Consolidation & Aggression

0:10 Programs in the scene real tight. RTB dreams light up the night. Adam says pivot. Garrett says scale. Together they weave one flawless tale. Tech talk. It's a Q&A. Buy side clearing the way. CEO spitting truth. No delay and owning the day.

1:02 Harvey, want anything special for your birthday? >> Just a decent cup of coffee. >> You're kidding. >> I'm serious, honey. Your coffee is undrinkable. >> That's pretty harsh. >> Well, so's your coffee. You know, the girls down at the office make better coffee on their hot plates. >> And he didn't even kiss me goodbye. You know, if I could just make a decent cup of coffee, I could relax. Oh, relax. Hey, great coffee. >> Better than those girls make at the office.

1:33 >> Honey, their coffee can't hold a candle to yours. >> What a world. >> Welcome folks. >> What was that? >> Real commercial. >> That was a commercial for like Maxwell House. Oh my gosh. Folders. >> Hey, welcome to Adel Folders. Welcome to Ad Talk everybody. I'm Gareth. I represent the sell side. >> I'm Adam. I represent the buy side. It is September 11th, 2026.

2:06 >> My office is right by the World Trade Center, which a vibrant vibrant district today. A lot of adtech in the neighborhood. I'm not there today. Don't want to be near all that stuff. But I was here on September 11th. I wrote my experience on Twitter. If anyone wants to read my 911 experience, which I never talk about, but I decided to write a Twitter thread, which according to Twitter, 25 people have looked at. But I thought I would start by talking about anthrax night. I was just telling you that about anthrax night.

2:40 Do you have any memory of an of anthrax night? >> I remember I remember two things. I remember the band anthrax and I remember >> that. Yeah. like one scare where there was white powder in an envelope that went somewhere and they thought there was anthrax in it. that's a lot. >> There's much more than that. How old were you in 2001? >> I was in middle school, so that was I was fifth grader. >> Okay. I was about 30 and there was there was a a couple of incidents of anthrax in envelopes, but then there was one Friday where there were several envelopes full of anthrax and they went to like NBC and ABC and a few other places >> and the news said they didn't have they could not say how far anthrax contamination went. Now, in my mind, it was two weeks after 911, but I looked it up and it was October 19th, so more like six weeks after 911. So, we were still very much in the aftermath of 911, which went all the way into the next year.

3:38 But, that night, October 9th, it was a Friday. I I me and a friend went out. I was a reporter, so maybe that's why I went ran out. But on 911, there were many different reactions. I wrote about this in my thread. You saw people freaking out. You saw people come. You saw people joking. You saw people crying. All the human reactions were on display. On Anthorax night, 99.9% of the people had the same reaction, which was to stay home. They were scared. It was the most scared New York City has ever been.

4:11 >> And I went downtown and walked around and like we saw ghost town New York >> in the pandemic in other cases. This was way beyond that. There was crazy nobody out and we were and the and bars were open because it was Friday. They were hoping for business. People had to work went to work. >> but we went from bar to bar where there were no patrons and we would like play pool and golf and just look for people to speak to and stayed out drinking the whole night like it was already a Friday night. So that's my reflection is why why was my reaction different than everyone else's that night? I do not know. People really thought that if they go out there might get struck by anthrax and it just to me it was like probably not and if it happens it happens. So there it was. Not 911 but the forgotten 1019 >> 2012.

5:00 >> That's crazy. I didn't realize that it was like that. I was in Texas at the time and I guess that like they really communicated. It sounds like it was similar to co and that we don't know how this is going to spread once it's out there like Uncertainty is what kills you. >> Yeah, anthrax is like you die. >> Anyone in the chat who remembers that night, whether you're here or not, >> we want to hear from you. We're talking about anthrax night on 911, just for a little variety.

5:28 >> Controversial. >> It is the most controversial night we can do, but you know, whatever. Our show is about adtech. So, let's talk about adtech. >> Yeah. I mean, it's, you know, it's a it's simultaneously our calm and like freaked out time in adtech. So, there's something that has emerged in publishing. Well, a few things that have emerged in publishing, but something that I think is worth talking about because I actually haven't talked about on the show very much, even though I wrote a Gareth Hates adtech article about it. and there are more and more threads about publishers finally figuring out how to advertise.

6:07 Now, there there's a term for this called buying traffic. >> Oh, yeah. >> That's the porative term >> because buying traffic historically means you go on Google, you type in get a thousand visits to my website and Sergey sends a thousand visits to your website from a >> or Yuri. >> Yuri. Yeah, exactly. The fact of the matter is there are very legitimate ways to advertise your content as a publisher online. In fact, all of the ways that advertisers use to advertise content online can be used by publishers.

6:43 >> you know, who knows this is TV stations. TV stations actually know the value of their own house ads and know how much it drives. So, one, we have a a product coming out for this, but but two, I have always So, this got lumps in with MFA, right? When JS first came out about MFA, buying traffic was one of the things that had you deemed an MFA publisher is most of your traffic paid. and that included paid traffic from Facebook, which never sat right with me. because Facebook's traffic quality typically when measured right when you're buying endemic Facebook even Augustine Fu will say that people who buy on Facebook not from Facebook audience network those are not bots those are humans >> right >> but the argument that it's MFA is that the ad is usually you won't believe who the number one sexiest model on our photo list is like you know not a very high quality clip >> leave it >> true but at the same time people read us weekly And I and and US Weekly is not typically labeled MFA. Yeah. I >> all media is foreign.

7:48 >> Yeah. Oh god. So much of experience. >> but I definitely believe that and by the way Chris Kane has totally changed up his definitions. He's like completely revisited this where like is there he released a whole playbook how to do legitimate paid traffic as a publisher. So that's why I'm bringing this up is the tide has finally turned. We've recognized that we need to let publishers buy traffic. And if publishers are good at this, the reason that I've always defended this and the reason that I think that this is such a virtuous cycle is that is it's well but all businesses are ARB plays in a way.

8:26 like all all buying all advertising should be a form of ARB where you put some money in you get more money out. the timelines can be variable, but like that's that's how it should work. >> And the reason I like it so much though is it's user exfiltration. That's what people miss about this. If a publisher, >> that word is not in my vocabulary. >> Exfiltrate. >> Yeah. Exfoliate is and infiltrate. Is it a mix? >> We're we're springing the users from Facebook. If you can get someone to click an ad on Facebook, come to your website, and then turn them into a subscriber somehow, what you have done is you have harvested a Facebook user, right? That's what you've done. You've pulled someone out of a walled garden and brought them into the open internet.

9:14 It's awesome. We should be like, so I'm like I'm honestly The first thing built was to try and help publishers do this. It was a lifetime value measurement system to try and help publishers do this. And every publisher I talked to about it was like Like literally just just like we don't we don't do this. >> we don't know where we would begin. >> Kakowski launched a startup. >> Kakowski. >> I did. I saw that. Same idea. Yeah, >> I know. Well, yeah. So, he's trying to like optimize the on-site experience to maximize LTV, which is, you know, a good idea, but it's going to be really hard to get into big pubs. And also measuring LTV requires like a significant amount of tribal knowledge. like that's that's that's grotesque tribal knowledge. Like this is like I was a pickpocket when I was in high school so I could unlock this door for us to get in here. like that's the type of tribal knowledge that it takes to measure LTV because because it's really gross, right? There are so many systems on the page and Google's monopoly caused this thing where no one knows what Adex is bidding.

10:15 >> Yeah. >> How you going to how are you going to measure LTV? And and we can do it, right? Like I know how to do it because I had to learn how to do it. But how are you gonna measure LTV if you don't have access to the bids? >> Yeah, you don't. You have to estimate it and you have to be if you're going to do a good job, you have to be really good at estimating it. and you have to deverage it per user and all stuff. But I digress. I certainly >> what's changed? What's changed?

10:41 >> You know, I first rolled this out as a product three and a half years ago. >> Wow. >> I think what has changed is AI. And I think that there has been significant panic inside of publishers where they have started to think how am I going to survive? >> What are they doing? >> And so what so what are they doing? >> Are there new tactics or they're just you know doing Facebook ads now? >> I think so the new tactics most of them have rolled out some sort of subscription function >> and we're still on the bleeding edge of this. This is still really early. I'm just starting to see it talked about again. so I commented on a post by Scott Messer earlier today. there I'm starting to see people talk about this is something that publishers are doing now. And what's going to move the needle on this is when a few agencies emerge who do this on behalf of publishers.

11:29 when there are a couple of agencies who pop out and say we know how to do this. We will bundle this up for you publishers. We will teach you how to build a subscription business and we will go out and we will buy the traffic for you and you'll give us the budgets and we'll take our percentage. >> You know it has to exist. >> What? It's another thing going back to 2001 again. Do you remember RSS feeds?

11:51 Do you remember how you I do you used to create your own scroll? Right. Facebook completely loved RSS feeds. >> Yeah. Facebook replaced this. But if you're too young to remember, if you liked the web page, you would click an RSS button and you would use an RSS reader. Google had one. I think I used Google's, but there were many. Me too. >> And then it would it would c curate out of your own choices what today we would call a feed. I guess we called it an RSS feed. So you could you could find pages you like and it didn't matter what it was, right? Didn't matter what format the page was in. It was mostly blogs at the time, but you could do any sort of page. And today, I think you would have substacks and beehives and regular blogs. And >> you know, I had this exact thought.

12:35 >> It could be so much better now, right? >> My buddies and I had a thought. We we play this game called $100 startups, which are like startups that like are worth $100. >> this would be a nonprofit. >> Yeah. Well, but a thought here and I' I've thought about this like a lot and I would I would not build this because I don't know how and I wouldn't start. But >> and it would fail as a businessman.

12:59 >> I think it would fail as a business. But I do notice that there is so much has migrated to email and email addresses have be basically become disorganized RSS feeds and that doesn't make sense. So you should have a you should have a RSS feed is a million times better. The experience is better. It can be ad supported, right? You could put ads in the RSS feed and revenue shares. >> That's awful. >> >> exactly. It's awful. So, how hard is it to build a plugin that sits on top of your Gmail and assembles all of your contentbased subscriptions into a feed in a different user interface that's pleasant and nice to use?

13:37 >> Yeah, I know guys working on this. Bill Simmons is working on something like this. Oh, seriously? >> Yeah. >> I this company should exist. I always thought like this experience is so bad. I click to open an email. I read it some. I normally click it to go to the website because I do that. And everyone who listens to this should click and go to the website. Give the publisher the money. If you're reading their content, give them the money.

14:01 >> It's a bad experience. >> Have to remember these URLs. >> No. When you when you Yeah. Get in the email. Don't just don't read the articles in your email. click through to the website when you open it. >> there was a time and there was a time that Facebook was a better replacement and me as a newsreader. Like when Facebook had news, you could actually cultivate a pretty good Facebook news feed that had, you know, a mixture of news from the web and news from your friends. Like that was the original idea, a Facebook news feed. And after they got in trouble for tilting the election to Trump, they got they stopped having news on the site. They didn't give a Yeah.

14:41 >> You know, I really think that this there has a problem. Yeah. Marley's bringing up Apple News, but I but Apple News is too curated and it's like it's it's I I >> I think it is >> too much of a controlled experience because Apple has this obsession with controlling experiences >> and they publish it like they are the publisher >> and they are the publisher. I don't like that. a an experience where publishers can push content into it and they share in the advertising revenue that comes out of it I think would work really really well.

15:12 >> How did RSS feeds work? Who was that actually what what I mean I guess someone like Google offered it. They didn't make any money from it and it just assembled all the content into a feed. >> Yeah, I think that it cost them money to maintain their RSS reader and >> it was just rerendered in HTML. They would just take the raw HTML and rerender it in the feed. no. I think that they would it would the RSS feed had like the the text abstracted like a like the way a crawler would and like it would render it in a a unified experience with like minor changes per article like you there was an RSS feed like spec. I'm sure it's an XML feed or something like that.

