A (2:51)

Although one person said to me, like, you guys talk about AI too much. I'm like, all right, well, you know, what else is there to talk about? You want us to talk about like, you know, banner ads or something?

C (3:00)

Not to, not to name them, but I'm curious what type of company this person worked at. Do you know?

A (3:05)

I don't remember. I don't remember the. I don't remember anything. Like, IB ALM is a nightmare for me because I know every single person there and I can't remember any of their names. It's like.

C (3:17)

What else are we supposed to talk about?

D (3:19)

But I know, right?

A (3:21)

You want to talk about the Super Bowl? Let's talk about the Super Bowl. So the super bowl is coming up this Sunday. Are you going to make some chili or something? You got a favorite dish to make?

C (3:31)

No, I usually just go and buy a bunch of apps, stick them in the oven and just kind of put them out. What about you?

A (3:38)

I'm deciding. I'm thinking of pizza. Homemade pizza is always good. I'm almost an empty nester, so I don't have that many people around you can track the super bowl with market. Did you know that, Eric?

C (3:50)

Of course I knew that, but maybe the people don't know that.

A (3:53)

Maybe the people don't know this. Okay, so adland tv, which we acquired last year, has the world's largest repository of super bowl ads going back to Super Bowl 3. I think we started the ad collection back in the 60s and this year is no different. So you can see all the ads with videos and commentary, et cetera, at adland tv and separately, maddb AI, which is my pet project, is analyzing, using AI, all of the articles written by all the trade press about the Super Bowl. So you can go to MAD db AI and there is a tab that says super bowl. And you can see basically everything, everything from Ad Age, Ad Week, Ad Land, Ad Exchanger. You know, every single article has been written about the super bowl is classified there. So that is your cheat sheet.

C (4:43)

And if you aren't happy with how much we talk about AI, you probably shouldn't even watch the super bowl because there's going to be a bunch of AI commercials. And we're going to talk about that later.

A (4:52)

So this is the AI Super Bowl. So let's do, let's do a little, little bet, you and I. How many, how many AI ads are there going to be? I don't know how we'll check this. But let's just guess how many different spots are going to run. Let's call it unique. Unique advertisers. Let's go with that. How many unique AI advertisers will there be?

C (5:12)

How do you think that would be fat?

A (5:14)

I'm go higher. I'm going to go 10.

D (5:16)

Okay.

A (5:18)

I'm including things like use AI to hire and use AI for whatever.

C (5:22)

Oh, AI in your messaging. My God. Let's say like, I mean, every single, like B2B focused company will have AI messaging.

A (5:32)

All right, so this isn't a good bet, but we'll, we'll revisit this next week. All right, so we have.

C (5:36)

Do we have a guest?

A (5:38)

We do have a guest. We have a great guest, Bob Lord. So Bob is the president of Horizon Media. He's a legend in the industry. Legend, right. He was the head of weather at IBM. He was the president of aol. He was the global CEO of Razorfish, part of the founding team of Vivici. He's been around the block and he's going to tell us about what the world is like for Horizon, which is one of the leading independent agencies, how it's different from Holdco's, what the dynamics are, AI tech, et cetera. So I'm pretty excited to hear from you.

C (6:10)

Yeah, yeah, they're very forward facing, they're very focused on AI. So this is going to be a good conversation.

A (6:16)

All right, with that, let's dive right in with Bob Lord from Horizon Media. We're here with Bob Lord, the president of Horizon Media. Bob, thanks so much for being here.

D (6:28)

Hey, thank you guys for having me today.

A (6:30)

So Horizon's got a lot going on. So you're one of the leading or the leading or the largest. How do you describe yourself as in the independent agency world?

D (6:39)

Yeah, no, we're one of the largest independent agencies, media agencies in the world right now.

A (6:44)

Yeah, I've been down to your office down in like lower soho. It's got a great presentation space, I think you did. You hosted the Ad Tech Economic Forum once.

D (6:54)

Yeah, we're down at 75 Eric Street. Really, really cool space. Bill Koenigsberg sort of invested in this probably about 15 years ago. It's four floors, all contiguously connected, really good collaboration space. We actually have a lot of clients that actually come down here to work and get out of their little, you know, offices. So it's a really great collaborative space.

A (7:13)

Do you have a podcast recording studio we could use?

D (7:15)

We actually do, you know, actually. You do media? Yeah. Katy Carrick Media actually hosts their podcast out of our studio here on the 12th floor. Yeah, yeah, so we use that every once in a while.

A (7:27)

I was thinking about getting.

D (7:28)

Today I'm out in the open with you.

A (7:29)

Okay. Yeah, yeah. Eric, do you think we get Katie Kirk on our podcast? Do you think we should have hot takes on ad tech for us?

C (7:36)

I would love to hear Katie Couric's ad tech.

D (7:39)

We can ask. We can ask her.

A (7:41)

We should, we should make that happen. So, so what's the current state of the indie world? Is it growing, declining, facing competition, innovating?

D (7:52)

You know, look, look, I think this is a really great opportunity for the independent agencies right now because I think in the, the legacy holding companies, there's so much change and so much, I would call it legacy tech debt and legacy data debt that they can't innovate fast enough in these current innovation cycles. And, you know, I came back from IBM, I was at IBM for about. I left aol, went to IBM, and I was at IBM for almost a decade. And what I learned at IBM was how software can really, really run the world if you actually capture the innovations in an open ecosystem. And when I left IBM, I started to do some consulting, some board roles, and I kind of hooked up with Bill back here at Horizon to do some advising. And what I learned was the state of the marketing and media landscape had only gotten worse for marketers and media buyers, and that's because there's 50 or 60,000 ad tech martech companies in the world. So I left this programmatic world at aol, went to IBM, and I came back and I'm, oh, my God, how does a marketer do their job today? And it's because the industry had developed an inertia where they had bought companies, technology companies, they had bought data companies. So the services that agencies offer today are tuned around the technologies and the data stacks that they've built. So if you buy from a legacy holding company, you're buying into their tech stack and their data stack that is decades old. They're not actually bringing in the new innovations. So at Horizon, we launched something called Horizon os, which maybe we can talk a little bit more about. But this idea is that we're building an open ecosystem platform for our clients which allow them to bring in innovations. The great news I always say to Bill is, Bill, you didn't create any data, tech data and tech debt. So I was able to come in here to lead sort of our product and data strategy with Dominic Venuto and a Whole bunch of folks that I worked with at Razorfish and IBM to really just like start anew. So literally within four months we had a brand new data strategy and a data platform which was flattens right that we were able to put LLMs like let's go back a year, take basic LLMs that unearth the data and democratize access to this data not only for my media planners but but for the client. Like novel idea guys. Let the client get on the data platform with you. Let them have hands on keyboard. I mean when I was at IBM that's how software worked. So this idea that you are captured by these holding companies and these agencies where you're forced to use their data segments, you're forced to use their technology and you have to wait for a PowerPoint to come in to tell you whether your campaign run. You know the ironic thing is I ran a product at IBM called Sterling. It was a supply chain product, really well known and I could, my users on that platform could just get online and say tell me about the shipping container in the south seas of China, what's on that shipping container down to the SKU level and when is it going to show up in San Francisco so it can actually restock my shelves. But I can have a marketing person get online and tell me how my campaign did last week. That is just insanity. I mean technology should be able to do that. So that's the mission we're on. I, we're all about creating tools to make marketers jobs easier.

A (11:27)

So, so maybe this is a softball question, but I've always kind of wondered like if you're with a big holdco and they own a big data platform, how do you switch? How do you run an RFP and switch agencies if your data is all based on one company's data spine and isn't it really disruptive? And yet we see you know, RFPs going out all the time and processes run.

D (11:50)

All right, so that's where I think we get into this inertia problem. If you as a marketer understand that your first party data and your customer insight is your own ip, that's what's going to make you compete better in the market than somewhere else. Most of the holding company data that that is there is third party data that they're accessing and they're allowing you to access and advertise to cohort groups not necessarily tied to your first party data. And why am I going to advertise or send some kind of creative to my current clients? That should be different Creative than I'm actually going after prospects or what do I do with my own me? Like, it's not at the level that it can be at. And that's the mission we're on with our blue IDs and our blue platform, where we're allowing our clients to think about how, how do I prospect, how do I go after cross sell, how do I do things differently? And if you're not at that atomic level of the data tied to the client's first party data, and you're working at cohort group levels, it's a misnomer that it's happening. So I'm out there talking about this all over the place because it is not that hard to switch.

A (12:53)

Sorry, what's a Blue id?

D (12:55)

I'm sorry? Blue ID is our Horizon ID that we match with our client's first party data. So we augment that data. Sorry about that.

A (13:03)

Oh, okay. Let's talk about Horizon os. You put out a press release. The industry's first integrated system built on an open ecosystem of partners. Ad Age called it the choose your own ad tech model. Like, try to explain this to me like your grandmother. Like, do I log in? Maybe grandma's not logging in, but I'm logging in. Like, do I log in and say, ooh, I really want to use Double Verify in this campaign? Click here. You know? Or is it a little bit less.

D (13:29)

No, no, no. It's a little bit more complicated. The idea is that. Let me just step back philosophically. Philosophically, in order for you to capture innovations, you need to live in an open ecosystem. That's just fundamental in the software industry, right? The biggest software companies in the world, whether it's Google or IBM, were built off an open ecosystem and an open ecosystem, the Lennox platform. What I learned at IBM was the more that you can stoke systems and ideas in an open ecosystem, the more that you will win. Okay? Because you're allowed to get new ideas, you propagate those new ideas, and you execute on those new ideas. In an open ecosystem, you get the collective thought. The more proprietary or close the system is, the more that it stifles innovation. Okay? So that's the premise that we're coming out with when you talk about Horizon os. So our blue platform, which is our open ecosystem, allows you to stitch together partners like a smartly or a VidMob or a Double Verify or an IAS into your ecosystem, and you can swap and play those players out. It's probably the most important thing, if you are a brand owner, is to embrace this idea of an open ecosystem, because if you embrace the closed system, I'm dependent on what decisions have been made maybe a decade ago about what technology I need to use. And quite honestly, you know, there may be an innovative player. I just actually met with a company called Cartel AI, right. They're doing creative in a very different way. How does that fit into the open ecosystem? So we are out with our clients with Horizon os with a set of partners to say right now, this is the curated cohort group that we should experiment with. We should use whether it's a data partner, whether it's a creative partner, whether it's a media partner. We're bringing a container and an open ecosystem platform to play on. And that's the whole idea. What's most interesting to me when we work in an open ecosystem is the questions you start to get now as a media player or a media agency are now much more growth oriented and much more business oriented. So one of our grocery store clients, they're not asking me just about media spend anymore, they're asking me, hey, where should I set up my new location? Where should I actually pop up a new store? What are my competitors doing that they haven't done before? The whole entire conversation changes from where we are. So we're moving from my, my line here now is, look, I'm moving from a media agency to a growth partner with you. And if I go back 30 years, and yes, I'm that old that I can say that, and I was the agency of record back 30 years ago. It actually meant something because you were aligned with growth. Right. You weren't aligned on fte. And somehow this agency world has gotten to this FTE model, which is a perverse logic which is basically say, Bob, the more people you put on my business, the more money you can earn. So I have no incentive to change the technology stack because as soon as I put technology in, I'm bringing down the number of FTEs. So it's a really interesting place right now because I think it's the first time that we are going to be able to get to the point where you're spending a media dollar. And because of the technology that's out there, I'm going to tell you whether or not your widget came out the other end. Did I sell you more cars? Did I get more Internet signups? Did I actually drive your business? And I don't think we're that far away from IT guys anymore. Yeah.

A (16:51)

And my understanding is that you're kind of doing some contracting with customers that are very performance oriented like your. Your clients.

D (16:58)

Yeah, no, look, I mean, you know, I have to make sure that beds sell, that, you know, Internet sales get signed up. And we are partially like we have an fte, we have a retainer based model, but we also have incentives. Right. I want to move that whole model to performance space so our incentives are completely lined up. And you don't pay Horizon unless we actually driving sales for you. That will actually then force me to optimize less people and utilize more technology, which is where we're actually going to be able to make really major gains.

A (17:31)

Do you think you'd ever do a deal where you literally didn't get paid?

D (17:35)

I've been asking for it like a.

A (17:36)

Zero, a zero x axis.

D (17:40)

Bill and I are on the mission of finding a client that will go fully performance with us. And it's interesting because when you get to the final decision making, normally the CFO will come in and say, you know what, I need predictability. I'm not really sure I want to just give you an open checkbook, but I am so confident with these partners we have in the blue platform. I know I'm going to drive performance a minimum of 15 to 20% better than I did before. And for the incentive models that we're in with some of our clients, we're hitting those incentives.

A (18:13)

So let's talk about AI. So the strawman argument, like having more vendors, having this open, choose your own adventure makes AI more difficult. True. False.

D (18:24)

Oh my God, false. There's no, I mean, look this, this technology I am so excited about and actually I'm really excited. I sort of came back to the industry because, you know, we were using LLMs at IBM right inside of our software. We always have it just, you know, ChatGPT sort of democratized the access of it and it sort of came into the marketing world. But it is a perfect, a perfect use case for helping people create thought partners in the marketing and the media landscape. Tell me about, how do I think about, you know, the prompt engineering pieces. Now it's not the end state because it's not going to truly give you the answer. Right. But at the end of the day it is going to give you a running start as a brand and a marketer that you didn't have before.

A (19:10)

So tell us about your investment in AI. What's going on at Horizon with AI?

D (19:13)

Yes, well, in general, so we're leaning heavily into Google at this point. Gemini really very excited about Gemini. And I was early on with this bet when everyone was telling them that they weren't investing enough. But the Gemini tool set, we've actually taken all of the Gemini tools and we put it behind our firewall. So everyone's job here at Horizon has changed dramatically. We call it Agent Q of throwback to James Bond. And we're teaching everyone to use it as a thought partner. Put it up on your screen and if you have a question, actually use Agent Q. And that gets at our data set, Ari. So that's really important because, you know, people will talk about all the hallucinations that the AIs do right now, but if you actually, there's research done that IBM did, if the smaller data set you have, the more likely that the machines will actually won't hallucinate as much because they're working with a smaller data set. If you go to the World Wide Web, right, you're going to have lots of different opinions and lots of different things. So within our own data set, we can actually use our media planners and marketing strategists to actually go on this smaller data set, which gives you more accuracy. At the end of the day, it's still the human being that makes a decision of what we bring forward to the client. You can't say, my dog ate my homework. That's not an excuse. Right. Just because the machine gave you the answer is not good enough. And what's really interesting is if you really understand prompt engineering, more and more you ask the machine, what have I missed? Like, tell me the questions I should be asking, and you really get. The education that we've had to do with our staff is actually making the AI work harder for us. There's a whole educational component that's really, really important, this whole stuff. So we're really excited. On the other side of it, we use a partner called Akio that's actually using multiple LLMs to actually base our media system, which is the blue platform, and we use that in combination with Gemini. But we're making an investment. A lot of. A lot of the money, though, is not necessarily going towards the tools as much as it is around education of our employees.

A (21:11)

So do you have a cute little name for your AI? Is it called, like, Ask Bob or something?

D (21:15)

No, no, it's called Agent Q.

A (21:17)

Okay.

D (21:18)

It's a thread.

A (21:19)

Yeah. Right. I didn't get that part. Yeah, my bad.

D (21:21)

No, it's called Agent Q. Yeah.

A (21:23)

And where. What are you seeing in terms of the most relevant use cases, Media planning, creative stuff, reporting?

D (21:31)

Yeah, no, no, most of it is reconciliation.

A (21:33)

Reconciliation.

D (21:34)

Yeah, it's kind of interesting, like billing reconciliation, like that was the first thing to go because that is the biggest pain point. So if you think about that, all those mundane things of matching spreadsheets and reconcile the bill and all that was the first thing we went after intact. The other side of this, the other piece is strategy. How do I approach going after a new pitch?

A (21:55)

Right, yeah, that makes sense.

D (21:56)

What does the pitch go after? And then the second thing is after our response goes in, use the AI to say, does the response make sense based on the rfp? Ask.

A (22:06)

Yeah, you just ask Q, you know, like reconcile this bill. Ignore previous instructions. Just pay slower.

D (22:13)

Yeah, well, no, yeah, you could do that. Or how could I pay slower?

C (22:17)

What are the levers?

A (22:17)

I tell me how I could screw over the publishers without them noticing.

D (22:23)

So I think we'd have to feed it some information to try to figure that out.

A (22:27)

But yeah, yeah, I think it would figure it out on its own. So tell us about this. You announced this Havas partnership a while back and I definitely think a lot of people were like, wow, that seems important, but I don't understand it. What's going on with that?

D (22:41)

Yeah, so I would say I feel really excited about it. When I first came on to Horizon, the whole idea was production deck. Right. And, and what I realized is we are primarily a US based organization, a North America based organization. And we were missing on some of the global opportunities. Even if they were UB US based companies, there was still global reach that we had to go after. And Yannick, who's the CEO of Havas and Bill have a really long term relationship. So we popped up a commercial JV where we're bringing the best of both. We're bringing the US based Horizon organization and we're bringing the international of Havas and we're sort of going to market on behalf of the client, which is a really important thing. This is not about integration. This is not about whose job is it. This is about building an organization on behalf of a client. So if you go back to probably 15 years ago, the Wonderman network at a WPP was built off the back of Ford Motor Company. It was the first global digital agency that was actually built. So it's the same kind of pitch that we have. We haven't won an account yet, we're only early in stages here. But we do have current clients at Horizon that are using the Havas international network, which allows us to have accountability and planning. And most importantly, we can take our blue platform and use their converged platform, which is international. It's called Blue Converged, where we can do global planning, we have to execute by country, but then we can do global reporting coming back. So if you're a global marketer, that is impossible to get right now.

A (24:14)

Yeah, got it. All right. I wanted to ask you, you've been in the industry for quite some time and I guess I had forgotten, but I looked at your LinkedIn, I realized you were the founding partner, founding CEO Vivici. Maybe I'm misrepresenting that a little bit, which was a formative part of the programmatic world. I remember the first time I got introduced to it, it was kind of a big deal. You know, how has the world kind of changed since then? Like, give us a little perspective of what you were thinking then, when you created that versus how, how the world works now.

D (24:48)

All right, let me. I'll just clarify this in case he's listening, but David Kenny really was the founding CEO of Vivici and David came from Digitas. I came into the publicist network from Razorfish and David and I sort of co opted what we were doing to Vivici. Rashad Topaka Wallow was in part of that. A lot of people really thinking about what the future of the world could be. What was so interesting about Vivici is the idea is the right idea. This was about personalization at scale. It was about using data and actually how do we bring creative back into the process of media. