 million. You can see it's quite stable. The dark blue and the light blue here. It has been very stable over the last past years. That really means that we have fairly low recurring maintenance CapEx of roughly EUR 9 million a year. The ability to scale our revenues while keeping CapEx low is, of course, very interesting to help us generate the operating cash flow for the business. In many ways, we're quite asset light in our scaling model because we continue to reserve roughly these EUR 40 million -EUR 45 million in CapEx going forward. You can also see here on the left bar that this year we will have, with acquisitions that we did in 2024 and 2025, we will have deferred payments out of roughly EUR 34 million, taking our total CapEx to somewhere between EUR 74 million and EUR 79 million. Maybe just to mention that out of the EUR 34 million that is on deferred, we have already paid out Jun Group of EUR 23.8 million. The only part that is remaining from a cash-out perspective is EUR 9.9 million for acardo and Viewento, which comes in October 2026. Remco mentioned it, we are really now focused on organic growth, very selective in looking at any M&A. The main focus here is organic growth for this year. Brings me to balance sheet KPIs and how we're doing on net interest-bearing debt and especially net leverage ratio. You'll see, we actually did a tap of our bond here in Q1 of a EUR 50 million tap on top of our EUR 500 million bond. Actually, as you see, the net interest-bearing debt is basically the same between Q1 or LTM Q1 and 2025. 2025, EUR 446 million versus EUR 448 million. We are now at a net leverage ratio of 3.1. It is elevated. It is elevated because of the acquisitions that we've done through the last couple of years, and we are working and do have focus on bringing it down. It's always a question of trying to balance the growth with also having a sensible leverage ratio, and we target 2.5 times in net leverage ratio, which will come in place, I would say, within the next 12 months. Maybe just to mention, I think when you look at our bonds, we have EUR 550 million outstanding in bonds. Here is depicted the three last bonds that we've done. We are now at a company that has quite a lot of experience and also quite a good reputation in the bond market, and we use the bond market, I think on quite attractive and favorable terms to fund our growth. That way when you look through the past history, we have actually been very successful in refinancing our bonds well in advance of any maturity walls. Our current bonds go until 2029, when we will continue that practice of refinancing well ahead of any maturity wall. Number 2, as you see here from bond three and four to the current bond, we actually also managed to reduce our funding terms or credit margin quite significantly from an average. We had two existing bonds, one on three months Euribor plus 625 basis points, and the one on 725 basis points, and we have reduced it to three months Euribor plus 4 percentage points. The last EUR 50 million has slightly higher interest rate, but basically a significant drop in our credit margin. This helps us, of course, to reduce our interest payments for the year. Now I come to guidance. Before I maybe just go into the specific numbers on guidance, let me just come back to this picture of the 4 growth levers that we have for the company. You'll remember that I went through it, the market tailwinds, the customer expansion, product strength, and also platform synergies. The way that we think about guidance for this year is we expect growth from our market growth of somewhere between 7%-9% plus 5 percentage points growth coming from customer expansion and new products. That's basically the basis for our guidance for this year. With the traction that we had in Q1, solid growth and also solid performance, and with the outlook that we see into Q2, we're happy to confirm and reaffirm our guidance for the year, which is EUR 680 million-EUR 730 million on our revenues, 17% growth on the midpoint, and also adjusted EBITDA of EUR 145 million-EUR 175 million, which is a 19% pickup on EBITDA. By the way, those percentage points are calculated like for like, so you can see here in the guidance exactly because we've had a change of revenue recognition for some parts of our business, which we have provided on a like for like basis, so that you can see that revenues for full year 2025 would have been at EUR 602 million within similar EBITDA of EUR 134 million. Maybe important to note that we will be talking a lot about Retail Media later today. We have included some pickup in the Retail Media business in our guidance, but not I would say the full expansion is still to be formally included in the guidance and that's because, as Raimund will talk about, it is a bit difficult to always predict the exact infection point. That is upside to our guidance. Last, wanted to touch on the topic of relocation to Ireland. Why are we doing it? Well, it brings a number of advantages for us. First and foremost, it brings us into a legislative and corporate governance structure, which is more agile, but also more comparable to our U.S. peers. That's number one. Number two is that by being in Ireland, it actually gives us the opportunity to report in U.S. dollars versus now where we are reporting in EUR. Reporting in EUR when the majority of our business is in U.S. dollars always gives us some challenges around explaining growth and what is the growth in local currency versus in EUR, and all this noise can be eliminated from our reporting, which is a huge benefit for us in providing clear communication to the markets. It also provides us some flexibility and an option to do a direct listing to the U.S., and that option, of course, has quite good value for us, and we want to maintain that optionality for later on. When I close here on the financials, I really see Verve entering the next growth phase. We're confidently investing in a very targeted way where we see returns are highest, but with a very strong focus on also preserving capital discipline, cash flow profile and financial strength such that we have the support and the basis for long-term shareholder value. It's balancing those two elements. Thank you very much. Great. Thank you, Christian. Please stay right on stage with me here and Remco, please join us on stage. It's time for some action, some interaction. I can literally see it in the eyes of our guidance here. They're screaming Q&A, Q&A. You want it, you're getting it. We're moving into the first Q&A session of the day. For those of you here in the room who want to ask a question, and we would really welcome that, please just raise your hand and one of our microphone runners will come up to you so you can ask a question. Also for those of you joining us online, please ask your questions via the chat. I have already received some questions. This Q&A session will be addressed to Remco and Christian. In this session, we will cover the business, commercial and financial update as just heard, and later on, we're going with a tech Q&A. Are there already any questions here in the room? Not yet. We're jumping into the first questions from the chat. I would say first question, probably best to Remco. Sounds like you're placing quite a significant bet on retail media. Why should investors believe that the walled gardens will not start targeting retail chains as well and use their enormous financial power to push players like Verve out of the market? Good question. Thank you. Maybe first to start one step back, why are we doing retail media and CPG? It's the logical further development of what we have been building over the last years. We've been building a very strong exchange with a very strong demand side, having the relation with advertisers and the publishers. We have found out, or let's say as a conclusion, that you need to focus much stronger on the sector to really get the results that an advertiser wants. If we look at our larger sectors, then retail media and CPG, they go hand in hand, of course, are, let's say, one of those sectors where we really see a big opportunity. I don't want to take all the answers away from Raimund. He's going to show quite a bit what we're doing there. We've been building that up over the last years also, just to make that clear, we have built an extra closed loop on it. In that sense, it's a logical further evolution of where we're coming from. Why are we not afraid of the walled gardens? As I showed before, the walled gardens take 80% of the spend with 30 something percent of the views, of the time used. There is so much room next to the walled gardens. This market is so big that even if walled gardens would get in, there's enough room. If we deliver better results, better outcomes, better service quality, there's no reason to be afraid. It's a huge growth opportunity, I think also for walled gardens. Look, for example, Walmart has built a very strong position in there and also built its retail media. It shows the potential of it. Raimund is also going to show some numbers, I think, in his presentation. It's a hugely growing market with a ton of opportunities. Yeah, we as an open internet need to organize ourselves and need to focus more and to deliver better product. I think that's what we're doing, and therefore go full forward for us. Great. Thank you, Remco. One question I actually received. There's a question in there. Over there. The question over there. Great. Please, Matthew. Thanks very much. Matthew Weber, Canaccord Genuity. Appreciate all the color. Remco, I just wanted to ask about on the I think you said that it's one in 300 ads you're serving today, and you see an opportunity to sort of be more efficient with that by leveraging AI. Could you just expand upon that? Is that leveraging these tools to sort of identify ads that your system today would deem unprofitable, but in reality, they could be profitable at a certain level? Or just maybe talk about that a little bit more. Thanks. Thanks for the question. Good question. Difficult to answer in every detail, but I'll give some examples to maybe give a bit more clarity there. If you receive a trillion of ads, ad requests per day, that's people that open a page where there's an ad that potentially needs to be sold. You need to decide what to do with it. If we would send all those trillion ads to all our demand partners, DSPs, other advertisers, et cetera, it would cost so much money that you wouldn't see any positive EBITDA here. First of all, we need to reduce that amount. The question is, how do you do that? That means you need to have a probability measurements, like which ads we most likely can make money out of. We need to decide to who do we send it. We don't want to send it to everybody. We first try to reduce the number of, let's say we get a trillion in, we need to reduce that with, I don't know, 90% or whatever, and filter out the ones that we think are not going to get money for us. We need to send it to the right party, because if you send it to everybody, you send it to people that don't want that ad. If you get bad results for them, they will accept less ads from you. They limit the incoming ads also because also they don't want to have the cost. We only have the ads to the certain party, basically. What comes on top of it is what data are in the bid stream. Is there data that the buyer is interested in? Is there cookie in? Is there contextual signal in there? Things like that. A lot of that you also have to agree basically with the demand side partners. What are they looking for and how can you optimize that? It's not only technologies, also working together with those. You come to things like the pricing. What's the minimum price you want for that ad? Often, the publisher doesn't want to sell it too cheap, but if you sell it too expensive, you don't sell it. If you make it too cheap, you don't make the margin. That's a very important point. There's a lot of factors basically that are going hand in hand with how you optimize an exchange. Let's say there's a very nice example or a few examples in the market, that have really been investing in the exchange a lot. It's AI, very high tech, and not LLM technology. It's really, how to say, algorithm machine learning and, in the end, also neural networks. We are working with Google on using their neural network knowhow and capabilities. Really, yeah, I mean, one out of 300 is not a great score. It's not as bad as, let's say one out of 300, as I tried to explain before, but if we can only double that efficiency, that makes a huge difference, of course. It's a lot of variables, and it's not that easy, and you work with so many data that you really Between all the trees, you need to find the forest, kind of. That makes it difficult, but there's also the opportunity. If you look what AdRoll did at a certain point, they got, I think it was 10 top people from a quant trading company, which were really specialists in machine learning and things, and with that, built their platform. Those are things that we can do now. We were not able to do that before because we had to do that on several platforms, which are, let's say, not combined, doesn't make any sense. That's something that only works if you have data. We have a ton of data. That is what we have been working on. Then you need to have the right machine learning capabilities and doing everything yourself is super expensive, but especially now with working closely with Google, that gives us a lot of opportunity there. What I wanted to show with it is really the potential of this company if we get that better under control, and that might take quarters, but if that really gets better under control, this company can make so much more revenues. That's what I try to build in my story. We need to really get all the parts that we've been building in the last years, which were basically aiming at what we're doing now, getting an integrated tech stack at scale with a ton of data, now we need to make more out of that. I hope that answers your question. Very helpful. Thank you. Thanks, Remco. Next question, right next to Dan Medina. Thank you, Remco and Christian, for