15:50 >> that's right. >> Yeah. Yeah, they were great. >> Things in a Kindle. >> Yeah, I loved my RSS feeds. Everyone I knew who like was a hardcore internet person curated their RSS feed and enjoyed using it. And like so many people lamented it when it got shut down. because it was a great idea. It's a great idea for an experience. >> It's weird. Yeah. I always think of technology as like a a ladder. Everything improving, but there's something really good that just didn't make it.

16:17 >> Yeah. That's so strange. I feel like it was a money thing because because it was free, right? and and Google did not monetize it as far as I knew. like they they didn't make money from it. Yeah, Marley. >> That's called capitalism. >> The new sound bites courtesy of local park slope actress Atari. >> Yeah. I I really wish that they had somebody had found a way to make money. Maybe they'll bring it back because I think the whole Substack thing, right, Substack, Beehive, all that stuff, there's there's too many platforms.

16:51 There's it's it's not a it's not a slick user experience to have things jammed into your Gmail. Like gross. >> And the format of of scrolling through a feed and seeing videos like that's the format. Like that's how >> that's how my kids spend half their lives. Like there's no question that is that's how people consume content now. And if in a feed, you know, it's video, but there's no reason they can't have text. Some of them do have text.

17:16 >> I mean, to be fair, I've I'm like pretty Twitter converted at this point. and I consume a lot of content in Twitter and >> yeah, >> I think that they've really nailed the experience of loading websites in Twitter while keeping you in Twitter. Like they did a really good job with it. >> Yeah, the Twitter browser. >> The Twitter browser, but the way that it's integrated is really nice and the Nikita when he was talking about it consumes content feeds.

17:44 >> Funny. but the way Nikita was talking about it was like this is going to make it so that when you post a link, we don't dep prioritize your post because we're not worried about you leaving Twitter. because we want to keep you in the app and it's this great middle ground and I I have all sorts of ideas for them for this. >> Have you noticed with the LinkedIn browser like when you try to go back it takes you through all these LinkedIn pages? It's another experience.

18:10 >> I I who is working on LinkedIn? Another one is whenever I click on an article in LinkedIn, I cannot comment on the article on mobile. It is not possible. Like literally it it's break it is a bug. It breaks. And so I >> let's save our complaints for a full stink in segment where we can we can unload a whole pile. >> It's gross. But yeah, >> they'll sponsor us. >> Maybe they will. That would be great.

18:36 but yeah, publishers are going to start to buy traffic soon. We're going to see some agencies pop up. going to happen and it's going to be great because we're going to see growth in the open web because if the open web stops thinking about itself as a lamprey on top of Google's search mouth that just gets the drex yeah lamp hang on they they eat what's left over you know and >> it's more of a raora >> raora that's what I mean sorry lampries are something else you're right do lampre eat what's left over >> I don't Oh, my my son likes this.

19:10 >> I know. They're gross. >> animal called a ten wreck. You ever seen a tenre wreck? >> No. What's a tenre wreck? >> It's like this little porcupine guy that only lives in Madagascar. >> Yeah, that sounds cool. Madagascar. Madagascar's weird, man. >> I could see Marley looking it up. They're cute, right? >> Yeah. >> Cute. All right. Well, I'll tell you what's going on on the buy side. >> Let's see what's going on on the buy side.

19:33 >> Maldo, could I have my first buy side graphic, please? So chat GPT put out a a gentic ads manager. Not a huge piece of the market in chat GPT. You know the whole ads in chat market is about five billion of which chat GPT is supposedly capturing a billion already. so not that big but this is what they decided their format that it's going to be a textbased chat format for managing ads. Is this been promised many times and not fully delivered and looks like Chad GPT is ready. So, does it does it take on a night? Who knows? But then the other shoe drops. Let me see. I just looked this up in club to try to get the date that it came out. So, let me see.

20:24 >> I have thoughts about this one, too. >> This has been out for a year. It didn't is not on my radar. Go ahead, Maldo. This has been out for a while, but what's changed is that Amazon is making very aggressive deals with agencies and publishers to set up Amazon Bedrock agent core as a a master UI for managing multiple agents, right? So the problem every big advertiser and agency faces is if OpenAI has a chatbased interface and Google comes out with a chatbased interface and Amazon comes out with a chatbased interface. I need an agent that speaks to all these agents and that's what Amazon Bedrock is for. It's a modular system for building them.

21:16 it comes with a lot of different pieces that you can add and subtract and of course it takes engineers to configure. But what's really changed on the buy side is Amazon is starting to go to these big publishers and agencies and making deals to bring in the people and build it right. So the agencies are a little bit out over their skis and promising agentic capabilities. >> There's going to be very intense competition between Amazon and Google over these builds and Amazon's kind of strike striking first. So what's the difference between this and a system that's using MCP?

21:54 >> This I assume this is all MCP but that you know there could be API connections too. >> Got it? >> I mean MCP just means you could >> the difference is just that they could be chat enabled instead of having to make an API call but if it's under the if it's under the hood it might not matter. Yeah, because I because I was reading through this and it does make sense to me that managing permissions becomes unwieldy. Like if you have 20 different platforms that you need to operate in identically, how do you do access control and how do you do things like that? The one thing that I'm a little confused about in this though is couldn't Claude be really good at the same thing? Don't you normally have like one lightning rod point where you do all of this from? And is totally wrongfooted.

22:40 >> They're like, >> they should be they should be here, >> right? Okay. >> Yeah, that makes sense. >> Microsoft, too. But Microsoft's history is waiting until the market develops and then coming in with something that, >> you know, wins on enterprise security or something like that. >> But I don't know if they're going to beat Amazon at that. And of course, Google has the best foothold into the agencies, though. Not not a huge number of friends at the moment. so if I had to bet, I would bet on Amazon capturing significant share with these agent core builds as a sort of a a bridge between what especially the holding companies have promised and what they'll ultimately deliver.

