All right, we're rolling. Hello, everybody. Thank you for coming. My name is Matt Koss from Morgan Stanley U.S. Internet team. Thrilled to be joined by Jeff Green, the chairman, CEO, and co-founder of TradeDesk. Thank you for being here.
Delighted to be here. Thanks for having us.
Of course. So quickly run through disclosures. For important disclosures, please see the Morgan Stanley Research Disclosure website at morganstanley.com slash research disclosures. If you have any questions, please reach out to your MS sales representative. With that, maybe let's just start with the big picture here. Over the past 12 months, you've kind of described 2025 as a recalibration year for TradeDesk and 2026 is a disciplined reinvestment year. So as you step back, where is TradeDesk in your long-term growth
trajectory? Yeah, so of course you always have to start with the TAM. So we always remind people that we have a trillion dollar TAM and that most of that is centered in the top 1,000 advertisers, if you will. And that over the 15 years that we've been a company, just over 15 years, and the 10 years that we've been a public company, we have really serviced the S&P 500 or the Fortune 500 brands. It's not uncommon to have sort of a Navy SEAL team go out and win that business and service that business. And as you get to sometimes something like a $3 billion revenue new business that you have to recalibrate to build a machine and what we've done in 2025 is recalibrate to essentially take especially our go-to-market team spread them out so that they're assigned to every individual account of course during our time as a company there's also been some pressures on ad agencies and the way that dollars flow is different the way that decisions are made is different and so in the same way that the market is moving so has our business moved to be something bigger. And we need to build that into a scalable machine to go from $3 billion to $10 billion. And that required some recalibration. So that's what we've been doing in 2025.
Great. So you've described how spend across CPG and auto has been more measured recently, and categories like tech and pharma are stronger. How should investors think about the causes of the vertical mix and demand trends that you're seeing? And are they more cyclical, or are there
are things that are changing structurally? Yeah, it's almost all cyclical. I would say that many CPGs and autos were not comfortable saying that tariffs and political pressures slowed down their business in 2025. And there were a fair number of headwinds, but 2025 was also a year where we added more supply to the advertising ecosystem than in any year in the past. and by doing both of those things essentially having the headwinds and and pressures on their business to the brands that have really built their their brands from traditional advertising and especially television advertising and it's really important to remember as much as everybody would like to just have the end of the funnel where people type in buy mercedes-benz and then buy a mercedes-benz without any ads that had to happen before that no one has ever done that that didn't first see an ad. There was a reason somebody typed that in. And that ads and brands are built with emotion, a connection to human beings. And television and radio ads do a much better job of that than little black and blue characters that only come after the real advertising has done its work. So those brands that are facing headwinds have often had dilemmas like oh i have to reserve hundreds of millions of dollars to make certain uh that tariffs don't affect my business and then i have to convince and one cbg said to us we had to convince the white house that we couldn't get these products in the united states and then in that environment how do we do more with less and so 2025 was them figuring out how to do more with less how to measure how to make certain that you give some credit to the top of the funnel if not the majority to the top of the funnel that is being taken by people almost exclusively at the bottom of the funnel and build the engines in a digital world and at the same time when the supply is exceeding the demand that does make it a buyer's market that is very good for us and for the biggest brands in the world but you have to parse through all the new supply assigned value to all of it in order to make it as effective as they'd like to and that's what we've done very well from them but in an environment where they're just counting pennies and trying to be cost conscious. Part of the reason why in our earnings we highlighted CPG and autos as those categories that have been sort of affected the most is it's the first time that we've ever itemized one of the, you know, two of the 11 sectors of the economy as sort of standouts. Because if you were to extract those, our business would be doing as well as it's ever done with all the other categories which is part of the reason why we call it out because we just wanted to point out that this is something that is cyclical in nature this is not a structural issue particularly with us this is something that is cyclical to our clients that we're helping them whether helping them get to a better environment so that we can of course help them get back to brand building which is the only reason why brands like coca-cola and nike and others have have endured so long as they they're the best in the world at creating brands right