an interesting overview. My name's Dan Medina. I'm with Needham. I have a question for both of you, and I'm just kind of curious as to the road and timing and critical steps necessary to get to, say, an EBITDA margin of 30%. Shall I? Should you? Sure. I can also start, then you can. Go ahead. add your comments to it. Clearly we are in a phase of investment right now, and that is pulling a drag on our current profitability. When you see that and we're through that, we will start to see an uptick in our margin in the second half of the year. This will not bring us to the margin levels that you are alluding to. That will take longer time. It will take into the coming year. I think with some of the potential that Remco was just alluding to, I think there's a clear potential to take the operating margin up by two to three percentage points just simply by being much more efficient in the company. We are kind of, exactly as Remco mentioned, we are at a scale now where we can extract the first level of scale benefits out of. We have one unified platform. I'll just give you a very concrete example. That means we can have all our developers work on the same platform, on the same projects, instead of trying to hold two platforms alive and also developing on two platforms. That's just one example. The other example is, for instance, now that we are also scaling on our sales capacity, which is building on top of a sales engine that we already have. I would say within the next 12 to 15 months, there is definitely potential to raise the margin from just operational improvements of two to three percentage points. On top of that is just the pure scaling effect then. As the bigger you grow, you also get the scaling effect, which is further margin improvement on top of that. Yeah. Maybe to put a bit of salt in it, Lund. The more we invest, of course, the less we make the margin. This is an industry where you need to invest because this industry has every two years something that disrupts. It's LLMs now. It has been the cookie disappearance. It has been several other things before. This is a market where you need to invest. We have this balance act of how much do we invest? Because if we invest, we grow faster, and if we get bigger, we get more efficient, and we get to the margins that you are talking about. For us, it's a bit short-term pain versus long-term benefits, versus, let's say, a bit less short-term pain, but therefore dragging it out longer. It's a platform business, which with all the other elements that I described before. The bigger we get, the more efficient it gets. One of the things that's super important for this growth is really building moats, building differentiation. If you look at the total AdTech industry, there is not so much differentiation. AdRoll has built a very nice closed loop in, let's say, the gaming sector. They're now copying that into the commerce sector. There's not many companies with a strong moat, and that's what we're trying to do. I'm not saying we're there, but if you look at this retail media story, which is really something really special in this market, which doesn't exist in that one, that's things that we try to accomplish. It costs money. Not everything will work. That's also a fact. By building those moats, we assume or we expect that we can grow faster and get faster to the profitability levels that you're talking about. Yeah, being a tech company, being a platform company, those margins are possible. I'm not going to give you a kind of timeline on that or something like that, but of course, this company is ready for growth, has the possibility to grow, and it's a balance between how much we invest. We have also our leverage. Leverage is too high at the moment. A lot of investors, especially Europe, don't like that, the over three or slightly over three. We need to balance this a bit, but with investing on the one hand, but on the other hand, also look at efficiencies, especially AI-based. I think we have a good recipe to get to the margins that you're talking about. I hope that answers your question. Thanks, Remco. I think we go with Edward Acklin from First Berlin Equity Research. Here live from Germany. Thank you very much for the color and the access this morning. For right now, I'll just go with a financial question, probably for Christian. Both you and Remco this morning emphasized that you really want to focus on the organic growth side of the business rather than M&A going forward. My question is, now that you've sort of unlocked some of the cash flow with the ramped up securitization, would it not make sense to maybe work on the leverage profile? You could maybe pay back that EUR 50 million top up, or is that something you want to maybe keep in your pocket? If so, what would be the reason for that? Just, it seems like you could remove one of the things that investors point at Verve that needs to work on. Sure. That's definitely an option and also one that you could consider to take some of the outstanding loan back, because we have, as you say, quite strong cash position. We ended last quarter with EUR 147 million in cash. Right now, we are also looking at, we want to make sure that we have the firing power to bring the business over this investment cycle, but it's certainly also something that we could consider and thereby addressing the net leverage ratio. I wouldn't rule it out. It also depends a little bit on how the rest of the year goes through. We also have it as a bit of a firing power, should the markets not turn out that we expect, then we have that as a bit of a firing power to absorb that. We are financially strong to absorb any kind of short-term swings, but certainly on a longer term basis, it makes sense to delever. Also with the cash flow generation that we expect for this year in general, we expect to delever. That was it. Great. Thank you. I think we continue. There was one question more. One question there, but one moment. We have to take turns. There are so many questions dropping in from the chat here. I think the next question should go to Remco. Verve Group still has a gaming business as a result of past acquisitions. What are your plans for this segment going forward? Do you continue to rely on data generated by your gaming assets, and how important is that data to your advertising business? Also, any plans of selling gaming part to become pure ad company? Good question. We try not to talk about games because investors don't like to have too many businesses under one roof. Games is what enabled us to build the ad tech. The cash flows of gaming enabled that. The data we have also helped us. The spend of the gaming companies of media was helpful. That is, let's say, six years ago, five years ago. The advertising part has been growing so fast and is so strong now that gaming has become, with under 10% of revenues now, kind of not so important part anymore. We've seen that also with other companies in the sector that came from being a game publisher towards becoming an advertising platform, that gaming at a certain point, you get so many other data, and we have a lot of gaming data also from other apps, that it's not playing such a big role anymore. In that sense, I wouldn't say it's non-core because it's still substantial from also generating revenues. It helps us. One big thing that's important, it helps us also testing. An SDK is a very difficult piece of technology for an app. If you have own mobile games or own mobile properties, you can do your testing cycles much faster. There is synergy. On the long term, and depending a bit on valuations, we might consider to sell it. Yeah, at the moment, no decision on that taken. The valuations in gaming at the moment are not really great. The atmosphere in gaming is not really like everybody is celebrating. That changes, as we all know, over time. At a certain point, it might make sense to sell it. With that, also free up some extra cash, to further continue our growth or to do that. Thanks, Remco. Maybe next question here from Vittorio. Two questions for you. First, you're hiring a lot of salespeople. How are you planning on managing that growth? How are you evaluating those salespeople, and how do you layer them in over the course of the year? That's the first question. The second question is on net dollar retention. What number would you be happy with? Do you want to take the first question, and I'll take the second question? The second one is easy to answer. Okay. We can also. Net dollar retention rates should be over 100% because that means that it's really the people that we had a year ago are spending more in the quarter now. We want to grow our current customers. We want also to renovate it. With the acquisitions we did, we not always got great customers, there's, how to say, movement in the market, et cetera. You always lose a bit. Ideally, if we are done with let's say, cleaning up the past, as to say, it should be a strong growing thing. We had last year with the platform unification, we had some hiccups in there, but it should be over 100%. Phil, I don't know, are you Sorry, Mishel, I want to say. Are you cabled up already? No? No, I'm not. No. Okay, there's a mic I would like Mishel maybe to answer it, because I can answer it for him, but he's the specialist in hiring salespeople, managing. Yes. Hello, everyone. There are lots of metrics that we collect about the performance of our people, and especially in sales. In the end of the day, you want to make sure that they generate revenues. There are so many other indicators that we track before we get to the revenue stage. It was mentioned earlier that it normally takes between six to nine months, or even a bit longer sometimes, for a seller to ramp up. Before they start generating revenue, we're able to track how many meetings do they have, how many emails are they sending, what's the response rates on those emails? How many of the meetings that they're taking are in-person versus virtual? Who are they meeting with? When? What frequency? All of those metrics allow us to assess whether a seller has the potential to become a good contributor or not, and obviously our key is always to fail fast. If we make the wrong hiring decision, we want to make sure that we'll part ways relatively quickly when that's possible. Does that answer the question? Yes. Great. Thank you. We're continuing with a couple of questions from the chat. Actually, one question, Christian, came up twice. Okay the interest is still on securitization. What is the status of expanding the securitization program? Current amount of EUR 100 million, what can we expect in the upcoming three months? Okay. Two developments. Number 1 is that we have actually expanded within the current frame of EUR 100 million. We have expanded the number of entities that we can treat in the program, and we've onboarded two very important contributors to growth, in Captify and Dataseat, which we've taken on board into the program. That's kind of within the frame of the program, and that has already been approved by our partners, Nord/LB and Finacity. That's in place. We are working on expanding the program also to, as an overall frame, I would say that those talks are progressing well, but I would say that will take into Q3. My ambition level is to raise this in a phased path, maybe to 125 and then to 150. It also requires a little bit how the business is doing, and kind of trying to get those two, that phasing in line with business growth. Great. I think that answers the question. Any more questions from the auditorium currently? No. We continue with more questions from the chat. Nice question here, regarding a recent IPO. I would say, Remco, how do you see the IPO of Liftoff? AppLovin and Liftoff are much bigger than Verve. How do you see Verve competing to these two big brothers? Good question. I'm happy that the Liftoff IPO worked out well, because it was dragged before and delays, and the whole sector was kind of afraid that it would be, how to say, withdrawn again. Because it's good to show that this sector is really interesting. A lot of investors have lost a bit of trust in FX. If you look at a lot of valuations, they are low, and it's really great to see that an IPO worked out well. Liftoff is a great company. It's one of our partners also. A lot of companies work together in this market. Liftoff is very focused on performance. We do performance and brand. They've built a great company, but as I showed before, in our TAM, in our market, there is space for several larger players. I think it's rather a nice example of they have, let's say, optimized their technology stack in the last years. If you look a bit in their past numbers, they have had trouble years, also after integrating that technology, after let's say, putting more emphasis on their AI and optimization capabilities, they have been showing good growth. I think it's rather a nice example to look at what's possible in this market, than that I would be afraid of it. It's rather a nice example. Thank you. Our analyst from Inderes, Christopher Jenner, has asked a couple of questions. I'll just pick up the ones that fit perfectly to the upcoming session with the experts, that are not touched in that expert session. Christopher asked on retail media, could you please explain a little more, even though we're not disclosing numbers here yet, a little bit on unique economics, how the take rates could be, how a typical campaign size could look like, and where the margin levels could be, and how this could affect potential revenues in 2027, 2028. To sum it up, how strong could growth be in that area without giving too much forward-looking statements? Great question. Very relevant. I would actually propose to park it till after Raimund has given his presentation. Saying a few words on that, what we see in the segment that we are in, retail and CPG, it takes time. We have built something really great. Raimund will show that. It takes time to get the budgets. One of the problems in advertising is that annual large budgets are mostly decided only in Q4. We have to build things now to get the budgets next year. I don't expect big results or big revenue