23:25 >> It's really interesting because also >> not not wrong. >> Okay. when I looked at the chat GBT interface, I was like, "Oh, if this becomes the homepage, you only need one chat interface, right? You don't need five chat interfaces. You need one, >> right? >> you need and then that one needs to be able to talk to the other ones. I using plain language is an interesting thought, but just accessing an MCP in a platform is also an interesting thought.

24:00 But I I I see the significance of having these two headlines next to each other because in theory they're kind of making the same play. Everyone wants to be in the middle. Like if you can be the one if you can be the one. If you can be the one that's in the middle, >> it's a huge win. >> Or on the top is how I'd say it. It's like it's like that game where someone puts their hand down and you put your hand on top of their hand. And yeah, I think Amazon and Google are the ones that have could be the top hand right now. Open AI is not in the position.

24:28 Anthropic if they make a play, Microsoft, sure if they make a play, but it's gonna have to be one interface to to control the others. >> But yeah, and and when do you stop playing nice, right? Like >> there's no incentive for people to cooperate in this like no incentive whatsoever. >> Oh, so you think if Amazon is you you think that if an agency signs up with the Amazon master thing that Google might say, "We're not going to talk to the Amazon agent." Like I don't think >> they do it with YouTube and DVD 360.

24:59 There's precedent for this. There's precedent like >> they can say hey you want to buy these things that we know you have to buy you will do it directly. >> I think all you can do is create another application layer where Amazon agent is saying the Google agent is saying you have to log in to blah blah blah to configure blah blah blah and like okay >> Google agent told us to off. Yeah, you must log in. I think that's what it's going to be like.

25:26 >> Yeah, I mean it's because we we have seen this, right? It's not like these companies play nicely together today. They do not. >> Well, I what I love is like not it's competition. We just haven't seen this is a this is a real threat to Google. Google's options in this if I'm using Amazon Bedrock agent core to build my master agent of agents, Google's options aren't great. Yes is not a great option and no is not a great option. And like I described like an in between thing that I think will be what they try. But what they're mostly gonna try is to compete at the exact same thing with the exact same audience.

26:03 >> Yeah. >> And that sort of competition we haven't we haven't really seen in about 10 years. I mean there's a a sad trade desk thought here which is the leverage point in these types of negotiations is always having some sort of something proprietary that you can't buy somewhere else right that is that is the only leverage these companies have chat GBT has you want to buy chat GBT ads you use this Google has you want to buy YouTube and search you use this Facebook has you want to buy Facebook you use this a neutral player even if they have the absolute best product could be iced out, right? Where if you if you have inventory that people have to buy, >> that's kind of your foot in the door.

26:46 It's a and it's a vertical integration. It's an ugly vertical integration, you know? It's it's tying. >> Yeah, it's tying, >> but no one's but no one has dominant share in the agent market, so none of it's illegal. >> Yeah. Yeah, it's this is going to be a weird one. I but I I have a lot of sympathy to you want to build the best platform that's neutral that manages all the other ones and these guys who have two businesses that they can tie together where you have to work with one of them have ugly leverage, right? This is ugly leverage. This is not good leverage. This is bad leverage. This is I'm going to hold you hostage with this thing that I have.

27:21 >> Well, I mean Tradeesk in theory could partner with Anthropic or Microsoft and get there. I mean would have starting a couple years ago would have been ideal. >> I mean that would that would be a spectacular deal. I would I would want I would want a bunch of trade deck stuff if that happens. >> That's if they want to maintain their position, right? Their position is to be on top of of another marketplace. But let's talk about Chavis' position because I this I'm looking at this as really good news because our struggle has been as a modular application. We haven't known where to build to live, right? Like we don't like we haven't really been able to say like we could build to do algorithms at scale but we haven't been able to build on how people will access the ability to build it >> the the like the the non-algorithmic part the integr layer. Yeah exactly the infra layer on top of our application and now like it's starting to clarify.

28:18 we have much clearer view of what demand is going to look like and you know it looks a little bit like there's going to be arenas where we're going to get get to compete and to the extent that trade desk is treated as an application and we're per advertiser modeling and they're some kind of pulled stack like you know I think we're in better shape >> well the market kind of came towards chalice like the market came towards the uncoupling of features and functions and the >> yeah well we're the only ones canvas.

28:50 >> I say we're the only enterprise AI company that bet on this kind of modularity being the the means and we're kind of all alone now. >> It's really it's a really new because because you can look at Cognitive, right? Cognitive is an older company. They've been around a while. there's a reason they had to build a DSP. >> There's no way. I'm sure I'm sure that was they that was them throwing in the towel and being like >> like we're tired of all these integrations and all these companies who want to make our life difficult. We're just going to build a DSP. We're just going to build it because we're so tired of this.

29:23 >> Yeah. Story before I was I was involved in them getting thrown out or banned from Trade Desk. I was the customer. >> Yeah. It's it's it's a new world now and it's still nent. That's the thing, right? Like I think that's part of why you feel a little lonely. It's still really new. The number of people who know about these architectures is still really small and the number of people who know about these architectures on the buy side is even smaller. because I know that like there are there I know that there are people at pre-bid, right, who are staring at this >> like they they stare at this architecture all day. and there are people at Amazon who stare at this architecture all day. And then there are a few people at exchanges. I would say two of the exchanges have people who have this vision like where they're like moving on it.

30:10 >> Yeah. This is the future. This is this is a because it's awesome for us, right? It's great. by the end of by the end of Q4, maybe by the end of Q3, maybe by the beginning of Q4, the Chalice models that go in ATRF will be in six SSPs. >> Really >> amazing. >> Six. Six. That was fast. >> I mean, it's lovely. It's Well, I I do think part of the reason it was so fast is because SSPs were like, "Oh my god, there's a way to us for us to survive."

30:42 >> You know who I'm changing my mind about? >> Who? >> Brian O. Kelly. >> How? So, >> when when Brian came out with like pivoted scope 3 into the agentic layer, I was like, "This is so distracting and annoying." >> Yeah. But now I think it was like he was he was giving a distant early warning of what big tech was going to do. Like he he showed his cards as you know as his vision. He wanted for whatever reason he did that or whatever he thought it was best for him but he was really showing big tech's hand and it was like the >> he he showed their ambition like what we said what Google buyer direct is what Amazon Bedrock is like the first we heard of this vision was from Brian.