so you've made some uh significant organizational changes over the past year so i wonder as you as you look at the impact that that's having uh what is your assessment of the ability to translate that into stronger execution and how is it positioning the business for renewed momentum
yeah so it goes back to that sort of navy seal like approach that i was describing before so when you're a smaller company, the way that you're scaling is you get a few, you know, a dozen people that are really good at doing deals. And then even when you have hundreds of other go-to-market team members, they're mostly executing after those deals are done. But as you grow from partnerships with Fortune 500 companies that are instead of measured in 10 million or even 50 million, but instead hundreds of millions of dollars, and in some cases billions of dollars, The decision makers spread out in those organizations. It's not just one media buyer. It's their boss, their boss's boss. And in fact, in many of these businesses, some of the biggest businesses in the world, our relationships go from the CEO all the way to the bottom. In order to service brands like that, you have to spread out. This has to be a system. So while we had some of the best deal makers in the history of our space, and we also had a pitch that is centered around objectivity we don't own any inventory we're not biasing towards our own the whether that's a youtube or whether that's amazon prime video or whether that's facebook inventory we're not biasing towards our own we're trying to help you objectively buy what is in the best interest of you that pitch is super compelling however uh we have to make it at every level of organization so what we needed to do in our go-to-market is spread out and we needed to assign the right number of people to agencies right number of people to brands assign them to every individual brand from the middle bottom top make certain that we hold each of them accountable where sometimes that the decision making is is shifting because slowly over the last 10 years it shifted from media buyers in some cases to the cfo who knows nothing about the complicated world of esoteric advertising or programmatic advertising and so in that environment we've had to spread out get clear roles and responsibilities and it is working in fact our jbp pipeline is the biggest that it has ever been but it takes some time to spread out get get new roles and responsibilities new people on board help them acclimate to a very complicated business which that complexity over the long run is a is a moat it is an amazing moat to our business but it does take some time to transform and we're laser focused on taking it from that three billion in revenue to ten billion and that needs a new sort of organizational structure
and given that the decision making is more distributed it sounds like inside of your your advertiser customers than it's been in the past i guess uh is there not it's not quite the same as the funnel but how do you know where it's working if you have these different touch points across the company i mean you know you could imagine something like is there a bottom of funnel sales rep at trade desk who's taking credit for the work of others how do you account for those multiple touch points in this more distributed uh way of going to market that's a
really great and nuanced question i don't know that i've ever been asked before uh from uh people outside of the our company um there is a lot of work to make certain that everybody is earning their keep and it is by measuring both the activity but of course the output and so we're an amazing machine now and because we're a data-driven company in everything that we do i mean to at the core of our value proposition is of course that decisioning and that decisioning is basically like the simple way of explaining the evolution of advertising for the last 50 years is that it used to be a spray and pray methodology where you called and asked what the rate card was to now a data science driven ecosystem where while inventory can be added easily in a digital environment so can decisioning about what works what's likely to work so that you can fine-tune exactly where you spend all of your money because that's in our dna that's how we got here we're data-driven people We're applying that exact same rigor to all the people that are servicing those companies to make certain that everyone is earning their keep as we spread out. So the harder part was actually creating the roles and responsibilities than it is to create the data and the accountability. It's more of a cultural shift because we've been so mission-driven, so vision-driven in what we do that adding the mathematical rigor to our own operations, to our teams, to apply that with their passion and their vision has been the transformation that we've been through in the last year, and it's working. Fascinating.
