LiveRamp Holdings, Inc. (RAMP) Earnings Call Transcript & Summary

Thank you for joining us. I'm Drew Borst, Head of Investor Relations. Glad you could make it out to RampUp 2026. We're excited to have you and have you at this briefing and Q&A session. So let me just kind of give you an overview of our agenda for today. We're going to start off with our CEO, Scott Howe. He's going to talk to you a little bit about why we really believe AI is a tailwind for our business. Second, we're going to have Travis Clinger, who is our Head of Partnerships and Connectivity, speak about how we're adding new AI nodes to our network. And then he's going to hand it over to Matt Karasick who's our Chief Product Officer, and he's going to speak to you about Agentic Advertising, how the marketing workflow and advertising ecosystem is changing by adopting AI agents to accelerate their entire process. And then after that, we'll go to Q&A. We'll have you out of here by 8:00, so that you have plenty of time to get to the RampUp programming. So without further ado, I will hand it over to Scott Howe.

All right. Are you going to give me the clicker? Okay. Perfect. Listen, I can talk to this slide. I would tell you, if you have a chance, go watch Matt Karasick's keynote when he was on stage yesterday and demoing our product because when you actually see like what it does and you hear Matt talk about it so simply and eloquently, it's actually more powerful than anything you'll hear me say right now. I always talk about how we're in the center of things, and we really kind of are. And that's true of the last 10 years. It's true of the next 30 years because every company needs to make good decisions and drive good results. I mean that's why they're in business. And the only way you can do that is by being more intelligent than your competitors. And the way you're more intelligent than your competitors is collecting and analyzing the right data. The problem that every business on the planet faces, bars, maybe Amazon and OpenAI is none of them have all of the information that they'd like to see. And the problem is that the information that each of them has is probably unique to them, their CRM files, their purchase histories, their web-viewing information. But the things that they want that would allow them to make an even better decision, well, that information is owned by someone else. And so metaphorically, you can kind of think, and I'm speaking to this picture here, LiveRamp's hosted a dinner party. We have a really big table, and we invited everyone on the planet. And by virtue of sitting at that dinner table, they have the ability to communicate and work with anyone else who has a seat at the table. So we've done integrations, set up the data governance, set up the Data Collaboration technology for anyone who has data to use it at the destinations that matter or ingest data from partners who may have complementary pieces. And so you can see whether it's brands, agencies, publishers, consumers, tech partners, data providers, all of them have been invited to the party. And if you've spent any time at RampUp yesterday, or you have a chance to today, you will see that come to life. because when you look at the attendees here, it is -- I think Meta has 20 people here, including the head of their business. Google is well represented. I met with OpenAI yesterday. We got Microsoft. We have Publicis. We have like literally every company that matters is here, represented at a senior level, and they're all talking to one another about what they should be doing together. And we're not even part of many of those conversations, well, at least not in the conversation, but our technology underpins everything. And so whatever they decide to do together out of their conversations, our business is going to grow. Now that said, we're at this inflection point, and it has a lot of investors a little bit anxious. I think as they take time to understand who the winners and losers in an AI world will be, they will come to the conclusion that we have built for a long time, which is we're going to be one of the winners. And what we see is 3 megatrends. Number one, consumers are changing how they behave. And all of you probably experienced this yourselves, like instead of going and submitting a 3-word search query buy luggage, instead, you're writing a much more eloquent query, probably 20-plus words. It's very descriptive. Hey, I