31:25 It's true. I and I definitely think he was really early, right? So, he is a gutsy entrepreneur. I I remember my calls with him and he locked in on this vision fast. He was like, "This is >> this is how systems are going to interact and just went whole hog on it." and he was so early because he's that's who he is. he was he out he he he front ran the market before the market could build like to the way it needed to be and because of that right like in some regards it missed a little bit right like saying it's going to replace programmatic I I don't think there's anyone even people inside of ADC CP who would say it's going to replace programmatic I think that that that chip has sailed and that people don't even say that anymore >> so that messaging was a little off but that was well I built programmatic so now I'm going to build this like I can I can see the through line I can see the thesis. but like there is a certainly now momentum in okay agents are being implemented here here and here and this is how programmatic is going to participate in this and the the convergence is coming.

32:38 >> yeah, >> that's great. >> But what about what about Scope 3 and Optimal and Swivel? Who who placed that bet at the same time? Now, aren't they now in direct competition with Google and Amazon? >> That's a good question. I certainly think that a lot of that Amazon the direction Amazon seems to be going is there. However, I wonder the extent to which specialization protects these companies, >> right? That's what they're >> and and that's and well, and that's the question is is specialization in ad operations and campaign optimization and all of that stuff enough to survive, right? Is that specialization enough or will the big companies eventually Oh, RSS still exists. Thank you, James.

33:28 It's on by default for WordPress sites. Yeah, man. I I think that I think when Google Sunset its reader is when it died to me. and >> James, what reader do you use? >> Yeah. but I Okay, what was I saying? >> specialization I >> What are their specialties? What are the specialties of Swivel, Optimal, and >> Well, they're actually different per company. Swivel comes out of CTV ad operations automation, right?

33:59 >> So, there is weird tribal knowledge and how do you maximize CTV yield inside of a publisher ad server that that's one Open AI definitely doesn't know that one. That's that's arcane arcane tribal knowledge that they know how to do. >> Is that are Google and Amazon blocked off from that knowledge? >> I they might be for a while. That that's real weird stuff. I find that intimidating. but then I would say that on the other side of things, >> I've I've talked to a few companies in inserting themselves at the holding company level and the large agency level that are like listening agents to manage four or five different advertising platforms properly and set up the reporting feedback loops and doing all of those things takes a significant amount of time and training, right? It takes a significant amount of time to build this out and they're hoping that the amount of time that it takes to set up a cohesive agent swarm that manages those things is their vote.

35:01 and that >> as of this September, they're going to be RFPd against Google and Amazon who are going to offer to do it for free in exchange for media commitments. That's that's tough that's a tough filled climb. >> Yeah. Here's the thing. I really like their businesses. I really like them. They're all lovely people. So, I don't know if they're going to get crooked, but whenever I would talk to my CTO about it, he was always he always said exactly what you said, which is the big guys will do this and they will be good at it. but maybe this is what maybe there's a weird co-opetition thing where they don't want to play nicely with each other, but they'll play nicely with third party middlemen.

35:46 possible, >> right? They they're like like I'm not going to help Google's business to do this, but this third party company like okay, I'll give them access. because you know Facebook, let let us come in here and >> settle this. >> Yeah, let us come in here and settle this. We don't have a we don't have a horse in the race. There's no like advertising platform. >> because it reminds me a lot of PMDs. Facebook used to really gate its APIs.

36:12 Remember that? That was the thing. Facebook did not give people access to their APIs. They did not do it. >> Yeah. Then they came out seven partners who were allowed to know how the APIs worked. Totally. >> All those companies >> made a lot of revenue, but their margin was capped by Facebook. So I don't know if they made a lot of money. >> Yeah. I I remember one of them, Liquid Social. and I remember them. And >> what happened to them? Media Ocean bought one.

36:41 >> Yeah. Who knows? wonder how those founders did. >> But it's interesting because because the history there can be informative of how this might play out where Facebook was super guarded. They're like, "We're not giving every Tom, Dick, and Harry access to this." And we're certainly not like, "You know who wasn't a PMD? Google, right? Like, we're not going to let Google set up campaigns in Facebook on behalf of their clients, even though I'm sure they would have loved to as part of their Google marketing offering."

37:07 because that just >> I remember Media Maths trying to do it. I wonder what happened with that. Oh, I'll have to ask Joe. >> That was a big thing when it came out. >> They chose that path. Media math was going to be a an overarching console into programmatic and and social. I don't think living that long. >> A lot of people have died on that hill, man. It >> is hard. >> So, that's a hard hill. It's a tough one, but it's back now, right? Because of all the agentic stuff. Everyone's back. I know a lot of companies that do this now.

37:36 >> try to reconcile them. Yeah. Us too, we would try to do it at the at the impression level. Yeah, you did your fancy way though. I mean just you know these are these are people having an agents going >> allocation budget allocation agent >> budget allocation that's what you got number up number down stuff like that like run a report what should I do today great go do it agent done >> let's move on now because there's some there's some publisher news from you about >> I'm glad James is watching today I want to hear James's James' thought on this >> well there we go I thought we were going to go straight fan. But I >> Oh, you're a fan. I'm sorry.

38:16 >> All right. Jag attitude. >> I don't know him, but I'm I like Justin Wall a lot who works there. So, this is a publisher. >> You said there were a few posts like this. A publishers complaining about an unnamed entity making publishers a deal they can't refuse an aggressive guarantee that their competitors think is fake. I read a couple of these and I was like, Gareth, what the is this? Who are they talking about? What is the entity that they're talking about doing?

38:45 Why are they mad? >> Yeah. >> So, I've known Jared a very long time. so, and just so everyone knows, like my first company, RTK, or my second company, RTK, was a pre-bid management company. So, most of these companies in inhabit the same space that RTK inhabited. and a lot of them sprung up while we were running RTK, like Freestar Immers while we were running RTK. Attitude popped up while we were running RTK. even though I pitched Jared RTK when he was at his last job actually. which is funny. He was my guy at PlayBuzz. And then, the other one's Playwire, which emerged from an ad network. And I've seen posts from all of those companies recently about an unnamed company that is doing a lot of revenue guarantees. and Jared came out. It's like super on the nose. And Jason Dubin from Playwire made a public LinkedIn comment on a case study from rapdev is what is what I saw.

39:34 >> Okay. So the CEO of Playwire just a public LinkedIn comment. This is a misrepresentation. >> So Raptive cont somewhat controversial company on adtech ad talk. They're CEO Paul Banister is very public. Always seems very reasonable. >> I think >> CO and used to be called Cafe Media. >> Used to be Cafe Media. And and their chief innovation officer, the chief chief R&D guy is Patrick McKent. So Patrick is who I chose to succeed me at PBI.js >> steward of pre-bidjs.

40:07 >> Yeah man you I I think Patrick is still running prejs and you can accuse Patrick in many things but man he really knows what he's doing and he's he's a smart guy and >> well he became controversial for yeah whatever he was >> he became controversial during the the TD ID transaction ID. Yeah he was he was the other side. Anyone can watch all those shows from last year where we covered that in great detail.

40:30 >> If you if you want to hit your re fire up your PTSD. >> so so what's Raptive doing now? They're going to publishers and saying >> so I I a I'm not certain that it's Raptive. That's just how it seems to me online. And B, the complaints seem to be that they're pitching revenue guarantees. That's that's from what I can tell from the three LinkedIn posts I've seen about this from CEO. >> How can Raptive possibly guarantee revenue?

40:53 So basically what they I'm sure they do is they go in they look at how much traffic a site has or they get those numbers from the publisher. They look at their CPMs for comps internally and they say publisher based on our internal comps here's how much money we think we could make you. And if you switch over to us for x period of time we will guarantee you the comp. >> And then when you hire aptive what do they do? They send your bid requests better.

41:22 Oh, that's a great question, Adam. >> I know what you do like you add data. >> You add data. >> The crux of the question because >> is how much of a difference can you make managing pre-bidjs and what are the levers that you have to pull to make a publisher's ad stack make that publisher more money? >> Right? Okay. So, there's the two evil ones might be in play which are bid duplication and ID bridging. the scourge of the open web. Twin scourges >> that everyone does. So rapid is not uniquely bad and nor everyone does those things. Everyone does.

41:59 >> Okay. Those are the two levers I know about. What are the are there angelic? >> So normally these companies have dynamic ad insertion. That's a thing. so that means they can add ads. >> Great user experience. >> Now here's the thing. the the LinkedIn post that Jason Duven called out was like, you know, they reduced ad experience. They had PMP revenue that juiced the numbers and they said, "Oh, look, this publisher made more money from fewer ads." And then the PMP ended and then they added more ads to the page.

42:26 >> Now, how >> I don't know is Raptive out on the buy side selling PMPs? >> They are. >> Okay. >> It's not a huge business. It's not huge, right? Compared compared to their overall pre-bid management business. It's not gigantic, but they're out they're out there. but there here's the thing. If there's anyone who knows where the bodies are buried and how to squeeze money out of programmatic, it's Patrick. He knows what he's doing. >> and like I would I I would not want to compete with Raptive. I'll put it that way. Larry, >> you do. But you do, right?

43:00 >> No, I don't compete with Raptive. Raptive is No, they're a publisher client of mine. I don't I don't do this business anymore. RTK competed with Raptive. Gamera, no. Absolutely not. These are these companies are all my partners. I love all these companies. >> >> but what is your play to the publisher that's different than theirs? Like you're gonna you're going to realize more. >> Oh, I don't manage their ad stacks. So, what I do is I do inventory quality. I do a few things for them. So, I do inventory quality core web vitals and other measurement and analytics on top of the publishers page. So, actually I have a partnership with a few of these companies and they implement our script as part of their package that they send to the publisher. And then I do like targeting signals. So if a publisher wants to set up a PMP of high quality inventory, we do all of the inventory measurement tax.

43:46 >> Why isn't that part of this? Why what's the difference of those two businesses from a publisher? They're both businesses that make me more money by sending better bid requests. Just one of them is managing JS and one of them is providing metrics and feedback. And >> yeah, they're they're complimentary, right? It's not one or the other. You can use you use both. You do use both. And that's why I I often implement via these companies. So like I won't necessarily do handto-h hand combat and close every publisher. I can close a deal with Jared at Attitude and and Attitude can leverage Gomera on behalf of all their publishers. perfectly lovely for me.

44:19 >> But >> it's very Yeah, I mean that's I didn't want to be in the supply chain again when I when I was done with RTK. Like I I just I don't like being in the supply chain. It's very >> Yeah. similar to me like like chalice were not in the financial chain which is the this real sticky political part of the of the buy side like with the money >> and imagine so these companies right take over publishers entire ad stacks that's what they do that's their business so the second something goes wrong on a publisher website even if it has nothing to do with these companies they are the first point of call for all of their clients and that so this is a services business and it is a hard services business because they're you're managing complex technology on generally unsophisticated publishers and the publisher goes, "Why is my revenue down? Is something broke? I I loaded the website. I didn't see an ad.

45:07 Something's broken. Go fix it." And you have a thousand people screaming at you like that. >> >> but yeah, I I certainly think that this is indicative of really aggressive consolidation in this space, which we have seen the inklings of before. Right. So James who's on here with with us right now because I can see his little image in the upper right. He worked at Sortable. Sortable was an early pre-bid management company that consolidated up I think into Freestar.

45:38 So I think Sortable became part of Freestar if memory serves maybe not. James help me. but there was always a thesis of all of these pre-bid management companies that manage the ASAX for publishers do something very similar to one another, right? because you're managing pre-bidjs and they're competing with knives right now and the reason they're so aggressive is actually because they're so similar. it is not because they are so different. It is because they are not so particularly differentiated from one another and when someone comes out making these revenue guarantees they sweep, right? It get they they it really aggressively conquers in the space and that's why they're all pissed. because one >> they're saying it's suspicious. It might. But you're saying it might not be suspicious. It looks suspicious.