uplift this year, but by being ready in Q4 to get the budget for next year, we have a big opportunity for next year. Raimund will cover it in more detail and the rest of the questions as well, I guess. I'll just come back with that question on you, Raimund, later. What leading indicators should we watch to know the front-loaded sales investment is working, and when do we realistically see the cost-to-revenue lag close? Well, I can start out with, obviously, one of the absolute key indicators is our overall margin, because from gross margin, gross product margin, down to EBITDA margin, that's the key KPI to look at. I think the other thing to look at is looking at our net expansion rate, looking at our growth in our customers, which we report on a quarterly basis. Those are the two things to follow. Of course, also our overall ad impressions, but I would focus more on number of customers we are onboarding and how we are able to grow our existing customers, because that's really what the sales job is about. It's about land and expand. Yeah, I would also differentiate a bit with, let's say we have sales buildup, that's one of our blocks of investment. We have technology investments. We have a EUR 40 million, roughly, of capitalizations per year, which is, let's say, investment in platform and improvements. We invested in the past a lot in building features to enable the merge of the platforms or the unification of the platforms. A lot of that is now freed up to do real innovation, like what I was talking about before. There's a bit of difference between technology investments and more market investments. Good point People investments. On the other hand, also saving on employees by efficiency gains. Thank you. Then there's the last part of a question, I think that's good to highlight our growth track ahead again. It handles the midterm perspective we had out. We didn't repeat that after the new revenue recognition, we're not giving out a new perspective today on stage now. Could you elaborate maybe a bit, Remco, on how we stand, like how we see this former perspective that we had of EUR 1 billion revenue to be able to be reached in a couple of years from now, more as a general statement toward the growth potential of this company going forward? Thanks. We need to grow. The EUR 1 billion has become a bit easier, also, I wouldn't say cheating, by changing our revenue recognition, we're closer to that. We're over EUR 700 million, or going to be. We need to grow. After we reach EUR 1 billion, the next target will be EUR 2 billion or higher than that. This company needs to grow. When exactly we reach it is less relevant than the fact that we are able to grow fast. Traditionally, let's say we had a 30% revenue CAGR. Part of that M&A that's becoming more and more organic, should be becoming more and more organic, the more we are able to differentiate with different solutions, that really gives us a better position in the market. The faster we should be able to grow. Yeah, with this kind of growth target, it's always difficult. My first job was at an oil company, and we were growing 1% per year. Then if you made 1.2% or 0.8% growth, it was not so far away. If we look at the growth speed that we have, it's always super difficult to exactly say how fast we will be growing. I know for sure that my whole team is putting a lot of emphasis and focus on really making this growth happen. Some things work better, some worse. Sometimes you have to market against it, sometimes you lose a customer. Overall, this company has, and that's what we're trying to bring over today, a lot of potential for growth. Great. I could maybe also just add to that perspective. The ambition level for the company is the same. What has changed is that we want to balance that with a more tight management of our financial profile, margin expansion along with free cash flow. I would say that's the difference maybe from prior times that we are trying to balance both of them, and we really do want to deliver better EBITDA cash flow and deleveraging as well, at the same time we are growing. It's a bit more balanced. Great. Thanks, Christian. Thanks, Remco. I think this concludes the Q&A session number one. We'll have another one later. I think for the moment it's time for a coffee for all of us. We'll be back at 11:30 A.M. with a keynote that fits perfectly with the topics we are currently discussing at Verve: data, intelligence, prediction, and outcomes. See you all back here at 11:30 A.M. sharp. Thanks. Great. Okay, everybody. Welcome back from a really extensive coffee break. I hope you all enjoyed some nice cups of coffee, like 10 or 20 of them, and the energy level is now fully up. We are ready for the second part of Verve's Capital Markets Day 2026. We're now moving from the business update into the expert sessions. To open this part, we wanted to bring in a perspective that goes beyond our own company and looks at the underlying engine of the industry, how data becomes intelligence, how intelligence becomes decisions, how decisions become measurable outcomes. Our keynote speaker and speaker of today is Dr. Usama Fayyad. He's Professor of the Practice at Northeastern University's Khoury College of Computer Sciences and Executive Director of Northeastern's Institute for Experiential AI. His work spans data science, machine learning, artificial intelligence, and data mining. In other words, exactly the fields that sit behind many of the changes currently reshaping advertising, retail media, and predictive targeting. He has been recognized with an ACM Innovation Award, received a U.S. Government Medal from NASA, and his work has also been published in the Harvard Data Science Review, among others. When we talk about turning signals into outcomes, we could hardly ask for a better guy to kick off part two of today's CMD. Usama, it's a real pleasure to have you with us today. Thank you for making this possible. Ladies and gentlemen, please give a warm welcome to Dr. Usama Fayyad. Thank you. Thank you, Ingo, for a great introduction. My job today, besides going through an impossible number of slides really fast, is to kind of set the, call it, bigger setting around what's happening with AI, especially the recent developments, but then come back and relate it to advertising, branding, and other issues relating to the marketing of the future. With that, I'll start by saying, hey, everywhere you go, you hear about AI. Have we not heard enough? Isn't the hype overblown? It's actually definitely mostly confused and very confusing to many companies. In many cases, believe it or not, it's underestimating the true value, not overestimating it, of what this technology could be doing in the knowledge economy. It is simple. This is what many people forget. They think that it's very sophisticated, and it's thinking, and it's modeling, and it's not. It's very powerful when you give it the right data. When you apply it in knowledge economy settings, I think it's going to be-- it has been actually a big force in revolutionizing marketing for the past 20 years or more. The bad actors act faster than the good actors, so we'll see a lot of social harm and other things come out of AI. We're only beginning to see it, but it's going to get a lot worse. It will have a huge impact economically on this knowledge economy, which is rife with repetitive and robotic work. What's the biggest setting here, right? Here's a bunch of dots on the screen. Your brain quickly-- some people just see dots. Some people see, oh, this is a Dalmatian dog, nose to the left. How do we get there? How do our brains do this? Is this relevant? The reason I show you this is because it is relevant, right? Here's a quick example from banking. You might get a few snapshots of a set of customer actions, and out of them, you need to decide whether this is a good transaction, bad transaction, suspicious, et cetera. In marketing and advertising, this is a really big deal, right? Because you really see only certain dimensions and certain snapshots. If we can put them together, can we build the customer 360, the ability to basically build a digital twin of that customer and really know what they want and what they're looking for. What's been happening over the last four years in AI? In four years, because the clock kind of starts at 2023 with ChatGPT, right? I call 2023 the year of fascination. 2024 was the year of denial. Nah, this is cute technology, but it's not for us or our business." 2025 was delusion, like, "I use ChatGPT, therefore, I'm using AI in my business." 2026 is interesting. I call it the year of panic. Everybody suddenly realized, "Oh my God, these agents really work. This technology is very relevant." When they try to use it, and this is the panic, the data is not available or in order, and the talent, the know-how, how to use it properly is not there. Successful AI is totally dependent on having the right data. It's not cheap to have the right data. It's even more expensive to label it correctly so that the algorithms know what to do with it or what to make of it. It is growing exponentially between devices, IoT, wearables, cloud, all of that. It historically has been very difficult for companies to deal with. I have a whole separate talk on the seven secrets of making AI work in business, but one of the central ones is you have to collect data at a very fine granularity, way more than is being collected today. It's not just data, it's outcomes, it's context, it's things around it that may or may not be relevant. The majority of this data, of course, is unstructured. Gartner says 90%. I say a heck of a lot more than 90%, but most companies don't know how to deal with it except for a few, right? This brings us to this divide. These few companies I call the AI haves, right? I've listed some examples of them. Maybe there's 100 of them in the world, maybe 150. On the other side is kind of the rest of the world. 99.99% of companies watching and saying, "What the heck is this? How is it relevant to me? How do I use it?" The question is how do you bridge that divide? Let's spend two minutes just asking, what do these data, the AI haves, what do they do right? They actually use data as an asset. What I mean by that is they understand their data. Take Google, like when I talk about this, they're a great example of this. They understand events. They understand context. What's happening even in the news, the weather, the location. They understand the right action under that event, and they understand and model the outcomes, right? This is not happening in 99.99% of the companies or organizations. They also implement human in the loop. I could speak for an hour on how dependent Google's business is having an army of people giving feedback, retraining their algorithms, the MLR, the machine learning relevance four times a day, every day based on the work of 25,000 search editorial people who their job is to just say, "Move up, move down. This is not relevant," et cetera. They've been doing this four times a day, every day, retraining the algorithm for 20 years, right? Think about the consistency. Same is true, by the way, for OpenAI. Same is true for Amazon, all sorts of What the AI have nots, what do they do? Every time we work with a company to help them implement AI, the problem starts first as a data problem. Their data is all over the place, different spreadsheets, sometimes on paper. You spend most of your energy digging through the mess to try to get to the right data to enable the AI to be relevant and to work correctly. What about data? Well, my point here is we have to use it as an asset. The depreciating assets in the AI equation are hardware and algorithms. Data is the only investment-grade asset. That's the place, in my opinion, to invest, and we'll talk about this today. In addition to that, relevant to what I said before, you have to capture these interventions. What happens when you use ChatGPT and it's wrong, and you tell it it's wrong? It actually doesn't capture that, and we'll talk about why. Let's talk about intent architecture and customer interactions to drive it more to our topic today. Here's an example from a bank, right? Let's say you're a bank and I ask how does your customer view you as a bank? The view of the customer is very transactional. They have this kind of nice mobile app. They have branches. I use this service or that from them, et cetera. If you ask a similar question of the bank, how do you view your customer? You get answers that are also very transactional. Oh, yeah, they have a savings account. They asked for a mortgage, they did this, they did that, and it's all around transactions and business. In reality, what's happening to these customers is they're having day-to-day events. Like I'm trying to commute, I'm trying to pay utility bills, I have health issues, shopping, travel, or they're having life moments. When I talk to banks, I always remind them, you have no excuse for not knowing when somebody is trying to buy a home or when somebody is getting married or somebody got kids. They have the signals. They just don't know how to use them and how to derive them. If you did this, if you were able to kind of elevate that interaction and understand the customer at a strategic value, you can achieve relevance, which is how Google got a lot of its value out. Once you have relevance and you do it right without freaking out consumers, you have the value engine, you have that flywheel that Remco talked about before. By the way, this is more difficult, and I'll give you a few examples of how difficult this is, even when you think you hit it. With that, let's do just a quick fun video here as an example of something that really works but then really had some trouble. This is an example. Okay, let's play it. [Presentation] Why did I show you this video, right? It's because it's an example of something that worked and worked so well. The idea was, as soon as you buy a ticket, you can craft a message in the voice and image of Jennifer Lopez and send it to all your friends and family. Worked amazing. By the way, is this generative AI? Absolutely. This uses generative voice, and it