Maybe shifting over to AI. It's a lot in there, but if we look out a couple of years into the future, and these tools continue to proliferate what ways will the ad industry look different and
in what ways will trade desk operate differently yeah so uh ai is such an amazing opportunity for the trade desk and for the programmatic ecosystem and for the open internet and i i want to just frame a little bit the the the challenge uh and and the opportunity because agentic has become such a buzzword that we often say a GenTIC is going to change everything and we don't need websites anymore. I think there's an interesting exercise to think about. Okay, a GenTIC will book all my travel from now on. So if you go to an LLM and you say, book a ticket, then it says back, what airline? And then you answer the question. Then you say, what day? What time? And you go through a series of dialogue that you basically just went through the web form on Expedia. and in fact at the end of that agentic experience you do that like 10 times and you're like you know what would really help this is if i had a visual and now you just rebuilt the ui for expedia so for doing simple tasks like that agentic is in my view not an upgrade now in our world we have this really complicated thing like if you were to talk to our ui designers they would describe their problem for the last 15 years as being, if you want to take a campaign from $500,000 to a million dollars, you have to change something in your settings in order to enable more budget, in order to spend more. And you could say, oh, let's change the frequency cap from showing two ads every 24 hours to showing four ads every 24 hours. Or you could expand your site list and say, instead of just showing on Hulu, let's add Netflix to the equation. Let's also add other parts of Disney to the equation. Or let's add some Spotify ads. Or you could say, instead of just showing the first ad in the commercial break, let's show the fourth. There are literally thousands, and in fact millions of permutations of changes you can make to make that change. And so if you think of agentic AI, if you want to simplify it, to an API or a UI that can reason with you. Not just give you options like in our Expedia case, but actually reason with you. Because there are so many possibilities, Agenda can change all of the ways that decisions are made. So what we've been enabling for the last 15 years is people to make those decisions, and we try to give them as much data as possible to make informed decisions. That data has been really overwhelming. So now, sitting between our UI, which was really hard to give them all the data we wanted to, and the user who couldn't keep all that data present in their minds, now can have AI as an aid to help them in the middle of them and can reason with them and give them those options. We can make recommendations way more easily in an agentic AI environment, which is why we've recalibrated our entire system to inject AI throughout the whole thing. So our business is more conducive to AI than almost any other software business that I know of. For instance, we look at 20 million ad opportunities every single second. We essentially have to make decisions in 10 milliseconds based on massive data sets from the past. And any brand, even the biggest in the world, are only going to buy a few hundred of those 20 million. How do you make those decisions in such a really rapid time as effectively as possible? I don't know that there's a better job for AI than that one. If you then say, okay, well now I know what I want to buy, but there's many paths to get there, meaning I could buy it through this reseller, I could buy it through this seller. This one has this markup. This one has this user identified. There's a whole bunch of different ways to get to that same thing. Again, creating millions of permutations that we have milliseconds to choose between. Again, an amazing assignment for AI. We've gone through every aspect of the way that we make decisions for our users. We've broken them into pieces in an infrastructure we call distributed AI. And we basically boiled them down to individual tasks so that we're not trying to boil the ocean with any of these, but individual tasks with a single question and an entire team trying to inject AI into it. And then there's inputs and outputs to every module of the system so that we can check each other's work. Because every AI system has a garbage in, garbage out problem. What if you're wrong? What if you make a bad decision based on bad data? How will you know? When Gemini or ChatGBT give you the wrong answer, you often make a bad decision and don't even know it. What we're trying to do is prevent any of that happening when we're repeating this process again and again by having all those checks and balances and all that distribution. So I take the minute to sort of detour and go into the details a bit on the AI simply to just underline what a significant driver of our business AI has already been and will be for as far as we can see into the future. Got it. Maybe shifting over to the
streaming ecosystem for a minute. So obviously CTV and video broadly has been a major driver of growth for Trade Desk for years. As more premium inventory becomes decisioned and authenticated, how is your transaction mix evolving? And what does the next phase of CTV ad monetization look
like? Yeah, so still the minority of CTV ads are transacted programmatically. And of those that are transacted programmatically, often the decisioning is very low or very weak. And so you're seeing, especially because of that surge in supply in 2025, that surge has made it so that it's not really possible to go get more revenue by just throwing more tags in the commercial break so you know now showing ad one of two was the norm of 2023 and now showing ad one of four or five is the norm of 2025 and we've we've reached the point where you can't really add to it without affecting subscribers without affecting ad efficacy it has some negative effect and so what all of them are doing, they're saying, hey, I'm too worried about any decisions that would negatively impact one of those things. So instead, I've got to figure out how to turn a $10 CPM into a $12 CPM, or a $20 CPM, or a $30 CPM. And the only way to do that is to enable decisioning for companies like us and the advertisers that we represent. So by making it biddable, by making it so that we have some metadata to make decisions. And incidentally, the data that we're looking for to make decisions is typically the advertiser's own data. So for instance, we're not asking Netflix to send us all of their data. Instead, we just want to make it so that we can use our data as we make the decisions and the advertisers that we represent. So more and more, they're all moving towards a biddable environment where they're enabling this metadata, where they're being more reassuring about exactly what we're buying. We want them to describe well what they're selling because there are so many companies that are not premium that are trying to pretend to be premium by obfuscating. And the only way for the premium to distinguish itself is to be very clear in what it describes, including like how many ads are in the commercial break. This is one thing that Netflix has done very well. They've kept the ad load light so that it is a premium and they've gotten premium prices as a result. Now to leverage that premium price or that premium environment optimally, I think it's best for them to make it biddable and to create competition. And I think