46:24 >> Suspicious. >> Yeah. I mean, here's the thing. It might be an unsavory practice or not. I don't know. >> Natalie, right on Q. >> Yeah. What triggered her? >> Said unsavory. >> But it's it's one of those things where there are ways to do it in an unsavory way. You make a guarantee for a short period of time and then the second that short period of time is over, you you change the strategy. but I I don't know, man. I there's a lot of mudslinging out there and I think that what this points to for me though is consolidation in this space. I think it's >> and aggression same as the buy side, right? It's like schools school has started, summer is over and big players are making aggressive moves.

47:04 >> Yeah. I'm sure a lot of the CEOs of these companies would love to come on though. So maybe I'll maybe I'll maybe I'll I'll get one on here and we can see what they have to say. because these are these are big companies, right? They're not that small. and I I think it's really fascinating because these companies are the actual last mile to the mid and long tail of the internet. which a lot of it is very unclear to the buy side. When they show up inside of exchanges, they're often labeled as resellers. But there is no universe where attitude is a reseller. Well, maybe there was at one point, but not anymore. There's no universe where Freestar is a reseller.

47:38 If you are buying traffic from Freestar, you are buying the traffic from the publisher the only way it can be bought. they are the exclusive representative of that traffic or the semi-exclusive representative of that traffic. >> Yeah, they're a tech partner. >> Yeah, they're a tech partner. They're they're who the publisher uses to do this. So, I I certainly think the consolidation there means that the person managing the last mile to the publisher as it becomes fewer and fewer companies, these companies start to get more and more leverage. and especially in light of ADCP, there's a reason all these companies are listening to ADCP stuff is they all want direct sales. they all want to be more direct to buyers, etc., etc.

48:15 >> yeah. >> So, Google can't really play in this for obvious reasons for because for antitrust reasons, they can't make moves. But what about Amazon? Amazon has publisher services. They're want to be one of the players that's standing in that last mile. So, where are they? >> Yeah, it's a great question. APS was a a shot at this. there was a misapprehension made by very smart product people that server to server header bidding management products would compete with clienttos server header bidding management products. And so what I mean by that is the way that Amazon works and and Google did build a competitor by the way Adam. Google built OB which was originally called EBDA exchange bidding dynamic allocation.

49:00 >> Yeah. And the these were meant to be serverto-s server answers to client side pre-bidjs. >> Yeah. >> and when push came to shove, they didn't work as well. They weren't as configurable. They weren't open source. They still aren't open source. >> Yeah. They got in trouble. Yeah. >> The the number of features they have is slow compared to pre-bid. Pre-bid's a machine. Patrick Patrick pushes features out in JS that will make anyone blush. The pre-bid moves fast. and like pre-bid ad server is a thing by the way.

49:32 I I was digging into this the past few days. They it's they're ready. They they are prepared for that future to come. and I think that they just all of those people who built those serverto-s server offerings, there have been rumblings on the publisher side that Amazon's going to shut down the publisher offering for an extended period of time because the juice isn't worth the squeeze. It's too much work to maintain it features for it. >> You know, I don't know. People have been saying that for like years at this point. It hasn't happened. So I I don't think they're going to pull out, but people people would say it, right?

50:02 Because it they they can they they don't want to invest so much money and it doesn't make them money. And at a certain point, the product manager makes a strategic argument like this is a position we need to inhabit. And then five years later, the CFO goes, okay, we're not exclusively in this position. This is costing us oodles of money to host and to develop. Why are we doing this? and I think that that's that's what's played out with most of the big guys answers to open source pre-bid and pre-bid just won. Man, this is this battle's kind of over in my eyes. The pre-bid battle is over. Pre-bid has won.

50:36 It is the open source that everyone uses. I do not expect it to change. and >> so the battle's just who's managing it, who's who's going to manage it? >> Y 100%. And I mean, even Magnet Magnet has a preven management product that they're stepping away from for the reasons that I just described. and they're as a standalone business, it just has to be more expensive. It's it's it's just a hard business. It's a services business. but it's super valuable. People underestimate the value of these companies. These companies are really it's hard work and they they do a good job. so yeah, it's fun to see this. It makes me wonder what's going to happen. Maybe we go from five of these to two or three. that would that's perfectly possible.

51:17 All right, let's keep an eye on this next story. It's gonna be Yeah, I talked about this week last week about did Pulysus have to leave the Coke pitch when they won Pepsi and it's complicated like they already have Coke in some regions and they were pitching for Coke in other regions. regions being like continents in this case and they pulled out of the ones that they were pitching for. So not totally not a totally black and white situation of picking out but this piece by Mark Ritson who's a measurement and strategy advertising legend was quite funny. I think Marley called it He's somewhat he's a very vivid writer and I love this point he made on the next slide.

52:08 agencies get a lot of heat for conflicts. As a matter of fact, the reason there are holding companies and multiple agencies in those holding companies is because of conflicts, right? They have to launch an entire new agency to serve Honda when they already have Toyota. That sort of thing has been happening for years. that's how holding companies start to hold many agencies. But this last paragraph, he's he sarcastically says, "If you need evidence of how how problematic these conflicts are, just look at Silicon Valley, where Google and Meta have Coke and Pepsi money at the same time. Hold both companies, media plans, and audience data, strategies, and optimize them with a single algorithm without any apparent strain or conflict." What a paradox is pointing out here. And cheers to him for that.

53:00 That is makes no sense at all and is a fact. >> Yeah, >> they don't have conflicts. They don't have >> So true. >> It's one of those ones we just kind of let happen. >> The minimums agency has to do is put different people on them. You if you have two, you have to have different people. when back at Avenue A in 2005, we had Coke and I mean Chase and Capital One and the teams had to be in different floors and even use different copy and facts machines, you know, whatever they call it, a wall between us and that's the minimum. But an algorithm can be sh >> You know, Adam, the poolled algorithm theory would say you actually want the exact same people on both accounts.