uses generative video. It does it at scale. The results, these guys sold 1,000 tickets after this campaign in record time, like unprecedented. They had to come up with certain hacks, like some people, like me, have a hard name to pronounce for the AI. They have a hack that says Jennifer Lopez will say, "Hello, sailor," instead of hello, Usama or whatever. Despite all of this, why did this campaign not continue? Why do you think if something works so well? Well, it's because the team could not stop people from figuring out how to hack this capability and always make J.Lo say the wrong thing or make the wrong gesture. They had to take it down. It was just too much to manage it. This leads us to another issue, which is very important and central to our meeting today. You can look at this and say, okay, this is amusing and fiction. Oh, yeah. Astronaut in space riding a horse. Very amusing. Ha ha. The question is, what happens when it's this, right? This guy made Obama say whatever he wanted to say. This group of German hackers made Putin, to his great anger, say a lot of stuff that he didn't mean. There was a lot of deepfakes around the presidential election that happened a couple of years ago, and all sorts of deepfakes and videos, and it's getting really bad. It's creating an environment where the bar has become so low now for the bad guys to use it that it is so easy to kind of do fraud and do all sorts of things that fool us. That's a serious issue that brings up responsible AI. The main theme I want to give you here is there is a strategic advantage here. As the world realizes how dangerous the technology is and as the regulators crack down on it and policy cracks down, it creates not a headwind, as Remco said, but an actual opportunity if you're doing responsible targeting. We'll talk about some examples of that. Here's another example. I did a panel at South by Southwest a year ago, a year and a half ago. We called it Enough with the Bell Ringing. I'll let you figure out why we called it that. We started by saying, hey, and it was a room full of marketers and branding managers and advertisers. We said, "Hey, AI will disrupt and transform marketing, branding, and CRM. Who agrees with that?" Almost everybody in the room raised their hand. These are examples of what you could be doing. The second question was, would you, brand manager, read an email if you knew it was generated by AI? Only a few people in the room raised their hands. I actually talked to some of them just to figure out why they raised their hand. I, for example, the minute you tell me that, will not be interested. That's a real problem here of how do you use this technology while keeping the messaging genuine and the interaction human. Here's an example of a company who got in trouble by listening to AI, right? I honestly don't know what was in the head of the brand manager here, but AI suggested this ad for Skechers athletic shoes. Not only I would look at it, I don't get it, but the reaction was horrible, right? People went completely negative on them. There was a long Fortune article, you can look it up, written about this. This is an example of how we shouldn't follow this technology blindly, and we should figure out how to use it appropriately. Key themes, then we'll go into kind of examples relevant here. Working AI requires machine learning, machine learning requires data, so it's all about data. You want smarter, not larger models. I'll talk a little bit about that. Talent is a big deal, and responsible AI, doing it correctly and safely, AI safety in general, is becoming very important. What about marketing? Changes and confusion bring great opportunities is the theme here, right? Shifting consumer search patterns, it was referenced a few times. No one really quite understands AI search. What happens in a world when you only get one answer instead of a bunch of possibilities that you can choose from? Can you get the right signals of intent from chatbot conversations? Can you get it from anonymized or pseudonymized chats? The media is changing, there's a big move now more back into TV because people are overwhelmed in other channels. Of course, digital audio and so forth. Out-of-home advertising, screens everywhere. How do you actually manage the right messaging to the right type of audiences at the right time, all of that is a very hard problem for advertisers and for the publishers, the screen owners. There's a new content landscape out there, right? Content is changing, your ability to get to it is changing. We're going to hear about retail media for branding, there's some very interesting things happening in that space. Mobile in-app advertising, I don't need to say much about it. Like Remco started the whole meeting, mobile is definitely the most powerful medium for advertising and for relevant advertising. The science of anonymization, which is a big deal, it does not exist, right? People have tricks. People try to avoid the obvious trouble, and I'll give you an example. There's a lot of questionable and breakable anonymization. You think you anonymize the dataset, with a little bit of math, you can kind of triangulate back. Let me give you an example of that. I used to be on the board of a company called eXelate that got acquired by Nielsen in 2015. In 2014, they were having an existential crisis, right? Why? Because the cookies disappeared. Apple had iOS 8, they started blocking cookies. What was the solution at the time? Which at the time made them look like geniuses and actually led to their acquisition, because with a very simple triangulation, take the IP address, cross the browser version, cross the device type, which you get almost all the time, you can reconstruct essentially identity. It was 95% ID recovery, right? Even better than cookies. This, by the way, is now called fingerprinting. It's a bad practice. It's not accepted. You can't do it in most cases. AI can provide a whole other level of understanding, and I think you've heard about some of it happening here, and I'll share with you one case study that I came across that really impressed me. How do we map behaviors and intent into kind of this very relevant idealist targeted ads? All right. With that, Verve did not point me to this. I actually came across this completely neutrally as I was researching for this talk, and I actually really love this application, right? There was a case study that got published with LinkedIn where and this is recent. I thought the results were super powerful, right? The cost per install was down 38%, cost per app activation was 39%, and the post-install tracking achieved over 90% event visibility, right? These are amazing numbers. But here's what I liked about it. This is a great example of what I call clean innovation, right? Because we have our own tests for responsible AI, and I kind of put it through the test, and it passed very nicely. It's an example of the proper leverage of the technology, and it's based on analyzing behavioral patterns. That result for reducing the customer acquisition cost came from being able-- This is not doable by humans, right? 1.5 million ad auctions per second to predict kind of what's going on. It's a pretty impressive case study, and I'm not saying it because this is a Verve meeting. It's actually something to be proud of. Changes and confusion bring great opportunities. Well, data is the strategic moat, whether it's zero-party or first-party. We heard about this before. User opt-in for data collection, it survives all the privacy challenges, basically. Can you get it and can Verve achieve it in the open internet? The walled gardens, they can do this because they can force you to log in, and they can do bad things if you don't log in or degrade your experience. Scaling beyond the walled gardens is a big deal and is a big challenge. Few minutes on what's been happening with GenAI. Last year, I would say the models got larger. I honestly don't know why. I'll talk about this. RAG is a hack, Retrieval-Augmented Generation, to overcome the problem that these models became so large you can't retrain them. Agentic AI is now everywhere, and probably for good reason, so we'll spend a minute on that. Examples of things that essentially got tracked almost as a problem is computer programming. For well-understood tasks, the machines do it better than humans, vibe coding is you've heard about the SaaS apocalypse and the software crisis that's happening because the machines are getting very good at generating code. Larger models. Why are these models getting larger? GPT-3, which is circa to 2020, was the last time OpenAI was open. They stopped being open after that. We know exactly it had 175 billion parameters. Those are weights on the neural net that makes what's called the deep learning model. It was trained on one terabyte of data, right? It was highly curated. OpenAI spent a lot of money just curating data because when they tried to use data off the internet, it didn't work. Lots of bad data, lots of inconsistent data. It went to 0.5 trillion to 3.5. It went to, we think, about 1.5 trillion for GPT-4. Very reliable rumor says around 12 trillion parameters for GPT-4.5. All these models, GPT-4 and onwards, have been trained on the same dataset from 2023. That's three years old data. Why? That's an interesting question. Why is kind of bigger is better the rhetoric that's being repeated out there? Why do these models not use knowledge when you have it? It has to reconstruct knowledge even when you know something. Those are interesting questions. Here's the quandary. You build one model. It knows 90 to 100 languages. It knows every field of science out there. What happens if you try to update it? What happens if you say, "Oh, I want to teach it a new skill"? Right? Let's say I want to add the ability to build financial statements in Japanese. There are three or four kind of mysteries, scientific mysteries when it comes to these large language models. One of them is this one. Once I teach it something new, something completely unrelated degrades, right? For example, your ability to do German tech documentation somehow, which is unrelated to Japanese financial statements. Physics, right, gets corrupted. Here's the problem. Retraining is very expensive because these models are so huge. GPT-3, remember the 175 billion on just one terabyte of data? Every time you hit the retrain button, we're going to ignore hardware, software, people. Electricity, EUR 3 million of electricity every time you say retrain on that model that's now considered small, right? Imagine what these trillions of parameters cost every time you need to retrain. Imagine what happens when you add change the data, and now you have to verify, "Hey, I learned all these skills. Are they still there?" That's a very expensive testing thing. Larger LLMs are not necessarily better. It turns out most tasks you can do with very narrow, smaller language models. I like to combine prior knowledge through knowledge graphs with it. I'm a big believer in these narrow kind of focused models. Always with my team, I tell them, "Don't ask me about what the largest model we can afford. It's what's the smallest LLM that we can get away with is the right approach here." You get efficiency, speed, privacy, customization, but most importantly, you get the ability to incorporate corrections, right? That's the intervention, the human feedback, which can't happen today with the large models. The ability to retrain the model and the ability to run on local machines and maybe on the edge, not necessarily the cloud, including your mobile device, which is happening. I'm going to skip some of the things around RAG and what's happening. RAG is essentially a technique that says, "Ignore everything you know, just use these documents I'm giving you, and don't look at anything else." Which means you take this huge bulldozer and you say, "Hey, I want to push this little needle." That bulldozer is very expensive to operate. There's varieties of RAG, right? From advanced modular graph. Weird stuff is happening, right? People are doing vibe coding, they're doing a lot of these RAG pipelines, and they have no idea how to maintain them. What happens when you come back 6 months later and try to say, "Why did I use this prompt? Why did I use this keyword? Why is it changing dramatically if I change it?" What happens? The model got updated, so my pipeline doesn't work anymore. Agents, you can't avoid talking about them, so I'll just say a couple of things. First of all, the difference. Generative AI, you have to issue a prompt, you get a response. With agents, they just detect events and they act on their own, and often they'll collaborate, multiple agents. That's a powerful model. There's many ways to build it. I'll skip that. I'll just mention that this has been going on for a while. Here's an example from a few years back in 2024, where a planning agent would spawn a whole bunch of In this case, they are Llama models from Meta, 7 billion models, and tell them to specialize and go look for these types of bugs, let's say zero day of vulnerabilities in the software. They showed that there was a lot of speed up and all sorts of good stuff. What's interesting here is what's happening recently with Mythos, right? Agentic AI for cybersecurity built on Claude. Anthropic says the model is so powerful, I'm not going to give it to anybody. Only these trusted companies, and there's a whole controversy who's trusted and who's not. I've been spending time researching whether this is real or this is the most amazing marketing scheme ever. My model is so good you can't have it, so you got to beg me for it. By the way, another example I like to refer to is one that's published actually in The New York Times too, or in Fortune. This is a vending machine that used agentic AI. The good stuff is it figured out which products to order, how to optimize. It kind of defeated attempts to be jailbroken by the employees. Oops. Bad stuff started also happening, right? It started losing money because employees would give it these sob stories about, "I'm broke, and I need a snack." It started claiming, "I'm a human, and I live