they're on their way to that. So talking about retail media for a minute. So as that landscape continues to mature, how do you see it evolving as agentic commerce develops? And how does audience unlimited expand the monetization surface for retail data yeah so uh i i think maybe the
single biggest opportunity that we're looking at to to elevate our business in in 2026 is all related to retail and that's related to retail data and the effect that that can have on measurement and those two things are just sort of creating energy like electrons chasing each other just creates tremendous energy for us to transform not just our business and not just the open internet but really all of advertising and the reason why I believe retail media and retail data in particular transforms that is that you do want that data to be used to figure out who to target So if you're saying this group of people is 25 times more likely to buy the product in the store, of course you want to advertise to those. You want to invest in those expensive impressions in order to buy them. And then if you do buy them, you will measure it using that same conversion data that informed you to go target them in the first time. And then it creates this virtuous cycle. What's happened in many platforms like Facebook or Google is that they've been taking credit for everything that happens up Funnel, and that's because of the absence of this retail data. Now, many of the retailers have been trying to monetize that data by joining it with their own ad network, by creating an ad network and selling some inventory with it. The challenge with that is you mark up the media so much that you're only going to get a sliver. So the average CPG who is spending billions of dollars, they will only buy a 10 million dollar buy when they're turning a five million dollar media buy into a 10 million dollar simply because of layering on that data so what our pitch to the retailers around the world and we we've aggregated a double digit percentage of all the retailers in terms of transaction data around the world by aggregating all of that data we make it possible for them to get out of that 10 to break that 10 million dollar ceiling and instead layer the data on the billions of dollars that these cpgs are spending across all of it so what audience unlimited is is a product that we offer to the biggest brands in the world including and maybe especially cpgs and auto who will then uh we've done to third-party data especially retail but what spotify did to music which is make it so that instead of buying one by one and making it hard to find we make it so that you have all you can eat and as soon as you have all you can eat then you're in the business of creating discovery again just like just like what spotify has done like if you think of the uis when everybody was selling at 99 cents a song it was much less about discovery and more about just putting things on the shelf in a data world where there's data segments that are just there's just billions and billions of data segments uh it becomes hard to scroll and click through all of that so uh we finally reached a place where uh the data ecosystem can can be what spotify has done to music and make it so that all the friction goes away i believe for the first time we'll have price discovery in the third-party data ecosystem at scale, and that will transform our business. It'll transform these retailers' business. It will make it easier for them to compete with Amazon. It will also make it so that in cases where we are competing with Amazon, all these retailers pooling together obviously are in part doing that to compete with Amazon. It makes it very clear where people are buying and producing results. And so I think we've created a way to accelerate our business in audience unlimited and the measurement products that will come out of it. Maybe sticking with the
theme of competition, I think you've said for a long time that you believe the end state would be five to 10 scaled global ad platforms with room for really one scaled independent DSP. So I guess how is the competitive environment evolving relative to that expectation that you've
articulated historically so probably the part of it that i've been the most wrong about is the five to ten part so far where there have really been fewer than that uh and it's largely because that walled garden strategy of trying to essentially take credit for the bottom of the funnel as well as uh to use really low cost inventory where the cost of goods sold is almost zero typically user-generated content to say because i touched the consumer with regularity i sold all of your products that's in a way what the largest success stories in advertising for destinations and i say destinations the trade desk is not a destination you don't advertise on thetradedesk.com but you do on instagram or facebook or youtube for the destinations that have had the most success in advertising. They have essentially unlimited inventory and have taken credit for all of that. At times, in order to attract more dollars, they've said, hey, and we'll also buy the rest of the internet. And not surprisingly, it's not a priority to any of those businesses. So I've been surprised by how long some of them stayed in the game. Like Facebook got out relatively quickly. Google is retreating some now. Like they're less of a competitor than they were a few years ago. Amazon has just kind of gotten started in that. And honestly, I think they have more conflict than any of the other two, as well as the advertising DNA is not the same as those other two companies. So I've been surprised by how few other companies have joined this space. But I stand by all the things I said before, which is it will be less than 10, and that the independent, we'll get the lion's share because objectivity matters more today than it did yesterday. And it will matter even more tomorrow. And as we add more supply and more conflict of interest and more competition in the world in general, the need for an objective partner to help you buy effectively is going up, not down. And so we've always been playing the long game, trying to win trust, maintain trust over the long run. And we continue to do that. We are unapologetically premium in the sense that we are not trying to build the cheapest platform. We are trying to build the best. And we make certain that every single feature earns its keep. It adds more value than it costs. And we measure that very carefully to make certain that we are always providing obvious excess or consumer surplus to the people on the other side of the platform.
So the conflict of interest you're referring to, I guess it really is a function of the existence of this vast amount of walled garden inventory out there on the Internet as opposed to as opposed to open web. And obviously, you, Trade Desk, have been champions of open web for a long time. As you see that walled garden inventory continue to grow, how do you see the competition between advertising and open web and advertising in that walled garden inventory evolving going forward?
Yeah, so one thing that I think is just important to define, sometimes I've learned over time that when you hear the term open internet or open web, that listeners have different definitions of what those terms are. So many of you hear open internet or open web, and you think about the ads that appear in a browser. And that facilitates a very different debate, because I think everybody here would agree that there is going to be a slightly less, maybe a lot less, inventory in the browsing internet because of the chat GPTs and Gemini's of the world. Although part of Gemini's success comes from a browser, so they may not want all of that to happen. but if you look at our business only about 10 percent of our business is display advertising which takes place on those websites almost all of our business is focused on that premium like I said we're unapologetically premium the parts of advertising that are really effective are the parts that connect you to uh to emotion and to memory like every one of us here could give a jingle that you heard in a commercial 30 years ago, assuming you're at least 30 years old or at least 35 years old, to remember a product from long ago because of the way that it made you feel. So what we're trying to do is help the biggest advertisers in the world connect with consumers so that they will feel something so that they can remember that and, of course, come back and buy the product. And that's what we're trying to enable.
I'm incredibly flattered that you think I might not be 30. So today your client base really consists of the largest brands in the world. How do you think about the opportunity for growth within that current segment of the market versus the opportunity for growth down market by creating an offering for maybe small and medium-sized advertisers?
Yeah, we approach the advertising business differently than nearly everyone else did, which was most of them started with the smaller businesses and have kind of worked their way up. And we recognize that it is a very robust sort of strong fat head and a very, very long tail. So if you were to draw the 50% mark in terms of dollars, it's probably somewhere between 600 and 700. So in other words, the 700th largest advertiser is the 50% mark where a half a trillion dollars is spent to the left and half a trillion dollars is spent to the right. And so by having so much concentration there, we have always said we want to focus on that head because that is the part that cares most about objectivity. That's where the conflict of interest from all the walled gardens come into play. And their strengths have been built from the offerings that they've given to consumers. That's why they've climbed from the end all the way toward the top. But they still make all their money off of the long tail. so we always set out to design a business that can make money off of the biggest advertisers in the world we knew that it would have to have a different margin profile we'd have to think about cost of goods sold differently than they do and we would also have to recognize that we have to get to scale and as a result we have to have a meaningful part portion of their wallet and that means to build tremendous amount of trust and align our interests with them. So we think we are better equipped to win a dollar from those sophisticated brands than any other company in the world in advertising. And so of course that's going to be the core. But we also recognize that our value proposition appeals to the mid-sized businesses. And that of course can appeal to small businesses. So we've always said, we'll start with the head and we'll work our way down. Exactly the opposite direction of the way that most of the others have, which is to start small and work their way up. But as an investor, I think that's incredibly reassuring because the biggest, most sophisticated, most scrutinizing companies have stuck with us for a very long time. And if we can win them over and perform well for them, it's much easier for us to simplify and do something similar for the mid-sized businesses, which are honestly the M's and the L's have more in common than the S's and the M's, but we still say SMB because I'm sure some consultant made up the acronym. But we absolutely recognize the value in continuing down that distribution and we'll continue to keep marching our way down. But we don't want to lose sight of where that nearly 50% of the entire TAM is concentrated.