want a new carry-on bag. I don't want to spend more than $300. I want something that's top-rated. And then lo and behold, Google AI Shopping gives you choices and you can buy straight there. And so consumer behavior is shifting. The second thing is the way that work is done is shifting. So yesterday, I talked about our Publicis partnership. Well, Publicis has built a whole AI layer called CoreAI. And the way they plan media, the way they distribute creative, that's changed. The way that Meta makes their ad-decisions, that's changed. The way that virtually every company that's represented in this conference, the way they're doing work is changing. It's becoming more streamlined, more intelligent, more automated, and more effective. But the third change is that data and identity are the fuel that make both 1 and 2 happen, and that's our business. We're the intersection of data and identity for all of these companies. And we just expanded the dinner party because we invited a whole bunch of AI applications to have a seat at the table as well. And so just as for a decade or more, we have been working with folks who have data and activating that data at the places where they want. Well, AI has a lot of data, and they want a lot of data to make their models more effective. And so we have 20 some -- you'll hear this, 20 some AI applications live already. I suspect in a year's time, that number is going to be in the hundreds. Our clients are helping prioritize which AI partners they want us to sign up. And just as we can activate data in Meta or Google or Microsoft, so too now can we activate data in these AI applications safely and securely, use it in clean room such that our clients are protected and their data is not misappropriated. Finally, I'll just say why is our moat really deep? Well, 4 things. Number one is identity. The fact that different companies have different datasets. It's really hard to comingle all those datasets. You need a matching key, essentially, a Rosetta Stone, a translation layer. That's what identity does. And we have the largest Identity Graph. And so whether you're an AI provider or a legacy provider, it doesn't matter. You need to have that translation layer that we provide. Second is interoperability. If you look at the companies that we work with, it's all over the map. It's every major cloud, it's every agency, it's all the major publishers, and we make things fungible across all of those companies so they can work together. Data Governance, super important in the world going forward. Everyone knows that they need additional data, but they're all scared to death that their data will go into the public domain and become part of the base LLMs, be commoditized and used for their competitors. And so they are all requiring safeguards before they do any Data Collaboration. That's what we provide. And then finally, but really most importantly, it's scale. Like we're one-of-a-kind. And even all of our competitors, we power. And so that is incredibly powerful and a real reason why no one ever takes a run at us because the time it would take for them to duplicate what we've built just can't be done. You can't put -- nor can you put a group of 20-some-year-olds in a room together and say, "Hey, build this using technology, because you actually have to go have the conversations with the gazillion different companies, earn their trust, put the safeguards in place. It's not something you can just immediately stand up. So we really like our business. We liked it a year ago. We like it even more going forward. Our margins continue to increase. Our revenue continues to grow. We're going to -- as we look forward, we think that continues, and we think it accelerates. I would tell you my biggest disappointment as a CEO over the last year is the fact that last quarter, our revenue is growing at 9%, and it kills me because I know that the difference between 9% and 10% growth is like one client in a few million bucks. The difference between 9% and 10% growth to Wall Street and valuations is ginormous. And our eyes are firmly on being a Rule of 40 company and beyond. To us, that means mid-double-digits growth, 13% to 15% locked in quarter after quarter and much higher margins than we have today. We said we'd be Rule of 40 by FY '28, and we're going to get there by then, if not sooner. So that's what I got, and I'll turn it -- give you the clicker, my friend.