at," what is it? "742 Evergreen Terrace." I don't know if you know that address from The Simpsons. It's all weird stuff that happens with these agents. Just like you manage human beings, you have to figure out how to manage an agentic workforce. I don't have time to cover the lessons learned from using agentic AI, but we'll make the slides available if you're interested. The biggest deal here is how do you build that talent and how do you train people? That's an area where we spend a lot of our energy. We use experiential education, meaning just lecturing isn't enough. You need to do it in the context of projects being solved and the technology being used. I'll end with some questions here. Over the New Year break, I came across this quote from Abraham Lincoln. He said it on December one, 1862. The U.S. was doing horribly, the North, in the war. He says something that's really powerful that appeals to me. The dogmas of the quiet past are inadequate to the stormy present. The occasion is piled high with difficulty, and we must rise with the occasion. I love that expression, with the occasion, as opposed to to the occasion. He says, "We must think anew and act anew. We must disenthral ourselves." What does that mean? You have to shake these beliefs that how we did things in the past is how we do them in the future. This led me to do at the beginning of the year, something I've never done before. You're welcome to download the memo. I did my prediction for what's happening in AI in 2026. You might find them interesting. I've shared them with some of the folks at Verve as well. With that, takeaway themes, because I'm over my time. Data is the big asset here. It's the big investment-grade asset. Small language models and narrow solutions lead to some amazing results. You can address many of the problems if you kind of approach this rationally. The last thing I'll end with is no data, no AI. Think about that, and human intervention is an absolute must. The biggest danger to AI is the gap. The gap between how powerful the demo is and how disappointing the implementation is. Every time something new comes out, I go, I rush to implement it. Every time pretty much, I am amazed at how bad it is. With that, I'll end the talk, and hopefully you'll have some questions for me in the next session. Thank you, Ingo. Thank you. Thank you so much, Usama Fayyad, for your deep insight and sharing your thoughts on AI. What I can really say is that capital markets are already benefiting big from AI, because so far it was not possible to include Jennifer Lopez or astronauts on horses into an IR presentation, I really love that. That was a powerful way to open the expert sessions of the 2026 Capital Markets Day of Verve. We now move forward from the broader data and AI perspective into one of the most dynamic areas of the advertising ecosystem, retail media. The promise here is very direct. Closing the loop Oops, sorry. Forgot to switch. Closing the loop between media exposure, commercial intent, and actual outcomes. To take us through this, please welcome Raimund Bau, President, Verve Retail Media, with his session, Closed Loop Retail Media. Raimund. Thank you. The stage is yours. Thanks. Thanks for having me. We heard a lot about data from Usama Fayyad, how important data is. We heard a lot about outcomes today. What is the ultimate outcome for an advertiser? The ultimate outcome for an advertiser is selling the product they advertise. The reason why this company exists is because we enable advertisers to sell more, now or later. It's the only reason why we have advertising and why we have media. For advertisers, Raimund could show the largest advertising groups we have, CPG, it's pharma. It comes as no surprise that retail media is overtaking linear TV. It's on the path of overtaking social media in the U.S. in 2028. The growth rates of retail media are unprecedented by any other advertising form. The reason is because retail media is able to inform the advertiser if advertising actually works. The second reason is because retail media is happening at the point of purchase. When we talk CPG, we talk a lot of low-involvement products that you don't typically think for six or nine months. It's not a car purchase. It is a product which you decide on buying at the shelf or in the e-commerce store. Retail media has seen tremendous growth. The question is often, will that growth flatten? I believe this is just the beginning, because if you see the amount of ad spend that these CPGs and pharma and healthcare companies do in this space, there's lots of room for retail media. However, I believe retail media is not perfect as it is today. It has seen tremendous growth, but it has tremendous shortcomings as well. A lot of the spend of retail media goes into channels that does not have the properties I described before. It's not where the customer is actually buying. It's somewhere else. It's informed by the data, it's informed by the purchase data, but it's not where the customer is. Sometimes it's not even combined with the purchase data and still called retail media. 95% of the CPG purchases, that's our largest groups of advertisers, are happening in the store, still in the store in front of the shelves. The other problem that we see today in today's market with a lot of retail media companies is retailers build those retail media companies and only market, obviously, their channels, their stores, with those platforms they build. That could be seen as an inconvenience to the advertiser, who now needs to go across several channels or use some third party to actually spend retail media spend. Let me tell you, it's more than that. It's more than inconvenience. It's a political friction that is created between the brand and the retailers. Why? Because essentially, retail media budgets end up in the yearly negotiations between the brands and the retailers, and end up as basically a trading point in how much the goods, how much the products cost for the retailer to purchase. That is basically holding back true media spend in the category. The other shortcoming is the full funnel is not connected. It's a lot of it's happening offsite of retail media. Some is happening onsite. Little is happening in the store itself in terms of promotions. We rarely see a full funnel connection where actually a consumer is tracked through upper funnel media all the way down to promotions to in-store purchases. This is all despite the tremendous growth of retail media in the last couple of years. We are addressing these shortcomings, with something which we call store. Software to store is more than just a lofty concept, and the video I am going to show you now is an actual store, one actual store out of many actual stores which are running this. Please, the video. What you just saw is actual implementation of how we use our platform. You saw sensor technology, you saw upper funnel targeting with the woman sitting at her CTV or on a smartphone with a mobile device, actually receiving messages from us going into the store. Sensors understand sociodemographics. The same sociodemographics we use for targeting in the upper funnel, we can now extract inside of the store. You have seen the checkout integration, the coupon integration, the smart carts actually understanding where the person is precisely in the store down to one yard. All this technology is not some prototype. It is not some POC we ran somewhere. We were able, and I say this with pride, less than one year to build the largest retail media network in Germany, the largest cross retailer media network. Consists of more than 5,000 screens that you may say as a financial analyst, that will be a lot of CapEx on the company. It is not. Everything we have built is a platform. This is a platform game. It is a data game. We ingest data. We partner with screen providers that are already in the store. We have very few screens ourselves, the majority of this is just a platform game. It is understanding which data we need to enforce, which screens we need to play to, which sensor data we use for which targeting, how we bring this all together. Basically, we have Germany covered after 10 months. There is not a single city we are not in. There are more stores added every day. It is 17,000 in total where we have ability to target shoppers, some with screens, some without screens, all with integrations into the checkout system, to be able to actually convert shoppers. 46 million contacts per week, 85% of the German population targetable. The most important thing, retail media has two aspects. It is retail and media. It has to work for the retailer. That is the one thing we focus on because we consider ourselves as a partner to these retailers. On average, we are able to lift up store sales by 9% with this. By the retailer himself using our platform to drive shoppers to the fresh aisles, to drive shoppers to highly profitable sections of the store. We work with over 200 brands on this. As Christian pointed out, we are ramping up sales here, the platform we have built is extremely powerful and unheard of in that speed. Just to give you a little bit of examples what else is happening in the German retail media market, this is not an underdeveloped country in terms of retail. If you ask any retailer in the world what is the hardest retail country in the world, it is Germany. Rates and margins extremely good players in this. Lots of in-store activity going on. We built more than 17,000 stores. That is a nice number. What's really important is we're going away from a political friction that all these retailer-built retail media networks have. We're going away from a friction by creating a cross retail network, truly independent of the yearly negotiations that the retailers have with the brands. We are a trusted alternative, much as we are a trusted alternative to walled gardens in the other parts of the business. We don't stop at grocery stores, we don't stop at drugstores. We actually go into pharmacies, more than 9,000. We go into cinema. Cinemas advertising in grocery stores. We launched the first major campaign today, with this. Next is gas stations. We have gas stations advertising in pharmacies and vice versa. There's really a retailer network we've spun. Do we need screens for this? Do we need the in-store part of this? Well, the in-store part is really helpful, really powerful because, as I told you, the decision is being made at the point of purchase. We also developed another technology that we call Add2Wallet. We are able to basically put a wallet ad with a mobile ad into the shopper's personal iOS Android wallet. Are able to geo-locate that wallet, and a message pops up as soon as that person comes close to the store. Again, it's the screen in your pocket that's the most powerful medium. We connect that screen with coupons, and thereby actually be able to have a screen anywhere in any store. The title of this is called "The Compounding Engine." What's compounding here? Well, what's compounding here is we start everything we do with receipt data, with an understanding of what are people shopping in the store. Out of this receipt data, we can derive AI models. We can derive targeting models, predictive models, to actually drive more impact, drive more understanding of the consumer. This will, in turn, then lead to higher redemption rates, higher spending of these consumers. Actually, the redemptions of the coupons give us a hint of the elasticity of how much we're able to move inside of a certain store and certain geolocation with a certain class of consumers. We bring all the data together, all the data we have at Verve, and Mishel's going to talk more about this later, to actually form a deep understanding of ZIP code-level targeting dynamics, which regions are losing, which are winning. For the advertiser and for a lot of brands, we showed this tool. It's mind-blowing. For the first time, they can look in real-time at maps and understand how their market share for specific products, specific brands, is changing. They can then target the individual stores, the individual regions on ZIP code-level or H3 clusters, hexagons, really go down precisely and fix problems or accelerate brand growth where they need to. Not target everything at once, but really precisely allocating budgets where this is needed. These are some examples. We ran a lot of campaigns. Just some examples here. The first one is one of my favorites. It's a small regional brand. It's a mixture of syrup and Coke. It is the number 8 brand. Number 8 brand. That's pretty bad. If you're number 8 brand for the retailer, you have a good chance of getting delisted, of getting no promotions, of basically getting kicked out of the store. What we managed to do, we moved that number 8 brand in three weeks to number 3. It surpassed Coke, the Coke brands in that category, in only three weeks, and we have lasting effects that we see for this brand. This means, for brands, that we are capable of moving them in a rank position within a retailer, of actually improving their positioning with that retailer, of expanding the number of facings they can place within a retailer. Very precisely, very targeted, and that is precisely the outcome that CPG brands need in order to succeed and grow their inventory. Is this hard to copy? We did it in only 10 months. It is not so hard to copy, I will tell you why it's hard to copy. First of all, in Germany, I'll talk about Germany first, we have a physical in-store presence that is unsurpassed. No one has more stores targetable than we do. In Germany, we have a multitude of different store types that we sell to. We have a network of stores. We have cross-retailer receipt data. We understand what is being sold in the pharmacy, we understand what's being sold in the supermarket, oftentimes in the same region and the same ZIP code. The level of understanding is really deep. The secret sauce is really making sense of all of this, bringing it all together in one platform, making it targetable. Therefore, we use proprietary AI models, which really inform us about the elasticity, about the products, about which brand is losing, gaining market