I want to close with one kind of overarching question. Before I get there, I do want to touch on the supply chain. How would you grade the efficiency of the ad supply chain in open internet today, and what are the steps needed to improve it from here?
So if I were to give it a grade, I would say it's a C minus. uh the the open internet is competing with walled gardens walled gardens has a less valuable almost infinite supply uh and that really uh less valuable inventory has still managed to do well in part because one of it's nearly infinite and two because the supply chain is really tight in the open internet there's a lot of incentive for people to make it inefficient or exploit it it's like if there's a factory producing a product on the other side of the world and then you're shipping it here it can touch 19 companies between it all of those represent an opportunity for tax or markup or in some cases obfuscation but in a digital world you're also not bound by the same rules of physics. You can send it around the world 15 times, and there's lots of ways to mark it up. So we're one of the only companies in the world that is obsessed with making that supply chain as efficient as possible. So on one hand, it's kind of disappointing that it's a C-. On the other hand, that's really encouraging because there's lots of efficiencies yet to be gleaned from it. We have to develop closer relationships, not just the relationships themselves, but supply chains, the pipes between us and the supply. So I'm sure you'll hear more about that from the next speaker in Netflix, where as they grow in the advertising ecosystem and in the business, they will build more of the tools on their side and we will plug in more and more directly as time goes on and that will have an incredible improvement or an efficiency gain for the ecosystem and those supply chains where instead of having 19 hops that go around the world again and again and again, you'll just see that truncate because it's better for the advertiser and it's better for those that are adding value to the ecosystem. So we made tremendous strides in 2025. We made more progress in 2025 than all the rest of the years of our existence put together in improving the supply chain. But the exploits were also growing at the same time because the supply-demand imbalance was so great. They have to exploit in order to even maintain revenue at all. So that created a tremendous both opportunity and challenge for us. we continue to win and make progress in it but it is a tremendous opportunity for us to continue to do it because the competitive environment that is the open internet will always be better than a walled garden in in efficacy because of the benefits of competition but also because everything premium sits outside of a walled garden that's where all of the the best of sports that's where the best of movies and television and music and news, all of that exists in the open internet.
And then just in the final moments, based on your conversations with investors, what do you believe is the most underappreciated opportunity and then potentially challenge for
you to execute on due to AI? Yeah, the most undervalued opportunity is the massive data set that we sit on top of, as well as the trust that we have developed with advertisers over the years in order for us to build this amazing AI learning engine. So, of course, we've already been developing that. Of course, it's already in production. We already use that. But every AI company is dependent on being fed data. In a lot of ways, that's more valuable than the AI technology itself, is the data that it's feeding. And those two things have to work together. so in order for somebody to build what we built with the biggest brands in the world i think the long pole is actually not the technology but actually the trust that you build and then the iteration that you build with that trust so in other words you get the data asset you learn from it you adjust and you do that again and again and again we've been doing that for 15 years. And the trust that we have built and the learnings that we have create a massive advantage. And the complexity of our ecosystem and the complexity of our technology also is an advantage in a world where the data asset is dramatically undervalued that we have. And when I say the data asset, I don't just mean the first party data that advertisers share. When you listen to essentially the entire open internet. We're listening to 20 million ads a second. You're learning from every one of those. That learning and that scale, which by the way, that pays for all the learning and listening is paid for by all the revenue that we're producing. In order to do that without massive expense, you have to have entered this game more than a decade ago. So that advantage is something that I don't think is fully appreciated. Like how strong, how wide, how deep the moat is around our business is something I think very few understand. Jeff, thank you so much.