Good morning, everyone. So I will talk a little bit about our AI partnership and ecosystem strategy. And then I'm going to turn it over to Matt, who's going to talk about what this industry looks like as AI modernizes every workflow. So diving into kind of where a consumer is spending time, we are seeing kind of for the first time in a while, a massive shift in how consumers interact with the world, how they spend time online and how they think about the world. And so you see this today with ChatGPT, with Perplexity with all of the different AI tools, but you're also seeing this permeate into the B2B workflows as we think about Agentic Workflows across the board, new ways to buy and sell media, new ways to do creative. And so we're really seeing a massive change here. And so at LiveRamp, we see this all as a massive opportunity to unlock new use cases. If you go back 10 years ago, LiveRamp's main priority at that time was how do we connect into social platforms, so Meta and Google. TikTok wasn't a thing back then. So it wasn't on our list. And open web, right? We talk about cookies and MAIDs. In a conversation in 2016, you would tell people how many cookies do you have? How many MAIDs do you have? I don't know how many cookies we have today because we don't even share that anymore. We don't ever get asked that. That time has passed. In 2018, 2019, we started to see the rise of FAST and CTV. And we saw folks saying, okay, how do I get on the big screen? And then you had some players out there, and they said, we're never going to do ads. Netflix was famous for this. We will never do ads on CTV, we're a premium product. Disney was famous for this as well. 2022, May of 2022, Disney announced they're going to do ads. Well, Disney became a key part of our network. Disney+ is a popular node today. July of 2022, Netflix announced we're going to do ads. We met with them about 1 month later. They hired a Head of Advertising. They said, one of our first priorities is we got to build an ad server. We don't know how to sell ads. We're like that seems fair. And then priority #2 was how do we bring in audiences and measurement. Today, Netflix is our fastest-growing destination. And so you had this wave of CTV. And then you had new social channels emerge. One of our top destinations now, TikTok who didn't exist a decade ago. influential, 1 in 3 Americans use TikTok every single day. They use it for an average of 2 hours a day. It's a lot of time on TikTok. And so as we look at the future, we see this as just another iteration of this journey. Just as our ecosystem has evolved from 2016 to adding CTV in 2018 to 2022, we're now going to add all these AI nodes. And so Scott mentioned Google Shopping AI. We'll do a little deep dive into that. But as you think about Copilot, OpenAI, Perplexity, Graph, all of those are consumer surfaces that we are going to enable as nodes for our brands to activate, measure and collaborate. We see them as activation but also Clean Room nodes in our network. But then there's a couple of other names in here, Scope3 and Chalice. And that gets to -- we see the AI network is not just limited to search experiences. By far, the most exciting thing here is that group on the top-left, those search experiences. We think those are meaningful and they're the most talked about. But there's all this other innovation happening in our network. So our brands are thinking about how do you change how you do creative. Creative is a really expensive process, right? You've got to take a product, you've got to A/B test it, you've got to have designers build it. We now work with a partner now who can take a picture of a product and make 10,000 creatives out of it in almost no time and then A/B test those in real time. This was not possible in the past, and they can do it for almost no cost. And so you see all of these companies coming out in the AI-creative space. And so dynamic creative kind of rotating into AI creative. These are all destinations. As Scott mentioned, we have over 21 destinations available using AI today for our brands. We see that rapidly expanding. We think AI is going to massively change commerce. This may merge with those search experiences. You can now execute buys on ChatGPT. As we'll talk about with Google AI Shopping mode, we're seeing more shopping move to there. We're super excited about that. All of that signal for activation, but also for measurement and for Clean Room. We're seeing a huge rise in Agentic Trading to us. Each of these would describe what they do a little bit differently. But what they're trying to do is redefine the media-buying experience. So today, whether you're buying a direct deal with a publisher or whether you're using a DSP, there's lots of inefficiencies. There's extra take-rates in there. There's extra people in there that are inefficient in that workflow. And so each of these companies is looking at it and saying, okay, I can make this step agentic, where a person picked up the phone today, call or bought a direct I/O, an AI agent can do that, and that can create this much savings. Well, I can create the savings in a workflow and take the take-rate down. And so each of these are partners where our brands are sending destinations. We have live case studies here today. We're also seeing Measurement. Measurement hasn't been disrupted in a while, and it's ripe for disruption. And so we're seeing new companies. Newton Research is one that's going to be on stage later today. If you have a chance to see them, they're a pretty incredible company. As someone who has no data-science background, you can use their tool and make an impressive Measurement report. And so they are making Measurement available to everyone. And they're doing it on top of clean-room technology like ours. And then you've got conversation tools. So we talked about the famous ones, the ChatGPTs, the Copilots, the Google AI Shopping. But you've also got all of these brands who are saying, well, hold on, if I'm an airline, people are messaging me every day, hey, what's my flight status, where is my gate? AI can answer all of this. What's the gate change? What's the flight status? Oh, you need to rebook, here's the next flight to your destination. But you also have so much data the consumer is telling you. Am I happy that you rebook me to the destination? Or was it 3 days later and I'm really annoyed I'm spending 3 days in a connecting city? Well, all of that can then be translated by AI. In the past, a human would have to read them be like what's the tone? AI can take that, segment it. That's more data for activation, more data across AI. If you're running ads on Google AI Shopping and the consumer just had a bad experience on your chatbot on your website, we link those 2 things together. And so we're seeing more tools like that roll out. This is the first iteration. I suspect next year, you're going to see -- we'll have to figure out how to fit it on the side, Wave 2 as we add more beachheads in our AI ecosystem. And real quick before I turn it over to Matt, I just want to highlight one of these kind of in reality. So Scott mentioned AI Shopping. This is Google AI Shopping mode. You're looking to find a suitcase. You can tell we all travel as all of our examples of travel on this. But here, you are typing in a pretty long query. Again, the average Google term is 3 to 4 words. The average AI term is 12 to 14 words, so a totally different search term here and then promoting that. We have a dozen-plus brands using this today already. We have our first case studies coming out. This is pretty exciting. We plan to do this across all of the different LLMs. And with that, I turn it over to Matt.

All right. So that's sort of the ecosystem side. But on the Enterprise side, what are our customers doing to augment how they do their marketing process, what their marketing stack looks like to do this. And so I tell this story when I first did it as an Eagles fan, the Eagles were still the Super Bowl champs. They're not anymore, but I'll still go with the Saquon story. And so imagine Saquon is in a marketing group at a company like Reebok, he has an idea. He's trying to gain some market share with a new product line, and he knows that the partners he needs to be able to execute his vision are the retailers who carry his products and competitive products and adjacent products. And so he has this campaign idea. What he is going to need to do to execute this to reach out to all of these retailers and to say, "Hey, here's my idea, what data can you make available to me? Can I use it? Can I execute these queries? Listen, we've done a great job building a platform that enables this to happen that enables this brand to work with many different retailers who have this data then to the media ecosystem where these ads would run and get a feedback-loop. And the way we do this is we then make it easy for everyone to connect their data. We make it easy for everyone to put the rules about what can and can't happen in that data for them to be able to go back-and-forth and iterate on these ideas. We then go and build our UIs, and we try to build the perfect 6-step wizards that will make it so that Saquon and his entire team can click through, do all these things and then someone on the retailer side, they have a different set of 6 screens for them to be able to do this or maybe Saquon is one of the handful of really, really sophisticated customers who can march in a software engineering team and say, I don't want to use your eyes, your UIs. I want to build my own 6-step wizards that are perfect for just our company. And so we'll all say that's going to be great because once they do, we're so embedded. But the other side of us can say this is going to take a while, right? And I hope they don't miss a comma, because then it won't work when they go to build all this. And so when you hear us get really, really excited and talk about AI as a real tailwind for us, it's because all of that whole process, all of that friction, all of these UIs and 6-step Wizards and e-mails and phone calls were in the way of Saquon getting access to the data that he needed to be able to have this vision come to life. And so AI is removing huge amounts of that friction and accelerating how much data he can use to gain these insights -- and so here, we have certainly more than 12 or 14 words for Saquon to type this idea out, but it's certainly fewer than 36 Zoom meetings and attorneys and redlines for him to be able to type in and say, here's what I'm trying to do. Can you help me? And so here, you can see this notion, and you may have seen this demo I did yesterday. But really what's happening behind the scenes to get a little bit geeky around it is he doesn't know this, but he's interacting with the first agent, this Orchestrator Agent who's saying, okay, I see what you're trying to do. I'm going to help you make a plan. And so here, this agent is coming back and saying, here's the plan I think you want. Here are the campaigns I think you're focused on. Here is the joint business plan summary that I'm looking at that gives me context as to what it is you're trying to do. It then when said, great, I now know that there are 3 particular Clean Rooms where partners have made a bunch of insights to you. Now, Saquon, you don't have to go and call your analyst team and tell them that the next 2 weekends are canceled, so they can go and crunch all of this data. We'll go and summarize all of this. And you can give that analyst team, hey, here's a set of recommendations that it's -- that this data is telling us. Then it takes these insights out, and you'll notice this magical moment that's going to -- we think it's so magical today. I promise in a year, we'll all just take this as a given. That insights agent handed it off to a Segmentation Agent, right? Because an agent that's good at insights is going to be very, very focused and purpose-built there. And then it's going to hand it over to a Segmentation Agent that knows exactly what to do with those insights. And so it moved on from there. Another really important piece of all of this is that as we build these agents, LiveRamp is not going to win because LiveRamp builds better agents than anyone else. That is not what's going to happen here. Much more, it is that what all of these agents need across that entire ecosystem that you heard Scott and Travis talked about is those agents need access to the right data. That is the most important thing here. And that data is in our network, and it's connected with a layer of governance in front of it. And so it is much more that we'll let the ecosystem, be it Saquon's’ team can go build these agents, Saquon's’ partners can go build these agents. And so here in this example, we had our own -- we built a Reference Look-Aike Model Agent. But more importantly, you'll see partners like Chalice that you saw on Travis talk about in our ecosystem, building their own agents that access the data in our ecosystem. And so this is the paradigm you're going to see. We're going to -- certainly, LiveRamp will -- in addition to continuing to build those 6-step Wizards, we will build reference agents that make it easier for users to do what they're trying to do in our platform. But also much more importantly, we're going to see our customers build these agents, and we're already watching this today. Travis talked about Newton. They've built agents on top of our Cross-Media Intelligence Clean Room. Measurement is a means to an end. There's 2 ends. First, it's to go tell their CFO, you gave me this much money. This is what happened. I can -- you should keep giving me money. But really, the other end is it's for optimization. It is for, hey, did this work well or not? And what would I have to have done better? And so Newton being able to build its own purpose-built agents, this team is Measurement Experts. They've built agents that know how to pull the signals that come out of those Cross-Media Intelligence Clean Rooms -- and those Cross-Media Intelligence Clean Rooms can only exist because we were able to get -- Sakequan is able to go to his entire media plan, social platforms, CTV platforms and say, I need to be able to measure and query across all of your data at once. And the only reason that these companies are able to say yes to that is because of the Governance that are in place, the rules that they're able to enforce and because that data is connected with that Foundational Identity. And because of that, that's a great ground for companies like Newton to then be able to bring and build their expertise on top of our ecosystem. And so this is why we're very, very excited is all of this using of this data and turning that data into value, what is mostly in the way is human processes. The amount of time it takes for e-mails and Zoom calls and redlines and AI is an accelerant that just allows you to use much more data and get much more value out of it.

So Matt, you just said we're excited. I will throw my editorial on, which is yesterday, Matt and I were on stage together. And it was kind of humbling because I probably had a follow-up conversation that went the same way 100 times, where someone came up to me and said, "Hey, I really enjoyed your presentation earlier. I said, well, thank you. And they said, Matt was amazing. And they always talked about 2 things. Number one is yesterday, he did this demo. It was a live demo. And what people were blown away with was that he did about 8 weeks of work in 8 minutes in front of everyone. And it was like demonstrating the processes that literally every marketer in the room has been doing for 30 years with armies of people, and he did it in 8 minutes. And secondly, we had a bunch of tech platforms in the audience, ranging from agency partners like Publicis to major DSP partners to a lot of the big cloud providers. And the thing he said that about enabling creators is a new thing for us because for a decade, our users have been captive to LiveRamp technology. They've been constrained by it. And now we're saying, you build whatever you want around it on top of it using our APIs. And so already, those partners are seeing opportunities to work with us in ways that they never thought was possible before.

Yes, it's a great tailwind. The -- it allows us to say we are open for builders, and it is drastically lowering the bar for who can be a builder, right? There's always been that sort of middle segment in every company in LiveRamps and every one of our customers and every one of our partners. You have the business and salespeople all the way at one end. You have the software engineers all the way at the other end, but there's that huge group in the middle of those people who are better at Excel and Excel formulas than anyone else. people who can think analytically in a data-driven way, but they couldn't actually build real product. That's changing. That whole middle group now can actually produce things and those things that they're going to produce need data. And so that's a great tailwind for us. And so in order for all this to work, this is why fundamentally, that differentiator that we have, our moat is the network. And so us basically telling the world, the agents, that Agentic Workflow that you're imagining, the data it needs is available. It's available programmatically, if it's able to get it. But it is only there and available. That network -- we've only earned the right to that network by being able to connect it with Identity, it only matters -- data from disparate sources only matters if you can join it. It only matters if you don't have to first say, all right, which cloud or which data schema format should we use? Those are blockers. And so we've abstracted away those differences. It doesn't matter your cloud, doesn't matter what your data stack looks like. It all just works. And the only way people are going to connect it is with Governance. If we can guarantee or those teams can guarantee to their leadership team, to their legal and privacy and Governance team that the only thing that will happen with our data is exactly what we said we can. There are technical guarantees that prevent anything else from happening. So this doesn't have to be because there's a contractual relationship. It doesn't have to be because there's a neutral third-party Switzerland that everyone trusts. Rather, there are technical guarantees that the only thing that occur is what everyone handshook and said, this is -- we're okay having this happen. And because of that, we have the Network Scale. We get to say the retailers who have the data you want are connected to our network. The place where this campaign will run is connected to our network. And so we can put all of these things together. And so I can't wait to be sitting here a year from now where we can all think back to some of these things and say, remember when we were at the beginning innings of all of these things because it is going to be amazing to watch how much more efficient and how much -- how many greater outcomes our customers can get from the potential of data. Everyone's always known there's more potential in the data. We're just limited by how many minutes there are in a day to extract it, and that ceiling is growing unbelievably.

I wanted to ask, you spoke a lot about what you can do with your data, what you can do on the back-end with all these new AI tools. And clearly, you've embraced that. I guess there's still a lot of worry on our end as investors. What's the entry point to the Internet now? And how are we capturing the user with this -- if people are moving to one-click shopping or their first stop is OpenAI or Gemini or whatnot, is that still allowing you to capture that interaction? Or is the Open Web just going to die and it's going to be difficult for you to do that?

Yes. I can start out on this one. I think the Open Web is going to evolve. So I think if you look at the Open Web today, you are definitely going to see disruption in the search-side. So, as Scott mentioned, we're in conversations with all of the major LLMs at this point, including OpenAI around how that does. We are one of the partners with Google AI Shopping mode. So that is an LLM-driven workflow that we are powering with data today. So I think our view is it's going to really disrupt search, but we're going to power it just as we power search today. Google Search, Microsoft Bing, Yahoo! -- are all top destinations of ours. We believe that AI-search will be the same. I think the other thing it's going to do is it's going to cause the evolution of the Open Web, the evolution of kind of CTV and social. And as the ecosystem disrupts, I think we are super well positioned for that. So we have over 300 different ecosystem partners today that we actively work with. If that becomes 500, that reinforces our Network Scale and our Network Moat. So for me, I think the disruption of the Open Web, like there's definitely going to be some websites that don't exist a few years from now. I think it's actually going to be good for consumers. We're going to see a more premium Open Web. We're going to see a better one. And I think we're going to see LiveRamp really thrive as we connect all of the different sources of kind of consumer-content out there.

Complete non-sequiter, but I just want to say it. One of the great things about working at LiveRamp is I think we have amazingly talented and committed people. And this man to my left, Travis, he is here at RampUp and he's having a gazillion meetings. Tonight, he's getting on a flight, taking a red-eye to Mexico, so he can get married on Saturday. So he's very unpopular with his future spouse, but I love this guy.

This is Shyam Patil, Susquehanna. So I wanted to follow up on Amit's question because it is probably the main question that we're getting from investors. If we kind of assume that the Open Internet does get displaced, I know you said it's going to evolve, but let's say it does get displaced and a lot of the traffic activity happens within AI Walled Gardens or Walled Gardens. Can you maybe share more specifics on how you guys will maintain your value prop and continue to operate in that environment? And the transition from here to there, is that something you feel like you figured out? Or could there be some pain along the way? If you could talk about that, that would be great.

I think 100% is something we figured out, Shyam. It's one of the reasons that starting a couple of years ago, we started to move to a usage-based pricing model, particularly for new clients because we think this is going to be a tailwind. And it may be that the way we think about activations really changes because what's going to be the case is all of the AI activity is going to burn even more data. And so by moving from a -- to a usage-based pricing model, it allows us to participate in that upside. We have seen for the last decade, as you know, like the tide has always risen. But underneath that, we've had a glimpse into who the winners and losers are going to be. So we know when Meta is going to have a great quarter or a bad quarter before they announce it because we see where the data is flowing. And so I do think that we're going to see social continue to soar. We're going to see CTV, which cannot be replaced by agents, continue to soar because just Netflix is better than going to broadcast. Kind of the comScore 50 open web publishers, they're going to be the ones that really have to rethink their business. And we know already 50% of their traffic is nonhuman because they're getting scraped all the time and their content is going into the LLM feeds. And as consumers increasingly use the LLM as the starting point for their journey as opposed to going directly to that comScore 50 publisher, there will be a disruption in traffic. I suspect that group may lose. But again, like we feel really good because the explosive growth that's happening in AI and not just in terms of where the -- where someone's consuming -- consumer is looking at content, but also in where the work happens. So it could be the dynamic creative. It could be chat. It also could be like the search stuff that travels. Each one of those 6 categories will contribute to our growth. So we feel really good. We feel also, though, that we need to start to publish a metric out to you that shows that adoption. In our last earnings call, we did some back of the envelope math and estimated that about 10% of our data now is going to AI applications. And probably more so to things that you've heard of like a Google, who's just replacing their existing algorithms with AI than some of these Greenfield Applications. But we'll harden that metric, and we'll start to report back. And I don't know if that will be the exact metric, but we want to show to everybody that we're benefiting from that AI such that we can just take this worry off the table.