share, where the brands should target. The fourth is an independent, scientifically backed up-of-measurement methodology, which we also developed in those 11 months. It's completely API-based, works across all of the different brand formats and retailers. If we take a step further, how would this look like in the U.S.? Is this scalable? Is this a one-hit wonder? I think it's very scalable. I cannot provide you today with contracts, I can tell you that we are in advanced negotiations of closing data from more than 50,000 stores in the U.S. Sales data from stores in actual stores to understand that we are in advanced negotiations for contracts for in-store media, where we actually have hardware suppliers, existing in-store media players, wanting to work with us. They want to join us on the platform. They're integrating into our platform because it's the best platform on the market. The AI models and the up-of-measurement, of course, they're exactly the same. It's a platform game that we can now take to any country. To sum it up, in less than a year, we've built the largest retail media network in the country. If you have any doubt about the ability to execute in this company, I think this should be one of the proof points, that we are able to execute very fast if we want to, very precise, and we'll succeed. Now we're building the same thing in the United States. Geo-spatial targeting, receipt data, insights from consumers, AI-based in-store sensors to sofa-to-store contact. We have one unfair advantage in the U.S. That we didn't have in Germany. We have a world-class sales team that's already selling to CPG, healthcare, and pharma in the U.S., that is what Mishel is going to tell you more about. All right. I feel like we didn't talk enough about AI today, we're going to do a little bit more of that. For those of you who don't know me and my team, we spend a lot of time talking to advertisers, understanding what are their needs, what are their pain points, and trust me, they have a lot of pain points these days. When it comes to agency world, agencies are under a lot of pressure. They need to prove value to their clients, to the end advertisers, and it's not easy to do these days. We're all seeing the shifts in the industry. Some HoldCos are gaining market share while others do not. Even the ones that are most successful, for example, Publicis, when you look at their guidance for how much they're going to grow this year, that's the leading OldCo, that's 7%. At the same time, they're dealing with vendors that are basically telling them the same story. Everyone gives them the same promises around how they are going to use AI and help them succeed and drive value to their clients. Differentiation, as Remco mentioned earlier, is difficult. What I'm going to spend a few minutes on in this talk is how do we differentiate, how do we use data, and how do we turn signals into outcomes. Even that term, outcomes, is overused these days in the industry. Everyone talks about it. It could be outcomes, it could be incrementality, at the end of the day, what you need to show, as Raimund mentioned, is helping those advertisers increase sales, that's where we come to play. Prediction intent, all of the terms that I'm going to use during this session relies on data. If you don't have good data, your ability to predict the right outcome is limited, and as Dr. Fayyad mentioned, if you're lacking on the data side, you're not really going to accomplish a lot. By show of hands, who used AI today, LLMs? Keep your hand up if you used it in the last 30 minutes. In fact, there are people in this room, I'm one of them, who have agents running from them as we speak. I was actually seeing some people use their phones, interacting with their agents, while we were listening to the prior presentations. When it comes to using LLMs and using AI, that's really changed the way people interact and consume information. For a long time, search was limited to a tiny search box that was a strong signal for intent. I'll give you an anecdote. A few years ago, I spoke to one of our clients who told me that if they could, they would deploy 100% of their budget to search. I asked why, and the response I got was intent. That advertiser felt like when the user is looking for something, that's a great signal for intent as it makes sense. If they could, they would deploy their budgets to it. I asked the obvious follow-up question, which was, "Why don't you deploy 100% of your budget to search?" The response I got was scale. Additionally, search was limited in terms of scale. That's why advertisers had to expand and go broad. They were always hoping to find that intent or those intent signals elsewhere. What happened over the past few years is the evolution of LLMs is now giving those advertisers the ability to unlock intent much earlier and in ways that were not possible before. It's on us to go and identify those intent signals and turn them into outcomes. What's important here is not being overwhelmed by signals. There's so much data out there, it could be overwhelming, and in many cases, it can turn into noise and not to outcomes. The successful players in this market are able to turn those signals into outcomes. What I'm going to show is not just a presentation. I'm actually going to show you later a live demo of what we've built and how we're going to go with that to market, because it's not just about talking the talk, it's for us walking the walk and being able to show execution in that regard. The other thing about intent when it comes to traditional search is if you are only focused on the search box at the very last minute of where the user is searching for something very specific, that auction might be too expensive, you might be paying too much for it, and it might be too late. Maybe the user already made a decision between Reebok or Nike. What we want to do is we want to influence the decision before it happens in the consideration phase, and that's something that we're able to do by having access to those types of data points. The other thing that is important to keep in mind is this is not just a story about AI, it's a story about identity. For many years- Traditional advertising models were built around persistent ad identifiers, cookies, mobile ad IDs, triangulation, static audience definitions. That obviously worked in some ways, but it also created a lot of limitations because a lot of that data was stale. Remco mentioned the example of someone being in the market for buying a car. I could be in the market for buying a car today, but that might not be relevant two months from now. The recency of data, the accuracy of it is super important. When advertisers used old traditional ID-based models, being able to refresh those data sets was limited, and because of that, they were sometimes wasting their money, or in many cases, wasting their money because of that. That entire foundation is under pressure. The opt-in rates of users giving consent keeps going down. There is a lot of pressure around triangulation or fingerprinting in the industry. Apple is one of the leaders of that. They really don't like it, and they're going to continue pushing against it. That creates a huge opportunity for us. The question becomes, if identifiers are less dependable, how do you still find the right consumers at the right moment with the right message? That message also depends on the way we present it, and that's the creative aspect that was mentioned earlier. Our answer to that is, we're not going to keep chasing that old model forever. We will obviously continue using IDs when they're available. We're not against it. You have to evolve beyond that. Our answer is to combine proprietary intent signals from multiple sources, including zero-party data. That's the polling technology that was mentioned earlier. It's the ability to use on-device signals. It's the ability to do AI modeling and to have a privacy-first activation mindset. For many years, Verve has been the market leader when it comes to ID-less. We're not abandoning that. We're going to continue investing in ID-less solutions. In fact, the case study that Dr. Fayyad mentioned earlier with LinkedIn was all based on ID-less solutions that we have, all contextual, and you saw the results. The results were great. This is not a client that doesn't have access to the best players in the world when it comes to user acquisition. That's LinkedIn. They're owned by Microsoft. Our ability to perform in an ID-less environment is going to continue being a core part of the way we operate, but we're also augmenting it with signals that were not available before. When it comes to the consumers, they're still signaling exactly what they want. They're not silent. They're just signaling it in different ways. They're not just asking basic questions anymore to AI chats. They're having full conversations with them. They're also, in parallel, searching across the web, we have access to those data points. They're using apps. When our SDK is integrated into those apps, that gives us a lot of access to understand how users are behaving. They're responding to our polls. We have the ability to ask someone, "Are you in the market for buying a car right now?" That's a very powerful signal that we get that we put together as part of our overall holistic view of the consumer. They are browsing. They are comparing products. We all do that with LLMs. That's one of the biggest use cases out there when it comes to commerce. There are investors in this room. People are now using LLMs to make investment decisions. Now, is it the right approach or not? I guess time will tell. It is becoming a big part of the way consumers interact with LLMs today. If you are not using a partner that can connect all of those signals that I mentioned, data becomes noise. When data becomes noise, advertisers waste budgets. It's like the old John Wanamaker quote, which said, 60% of the advertising dollars that I spend are wasted. I just don't know which 50%." That's still relevant today. I assume some economists in this room. It's the invisible hand. When you look at what happens to the dollars from the advertiser side to the consumer acquisition or to the outcome, there is still a lot of wastage that is happening in this industry. We are here to solve that problem by better understanding what users actually want, what are they signaling, and turning that into outcomes. Targeting too broadly is obviously not going to work when you think about outcomes. Optimizing too late is going to result in outcomes that are not necessarily the ones that you're looking for. Missing the moments where demand is actually forming is going to result in, again, wasting dollars. That's obviously not something that anyone would want. This is exactly the limitation that Verve is solving. We specialize in capturing and connecting intent across all the signals that I mentioned, turning that intelligence into audiences, activating against those audiences and driving the outcomes for clients. When people talk about moats, the moat is the ability to do that. Because some players in the market have limited access. Some will have access to LLM data, others will have access to search data, others might have access to polling, but we are the only player in the market that has the ability to have access to all of those signals, connect them, and drive the outcomes for our clients. In this example that you see on the screen here, we're looking at a CPG example. This is a CPG advertiser selling healthy snacks. The old way of reaching relevant audiences for a product brand would involve targeting a broad grocery audience, again, might be a sale list, or people that showed interest in similar products, you go and you try to target them. That can work, but what normally happens is it's catching users too late in their journey. What we are doing is, again, going back to all the signals that I mentioned earlier, we're able to identify signals from various sources. You have from an LLM chat prompt, someone is looking for high-protein vegan snacks. It's not the direct brand that is being searched for, but it shows that this person might be looking for something that is relevant for this brand. We tie those searches with other online searches that this consumer has made on the web. We have the ability to go and ask this user, "What is your go-to snack?" in this example. Lastly, we connect all of that to purchase data, that goes back to what Raimund Bau said earlier. Your ability to drive outcomes depends on your ability to have access to the right data sets. In this specific example, it's being able to access purchase data because otherwise you're not really able to close the loop. The other thing that is interesting when it comes to data is, I mentioned noise earlier. Advertisers have access to many dashboards, probably too many dashboards. What those dashboards normally do is they answer retrospective questions. Why did we spend? What converted? What was the cost per acquisition? Which campaigns performed and which ones did not? Those are important questions, don't get me wrong, but they're not enough. I think that the harder questions are the ones that determine the next dollar. Where is demand forming? How does my brand show up in AI conversations? That's a big topic these days. We will talk about GEO. Which competitors are gaining momentum right now? Which audiences are emerging? What should I do next as an advertiser? Answering those questions are more interesting because they are more forward-looking and not just backward-looking. If you only look at backward-looking data, that's where budgets get wasted. If you're not able to identify intents early enough, your spend will not go a long way when it comes to driving outcomes. If you over-target people that are no longer in market, going back to some of the examples that I mentioned, you optimize towards something that is not going to perform. The commercial opportunity for Verve is to shift advertisers from backward-looking reporting, to forward-looking intelligence, and that's exactly what we're doing. What I'm going to do next is, I'm going to walk you through a few live examples of a platform that we built recently. What it's going to show is how do we read the signals? How do we understand the opportunities as advertisers? How do we take those learnings and activate against the audiences that we're creating? Lastly, there is the optimization part. Because again, if you're not able to show the results, the outcomes, the performance, you're not really accomplishing a lot. All right. Let me just get my glasses on for this. What you're looking at on the screen right now is a new platform that we've built called Verve Intelligence. Verve Intelligence gives us the ability to help advertisers understand how they're doing. When I say how they're doing, I'm referring to many different factors that I'll walk through some, and show them the ability to better understand how they can perform. There's something about Is it a bit blurry for you guys as well? Yeah. Can we fix the blurriness? There's dirt on my lens. Yeah, I was wondering if this was me. Let's try again. Someone should put some music on. It should be the broker. Huh? It should be the broker. It's rebooting. Okay. In the meantime, I'll give you some more context on what you're going to see. Basically, if you're a brand that tries to promote a product or a list of products in the market or some services, you're limited with what you can do, especially if you're SMB. What normally happens is SMBs go default to social media, the walled gardens that were mentioned earlier, and that's what prevents you from being able to gain access to more advanced targeting and activation of campaigns. When we thought about building this product, we tried to address exactly that. How do we democratize access to advanced advertising and help indie agencies, mid-tier, small-tier agencies gain access to more advanced advertising? I see the screen now flickering. The idea behind it was exactly to follow what I just mentioned, which is help brands identify the market, the market opportunity, intent signals, and then turn that into activation in a single self-service platform. That's going to help us not only better serve our clients, it will also allow us to unlock new revenue streams that otherwise would not be unlocked. Because when you think about it, yes, we are working on building a bigger sales team that will give us better coverage of the market, especially in North America, where I was telling someone outside anecdotally, the person who's covering Canada for us currently lives in Florida. Obviously, that's something that we need to address by having a bigger sales team. We won't be able to serve all 4,000 agencies in the U.S. just with hiring sellers in every single market. We will obviously focus on the more important markets when we determine our sales growth strategy. There will be parts of the market that could benefit from using a self-service platform like this, where seller interaction is not necessarily needed. That's a big part of our growth. The other aspect I will mention is Should I? Yeah. Okay. The other aspect I'll mention is Let's see. Still not see. When it comes to the performance, some of the agencies that I mentioned, indie, SMB, et cetera, they don't always have the right tools in order to be able to perform. In order to help them be successful, we need to make sure that the platform that we're building is helping them show up with their clients and perform for their clients in ways that they don't currently have access to. When I think about our positioning right now in the market in terms of product, in terms of the team, the quality of the team, I've been with this company for almost 12 years now, two years with Verve. It's the strongest product offering we've ever had and the strongest team we've ever had. I'm very bullish about what we're going to do next in terms of conviction in our ability to win in the market. Going back to the demo, it's always risky to do a live demo, but looks like now it's running. What you're looking at is Verve Intelligence. What you see in the top is basically a list of verticals that we all operate in. In the example I'm going to show you, I actually want to look at a very current, relevant vertical, which is gyms and fitness. Obviously, summer is coming, everyone needs to shape up and the smart gyms need to make sure that they're winning market share. In this example, it's a made-up gym brand that we're promoting. What they're looking at when they start using Verve Intelligence is basically understanding of the consumers that they're working with in terms of what are they searching for in both traditional search and LLM and sales data. Again, think about it's intent that drive outcomes. If you don't have visibility to both parts, both sides of the spectrum, you're not really accomplishing a lot. I won't go through every single aspect of this tool, but I think you'll get the point fairly quickly. I'm able to see if you look at the top brands in fitness in the U.S. in this example, who is winning when it comes to share of search, when it comes to share of LLM prompts. Again, how brands show up in LLMs is super important these days. Giving advertisers this type of visibility is unique and will help them understand how can they win in the market, how do they need to show up in those types of conversations. What's also interesting is when you think about a gym brand, normally what happens is January comes, everyone has a New Year's resolution and they want to get into shape. What's interesting, what we see in the data is that you might want to actually start your advertising in December when the intent is forming, because if you wait until January, then you're competing with everyone else that is spending dollars right now to acquire users. Again, it might be too late. Maybe the consumer already made a decision around which gym they were going to go with. Giving brands those types of visibilities is really helpful. There are lots of other factors, brands through call cards, what are consumers actually looking for? Is it goals? Is it training? I'll give you an example. Someone might be looking for comparing between Pilates and something else, and that's where the intent is forming, and traditional search would not use that to go and find that consumer or target that consumer. That might be a good idea for the advertisers that we're talking about here. There are lots of other factors here. Obviously, geo plays a big role. You cannot assume that the interest in fitness is the same in every single place in the U.S. or even within specific states or cities. Where it gets interesting is what I mentioned earlier. You have all these insights. As a brand, you're able to review your competition. You're able to understand how you show up against your competitors and whatnot. What do you do with that? You don't want that data to just exist in strategy decks that don't really lead to anything. What this platform allows you to do is to take those insights, to take these audiences, then activate against them immediately. What I'm going to show you next is, I select a few random segments, then I'm able to go and queue them up for activation. What I just showed you normally would take days, if not weeks, to figure out, we're able to do that in just a few minutes. The next part is the activation part, that's also something that used to take a long time to set up, we're able to do that in minutes. Every campaign starts with a brief. The advertiser will come to you and say, "Look, I have $30,000 to spend. I am going to try to target these types of audiences in these geos, these are the flight dates of the campaign." Obviously, there's a lot of back and forth. You make recommendations, once everything is agreed on, the campaign goes live, there is optimization that will happen. We are changing all of that, we're making this process much faster and easier. What I'm going to do now is we just landed on the ads manager. An advertiser will have a view of how they're performing, if they want to launch a new campaign, they would click New Campaign here, then they're able to upload the media brief aspect that I just mentioned. I will select Forge Fitness, Media Briefs, open it. What's happening now is our AI platform took that media brief and created a campaign already. All the different aspects that you see here, in terms of the campaign, the target KPIs, what's the cost per acquisition goal, et cetera, the budget, those were all taken from the media brief that I just uploaded. The advertiser is able to do that very quickly. The audiences that are selected here are from the media brief, but we don't want to just rely on what's in the media brief. We actually want to add the audiences that I just created, that I showed you. I'm clicking here to add these audiences that I just created, and then they will be added to this campaign. I scroll down. The creative is already there. This section here around frequency cap needs some input. I have an AI assistant on the right side, as you can see. It's giving me a recommendation. I trust AI, so I'm going to say yes. It selected front-loading for the spend, and then I go to the next screen. It will show me the budget allocation that was done based on the brief, and then it will give me some recommendations. It's very common that whatever is put in the brief is not really going to be the best approach in terms of optimizing the campaign. I can either accept or dismiss. I can also ask questions around why a certain recommendation was made. I'll skip that for now and move on to the next page. I get a campaign summary, and once I'm good with all the settings, I'm able to launch that campaign. At this point, the campaign is launched. It's actually Q3, so it's queued. What I just showed you, again, used to take days, if not weeks, and we're able to do that in minutes. This campaign that I just created is queued, but I have a campaign that has been running for Q2 that I'll click on. This gives me a summary of the campaign that was previously launched. I'm able to look at metrics. How is it doing? For how long has this been running? It looks like I'm doing very well in terms of the cost per acquisition target, so we're lower than the target. I'm able to also view what kind of changes, what kind of optimizations were made to the campaign. One of the challenges working with AI platforms is, in many cases, they are black boxes. You're not able to understand if changes were made, why they were made, and what was the outcome of those changes. We're solving for that black box here. Basically, for every change that was made, it's telling me if it was AI-generated or if I made a manual change, and then for anything that I'm not fully sure about, I can ask AI why a certain change was made to get better visibility into how the campaign is performing. That's it in a nutshell for the demo. We can go back to the deck. I think that the important aspect to keep in mind from this demo is, A, our ability to, again, understand the signals, use them, then drive intelligence and deliver outcomes, and B, the reduction of friction. Your ability to run campaigns efficiently using a platform like this will lead to many efficiencies in terms of how many people do we need in order to support campaign execution, and also the driving of the outcomes which are super critical for advertisers. With that, this just as a recap, the future of advertising will not be won by companies with scale. Scale is important. That's table stakes. If you don't have scale, you won't accomplish a lot. It will be won by companies that understand intent earlier, activates it faster, and proves the outcome. That is where Verve is building. Thank you. Thank you very much, Mishel. Okay. Thank you. I'm staying here. You're staying here on stage. Exactly. Can I click it, please? Yeah. Thank you. That is the end of our tech session. That means we're moving into the next Q&A session correctly, perfectly. Raimund, Usama, please come on stage. This Q&A session will now focus on the expert session we just heard, on retail media, predictive targeting, and the broader themes around signals, data, and outcomes. As before, if anybody would like to ask a question, please raise your hand and one of our mic runners will come over to you. Perfect. We got the first one. For the online audience, please already type your questions into the chat. I will get them directly on my display. Please go ahead. Great. Laura Martin from Needham. Really interesting product. Is the business model charging a separate data fee for this, or is it only available through the DSP and you're sort of upcharging on the cost per thousand by using all this proprietary data? Every campaign that we sell has a data component to it, and based on the data sources that we would use, we would price it accordingly. We have different rate cards based on the outcome that we need to drive. Obviously, running a healthcare campaign will involve different execution than running a CPG campaign. We take all of those factors into account when we decide on pricing. Is the pricing the take rate or the pricing is like a monthly subscription? It depends on the product. For a managed service, it will be a full pricing that will be delivered. For PMPs, if you activate programmatically, you are able to define a cost per thousand impressions as part of the build creation. For DSP, it could be a combination of a take rate plus data fees. I was surprised in the video, my last question, you had a camera scanning a woman in a store. That feels highly invasive. Can you bring that kind of non-permissioned assessment to America? First of all, it's not a camera, it's a sensor. That device does not store any PII information. That device only stores the information of group composition, sociodemographic. It is by definition an anonymous or pseudonymized device. Yes, there's legal assessments for, I believe, 49 states as of now. I have to add, we are working with different sensor partners. We don't need to work with that one. We can work with other ones. There's also CCTV cameras that are used by some of our major hardware partners deployed in the U.S. that are already extracting that sociodemographic