53:48 >> That's what they do. Yeah, Google does that. They want the exact same people. >> The industry expert in each Google pod works with all the competitors and knows all their strategies and all their budgets and somehow that's not a problem. >> I don't think they'll let Amazon do that. I think that they Google and Meta can get away with it. I don't think Amazon's going to be able to because all of a sudden they're going to start seeing an Amazon version of all of their products popping up. They're going to be like, "Wait, wait a minute.

54:14 >> Their history of stealing products might >> wait a minute. What's Amazon Cola? Like, where did this come from? Why is it cheaper and why? >> It's the everything store. They've committed to selling everything. >> Yeah. No, it's it's fascinating. I think agents are really going to facilitate the move to like more siloing of this information. They've got to do it. Agents are so useful for it. they make life so much easier where we want to have modular infrastructure that is unique to us.

54:48 >> Get rid of this pulled pulled easy to log into UI thing where I have I don't have to train people anymore. I don't have to do this stuff. I just go type it in and it just goes and now I want my own private Idaho where I have my sequester data that you don't get access to that that works this way. >> live ramp we could we could do Coke and Pepsi with the same people, right?

55:10 Because we every model is per advertiser using only that advertiser's data. It could be completely fair. >> The battle if they use the same people and they have the same prowess, it'll come down to whose creative is better. >> Totally. When there's like when it's machines doing the estimations and there's numbers involved, as long as you don't bootstrap any learnings, like it's it's very easy to keep totally separate. you can even keep >> that's what the platforms work. The platforms pull the data and average it.

55:38 So >> yeah, >> and it's really interesting. I mean, I'm just thinking about it in terms of God, I I always thought clean rooms were so dumb. Like just so dumb. But like there is a there is a clean room component here where if you can keep your data in your data lake and you can synchronize it with Chalice without sharing it with anyone >> and then there's a way to ch for chalice to activate it with minimal exposure of the underlying points that informed the bidding strategy. I guess you could you could try and derive it on the other side based on what the bids are, but it's a lot harder.

56:14 >> yeah, we're not sure that that's a piece of demand. We still really don't know who's going to want us to do it that way and who's going to make us like stand up our entire infrastructure in their walls like Palunteer. >> Yeah, it could happen though. If I if I were big and sophisticated, that's how I'd operate. I'd be like, "No, no, this comes in. >> This comes in. >> They could do it. They could do it. they have enough money to cover that >> and we're building yeah to be to be pretty modular and you know there's a there's a number of environments we have to be able to build into but it's not infinite it's like going to be like three to five >> oh totally and especially if somebody pitches an infrastructure play like these companies are all starting to do where it's I will stand you up your own set of infrastructure that you control And >> right >> you and you can bring vendors into your instance of our infrastructure that comes pre-integrated with things and have >> it's going to have to be pretty standardized. A lot of it's going to have to be standardized.

57:17 >> IB Tech Lab's going to be busy. >> So busy, man. And it's so interesting because like the >> I think Amazon is really early here. I think some of the exchanges are really early, but this is like real 5 to 10 year out stuff. But it's but who then again who knows with AI >> it's good we're gonna see a lot of bets come down now we're gonna see we're gonna see a lot of more investment now >> I have to right >> in my six years if nothing but nos in investor conversations or that seems interesting or no feedback or criticism we're going to see some we're going to see some serious funding now that the smoke is starting to clear >> that's right >> that's right and some >> fun out Yeah, it's getting fun >> this fall. This fall. Yeah, this back to school season. It's always a bit of a starting gun coming back after summer, but this was like a starting cannon.

58:11 >> Oh, it's awesome. >> I'm insanely busy. >> Oh, it's the best. Or >> miss it when it's gone. >> You miss it when it's gone. >> Yeah, of course. >> You need sleep. >> I need sleep for a different reason. >> What? >> I need sleep for a different reason. I have co in the house. That's great. >> How's the first day of school? Do you have you had kids in public school? >> private school. it was great.

58:38 >> Yeah. >> My daughter's happy to be back. She's having fun. She she enjoys school. It's nice that she likes school. And >> yeah, >> I if if it was a battle to get her on the bus every morning, that would be a thing. >> yeah, our son hasn't started yet. He's he starts soon. Yeah. I know his his preschool has this this weird phase in thing, man, where he goes like an hour a day for like a week.

59:00 >> >> but we're getting there. How your How about you guys? Good back to school. >> Yeah. >> No illnesses yet. >> The girl loves school. The boy, it's a little more touch and go. Yeah, everyone's healthy. The boy had a playground injury yesterday, but it was >> no stitches. No, >> injuries. No stitches. It's fine. No glue. >> Just a It looks worse than it is. It's not It's not It's not not in Marley's league with her playground injury record setting.

59:27 >> Yeah. >> Everything Everything all right, Marley? Kids okay? >> Yeah. She's going to get the pins out of the back of her arm next Wednesday and she's like, you know, Christmas chain counting down the days to getting her pins out. So, >> there's no weirder feeling pins being taken out as a kid without pins in his hands. >> Does that hurt? No, it's just strange. It's just weird. because it's deep inside your bone. So you like you can't feel in there.

59:58 There's not nerves really, I don't think. But like it's it's weird. you feel like the the depth of the >> Yeah. >> Show her this, Marley. >> Yeah. >> She does not believe me that this is going to be a low-key experience. I actually feel like what you described. Yeah. Not bad. >> No, it is. It's the low-key experience. It was really interesting. It was super casual. He was just like, "Oh, come on in." And then we were out in like 5 minutes.

60:23 >> That's pretty harsh. >> Well, so's your coffee. >> All right, guys. I got to jump to a one. Thank you everyone for tuning in. Thank you, James, for coming. Thank you everyone else in chat for coming. We'll we'll see you all next week. >> Thanks, Marley, David Maldo, and Sasha. Thank you, Darth. Darn. That's the >> Yo, the crowd is pack. The lights are dim. Front rows buzzing. It's about to begin. Ad talk to taking the stage. Adam and Gareth on a knowledge rampage.

61:02 >> Tech ad talk. It's a live Q and