I can add 3 stats here to just add a little bit of context. If you look at our top 2 destinations, none of them are Open Web. So like today, the Social, CTV. Our top 50 destinations, 70% of those are CTV. And if you look at our fastest-growing destinations, we have things like Netflix in there. There's not a single open-web property in there or open-web DSP. So today, when you look at the growth of our network, there's definitely DSPs in there. I don't want in any way to diminish them, but they are not the part of our network that our brands are most interested in or the growing part. Brands are focused on Social, CTV, and as Scott mentioned, the new emerging AI destinations. That Google AI Shopping mode got a dozen customers in just a couple of months.

Well, even in the DSP space, again, we can see who the winners and losers are going to be. And I will tell you, the winners are the folks that have a lot of data. And then they're looking to us to supplement that. So like why is Amazon growing so rapidly? It's because they have a data advantage. And so you will see a mix-shift in the DSPs, and you're already seeing that play out as they announce their results.

Can I just add one, you just talked about some of the social platforms where spend is going. even as Travis talked about, we've always sent data into Meta and kind of some of the social platforms, but some of that spend kind of goes to Meta Advantage or TikTok plus us or Google PMax. We have a value proposition as spend goes to those AI-driven platforms. We're able to connect our data from the network. There's models built in that to create like a scoring that actually influences the bidding decisions in those platforms as well, which is available through us because the data comes from us into those platforms as well as you actually have to measure the outcome. And those platforms can't measure the outcome. It only happens off platform kind of through our network. So we can both influence how the spend happens in those -- where those big dollars are kind of going as well as measure the outcome and optimize the next step in the whole journey. So the whole point is that we'll have a continued growing role even if kind of spend continues to kind of grow in those platforms.

It's Mark Zgutowicz with the Benchmark Company right here. Scott, you talked about an expanding AI Ecosystem. And I'm curious if there's a linear relationship to that expanding ecosystem to revenue. Meaning is there incremental data that you get usage that you'll be selling as these expand? Or are these incremental AI partnerships just going to be table stakes? And -- or are there other subscription tiers that can evolve here as well as you expand your AI Ecosystem?

Yes. We think it's going to expand. If you saw me speak yesterday, I had -- I talked about how data models are built in one section. And it always looks like an S-curve that when you start cooking data, you get very little lift initially until you get to sufficiency, in which case, you get a massive performance improvement, but then you start to flatline because the incremental piece of data just duplicates what you already know. And this is the problem with the LLMs is they are all built on publicly available data. And so they've already started to reach the limit of how effective they can be. And I had a quote from Satya Nadella, who talked about this, but Larry Ellison has talked about it. I mean the CEO from Databricks has mentioned this as well. Everyone is seeing this play out. And so there's a war. It's a war for signal because what everyone recognizes is the best data isn't the stuff that's in the public domain that's commoditized, it's available to everyone. It's all the stuff that sits in the private domain. And so if you can control the usage and limit it to enterprise use as opposed to allowing it to go into the public LLM, there's a tremendous opportunity here. Think about it as the fuel that feeds all those AI models. And so like when we talk about the growth of AI, like for anyone to think that, that doesn't mean the entire private data sector is the biggest beneficiary, like that's the only reason the LLMs work. And so I think in ways that we probably can't even anticipate right now, like I don't know like in measurement versus search versus chat, which one of those is going to take off most rapidly. They're all showing explosive growth. And my guess is it's going to be a decade of crazy growth driven by AI.

Aaron Samuels from Susquehanna. I had a question yesterday at the keynote, I think Matt and Scott, you had the 3 things that are going to change or evolve in the future. And one of them was the Data Marketplace going from licensing Audience Segments to predictive -- licensing the models to do Predictive Audiences. And then it kind of ties into what Dave was saying about how you can measure in the platform, even if the spend goes into these platforms, you're still involved. I'm curious how that changes financially, how that -- how you charge for that, how that accelerates the business? What does it all mean really strategically?

Yes. I'm happy to start. The uptick in the marketplace is additive. The business that we have in the marketplace today, the predominant usage of people licensing or getting access to segments of users that are good acquisition targets for them that they don't have already in their own system, that is not going away. That is healthy and growing or people licensing data conversion signals for measurement if all of the conversion for their product doesn't happen in their own store or app. And so they need to use a partner to get access to that data. That isn't going away. That is growing and healthy. These are just new expanded use cases that we haven't had historically or I should say, happened in very, very small pockets versus something that was becoming much more mainstream and is a category worth even talking to you all about, which is, hey, now I might want to license data simply to train or tune a model. And maybe that's in marketing use cases and maybe it's beyond marketing use cases or the licensing no longer of a dataset, but of a model itself. I may -- I have a particular new product category, and you can look at this in health care, in financial services, in automotive, where I think I know who I'm trying to reach and rather than get a list of 10 million identifiers to say, "Hey, why don't you go show them this offer? I'm much more interested in someone who has built and trained a specific model for what it is I'm looking at, and I want to license that model for a variety of use cases. Or you saw the likes of Chalice and Newton. -- maybe I don't have a Data Science and an Engineering team myself. I don't know what to do with the model. I actually need the whole application that is powered by a model. I still need a 6-step wizard. And so someone being able to just more easily go in and license that application. And so in the end, it all ends up still just being utilization and usage based of the underlying data. These are just new consumption modalities that make it easier to meet our customers where they are. And so again, these are just expanded use cases.

And from a pricing perspective, similar to how we price Data Marketplace utilization today, it's a take-rate paid to us by the data provider or seller of that data. And so maybe to connect to Mark's question, I think our Data Marketplace is where in the near term, there should be a very linear correlation between kind of AI usage and revenue growth.

Clark Wright, D.A. Davidson. I wanted to touch on identity and your guys' Identity Graph. As we think about -- the prior questions have really been asked around the consumer-interface piece, but understanding the changes that are happening to identity and who is you. A lot of the conversations I had yesterday with customers were trying to understand how your guys' Identity Graph is inherently changing in order to account for that change. I would love to understand your guys' thoughts on that matter.

Yes. I'm happy to go first here. There's 2 pieces to it. One is Identity is always a means to an end, right? The end-result of Identity, resolving Identity or finding out who you is doesn't accomplish anything. You're doing it for a reason. You're doing it to impact personalization, to impact Measurement, what happened. And so that is the outcome. And so what all these surfaces need, whether it's a closed Walled Garden, whether it's a Social Platform, they don't have all of the information in order to get to that personalization and in order to have it be measurable, whether that is a person or whether that is an agent on their behalf. And so our Identity Graph in and of itself, it isn't changing beyond us just having far more signals that are getting fed into it. The thing that's changing the most is it used to be that the way we power this entire ecosystem, is by making sure we can have all of these identifiers be joinable, so that on that surface, someone could go do that look-up to do that Personalization or add that row to that log that would resolve later for the purposes of Measurement. It is just now how much more information beyond just the identifier that can get passed around. And so this is where you see industry standards emerging. And so you've heard us talk about what we used to call User Context Protocol. It's now Agentic Audiences, which is something that the IAB is powering. We developed it and gave it to the IAB to push for industry standards. And that is around how the ecosystem can now be passing more than just identifier signals to add far more context to lead to that more Personalization and to that Measurement. And so it's less about what are the rows and columns changing in our ID graph, and it is much more in addition to the ID graph, what other context signals can get passed around in a governed way, and that's what's evolving the most.

Alec Brondolo from Wells Fargo. I think a lot of what you guys demonstrated today is improving the ease-of-use of the tool set. I also think that you've implemented a new pricing model that's supposed to lower the barrier-to-entry for new customers. Scott spoke to a 13% to 15% growth, I guess, aspiration. Can you maybe help us think, could new customer growth more significantly contribute to that 13% to 15% than it has to growth over the last several years?

Yes. And I'll just -- I'll clarify because what we've stated publicly is 10% to 15% revenue growth. I think, Scott, just kind of upped to the ante, but in the near term, we're targeting to get back above 10%. And as we mentioned on our past earnings calls, we think we've got clear line of sight to that in the near term if we can continue to execute in the way we have over the past handful of quarters. To your specific question, I do think New Logo has the potential to be a much bigger growth-contributor in the near-term than it has been over the past handful of years. In our most recent earnings call, we talked about net new logo growth or Customer Count increasing pretty meaningfully, and that was a nice trend-change from what we've seen in prior quarters. That was in part driven by the new Pricing Model, which introduces a lower upfront fixed cost and is particularly appealing with mid-market customers. So if that continues, we do expect New Logo to be a more meaningful growth contributor.