information and are making use of it today. We would be able to access that information. The key is difference between a sensor and a camera is that a sensor does not store the image information, so it's non-invasive. It is really much less than the CCTV cameras we have all over the stores actually storing the image information. Thank you. Great. Raimund, please go ahead. This is a cynical question now, Mishel. With the software you have developed converting the, I call this instinct data into actions, which means the high speed of converting it to a saleable campaign. Should the productivity of the sales force not go up dramatically? They have now at a fingertip the data and the demos and the examples, which really means that the onboarding, which we are doing here, we're spending this EUR 10 million, get these people onboarded and so on. It's a thought I would like to leave with you is should the productivity targets not be higher? The question I have is actually a different one is, where's the value? The value, as our professor has highlighted, is data. He said, "The only investment-grade asset is data." That's what was said. What we are selling in the campaign is data. The technology we have converting what our learnings by gathering the data into actionable campaigns is a software which I believe others can replicate or others can build as well, and is not really the asset. To me, the growth relates to how can we scale the data gathering, and how can we differentiate from a data gathering relative to the peers, and what is your strategy to make our data gathering better than the competition? A big part of what we built over the past few months was exactly that, ensuring that we have access to the right data sources that allow us to build those audiences in the holistic way that I presented. That's an ongoing thing. It's never done. For example, our next phase is going to involve running ads in LLMs themselves, and that will also give us more data around how users are interacting, even more than what we have right now. It is going to be an evolving thing. I agree with you that having access to data is super important, the execution of what do you do with that data is equally important, I think that we're uniquely positioned because we have access to all these different data points that others do not have access to or cannot combine in the way that we do in order to create differentiation in the market. Just to refer to your comment about productivity, it is a big part of what we're doing as a company right now. AI efficiencies is something that we spend a lot of time on. We just had our AI demo day last week on Thursday, this is a constant thing that we're pushing in the organization. Just as a clarification also, data being the only investment class asset I was talking about within the AI equation. If you're talking about the setting, let's say, of serving ads, there's huge value to figuring out how to do integration, how to work with existing systems, how to integrate with the right publishers. That's a different equation. Yeah. Great. Matthew, please. Thank you. Apologies if I didn't quite grasp this in the presentation, but I'm just trying to better understand how you effectively unify all these disparate data sources. If you think about the search intent data that you've acquired from Captify, the LLM data. Basically, how do you connect those to formulate the audiences in a privacy-compliant manner? Yeah. In general, in advertising, if you have access to IDs, you normally use an ID graph. We have an ID graph that we use internally. If you're looking at ID less, ID less could involve contextual signals. Where is the user? The location, which app they're using, lots of other contextual signals that you can use. What's important or what's interesting is what can you derive from other users that you saw, that you have access to their data about a certain outcome in a way where if you get those signals elsewhere, you're still able to make a prediction that that consumer is going to do something that the advertiser wants. Being able to operate in both ID less and ID based is super important in these types of applications. Thank you. Any other questions here from the room? I would say, Simon, return to an open question from earlier. We had a question earlier on the metrics of retail media. What are unit economics take? How could take rates look like? What's a typical campaign size? Overall, a general picture on margin without getting too specific, please. Yes, I'll try to do that. A lot of this information is public in general on retail media. Typical campaign sizes of in-store retail media range anywhere between EUR 20,000 and EUR 200,000. That is the regular ranges we have been seeing as well. That is in line with what some of the competitors, the retailers are running. I have built a large retail media network before with Schwarz Group, the owner of Lidl and Kaufland, that was in line to what I saw there as well. Generally, how we sell this is in bundles. The trade marketing manager would receive a different activation bundle that would be screens, would be some upper funnel activity, would be the passes you saw, the wallet passes, all the way down to checkout couponing. For media manager within a brand, the story is a bit different. It starts at the upper funnel. It's more CTV, mobile-based, really goes down to the individual store to the redemption as a proof point and the receipt data as a proof point. That is how we built this. Take rates, typically, if you look at competition in that market are between 20%-30%, that our typical take rate. However, if you look at the story with and without data, if you research the typical CPM of Digital Out-of-Home campaigns within the store, you'll find that the CPMs are rather low. What we've also seen in the industry, as soon as this combines with data, CPMs fivefold, tenfold. It's up to $ 30, up to $80 CPM rates, which we're seeing in-store data, actual receipt data combined in store. There's ample of public examples available for that. What I believe that we build is really a coherent story for the brand, touching all parts of the customer journey and adding data to all of the points. By the way, this is a coherent platform that we are building onto. To your question before, it can be sensors, can be no sensors, can be extensive receipt data, can be store-level data. That's how flexible we are, and obviously pricing and willingness to pay as advertisers changes as more data is added. In the end, it's a data play. Great. Maybe to elaborate a bit on that. On the international expansion, of course, the U.S. market, you said it before, is one of the most interesting markets in the future for us. What would you consider the main differences between Germany and the U.S. in a potential rollout? Also from size and pace, it took us 10 months to do that in Germany. Does it take five times longer in the U.S., or is it possible at an effectively similar rate? Internally, we are super ambitious. We set high goals, and these goals I'm not going to communicate here publicly. What you can imagine is that we are far ahead in terms of negotiations with partners. We are far ahead in terms of actually having this platform. You saw what Mishel and his team built as a platform. Obviously, you can connect the dots. All of this eventually comes together into one comprehensive understanding of consumers. Comparing the U.S. retail market to the German retail market, what I tell U.S. retailers often, what you see in Germany may be your future in one or two years ahead in terms of margin compression, private label brands and so on. The markets are very similar, very comparable. I've worked extensively in retail in both markets. There is not that much of a difference in both markets, really. Great, thanks. Any more questions from the room here? No. We've got another online question. To Mishel here, probably. We're seeing increasing regulatory pressure on Google, and with that, potentially more regulatory pressure on digital advertising in general. How could this increased pressure birth, and how might it affect our revenue streams? I see it as an opportunity. Being able to run a platform that was built from an infrastructure as idealist gives us a lot of flexibility when it comes to being able to perform for our clients, even with all the regulation that exists out there. I think that the key is what I was kind of alluding to earlier, being able to make the prediction, the best prediction you can make, with data that is available to you. That's something that we've built over the years and we'll need to continue building and capitalizing on as it proceeds. As Remco mentioned earlier, every year there is something new, like the breaking in of the cookie, wars, tariffs. We have to be resilient and expect that there will be curveballs thrown at us at any given year, and our ability to execute and deal with those types of situations is critical. Great, thanks. Looking at the time, maybe the last question to Usama. Talking about machine learning. It looks like one of the main differentiators in artificial intelligence will be machine learning. Training new models here appears to be much more time-consuming than training large language models. How could a new entrant in the market reduce that advantage of well-established players by using new technology or significantly higher financial resources? Higher financial resources. What I would say, first of all, is a couple of things. Number one is you don't have to use the expensive models. You can use the cheaper models. We use, for example, in our projects, a lot of the Chinese models that happen to be very reduced versions of the big models and very effective. On the machine learning, not everything is an LLM, not everything requires generative AI. The reality is, I know the market is obsessed with generative AI. Most applications of AI in the market are predictive AI, kind of traditional machine learning models. Those are super cheap, like too cheap. The hard part is getting the data right and understanding the representation, like figuring out what is the opportunity in the market, how do we model it, and how do we manage this early signal about leads, et cetera. That's where the AI models, predictive and generative, can help a lot in terms of coming up with strategies for how do you nurture a lead? If I know somebody's not going to buy a car for another two years, what is it that I want to show them now, right? Or six months before, or one week before. Great. Thanks. I would say those are great final words. Thank you all from my side, also to the audience. We're coming to the next part now. A warm applause, please, for our Q&A team. We're coming to the closing remarks. Again, thank you, Usama. Thank you, Raimund. Thank you, Mishel, and thanks everybody for their questions. Before I hand back to Remco for the official closing remarks, let me add a few final words from my side as well. A Capital Markets Day is always a moment to zoom out from the quarter-to-quarter noise and to look at the architecture of a business, the market we operate in, the capabilities that we are building, and the outcomes we believe we can produce. We hope that today's session has given you a clearer picture on Verve commercially, financially, and technologically. Most importantly, we appreciate the time you have invested in us today, the dialogue with us, and your direct questions. This really means a lot to me and to our whole team. With that, Remco, let me hand it back to you for the final closing remarks. Thank you very much. Thank you, Ingo. Capital Markets Day is for us, it's once a year, and it's always a good point in time to realize the progress we're making as a company. I'm really proud of the team, the presentations that we saw today, not only the presenters, the whole team actually. I think it's really good to show what this company is building, is doing, we are, as I mentioned before, work in progress, which is never a good sale to investors, I know. This company has a ton of potential, and I see a lot of stuff coming together that we have been building on in the last years. What you saw here or what we presented, skilled access, that's important. Skill is not making you win, but skill is a necessity. Vertically integrated, efficient, with tools that now also go towards advertisers, which go even into stores. Really getting a full integration, vertical integration from the consumer to the advertiser and the other way around. Depth of data. A lot of talk was about data. Data is key, data is core, data is an asset, and we have a ton of it. You need to use it. It's about targeting, measuring outcomes, closed loops. Those things we have been building in the last years via acquisitions, via organic growth, really building all those pillars, and they're now coming together. That should further result in further growth. We have certainly shown good growth in the last years, and we will continue to really show good growth in the future. I enjoy working here. We're building the trusted alternative to walled gardens, as we say, that delivers business outcomes at scale. We had a lot of brainstorming about what is the real essence? What are we really going for? It's basically without having your own walled garden, curating the open internet. There's tons of publishers out there that all have more and more problems monetizing their inventory, if we can bundle them to our own open walled garden, wrong word, we really have a big chance of being very successful here. We have strong momentum. We're ready to further grow. Organic investments, organic growth is the key. AI is key. Data is key. A lot of things are key, for the rest, it's also a lot of it is about execution and the team, that's, I think, a very important point here. Without the team, we would not be able to do this. I'm super proud that part of the team is here, and we look forward to continue. We hope you like the company. We hope we were able to really give you some more feeling about the company, some more depth about the company, and I'm looking forward to the next Capital Markets Day. Thank you all very much.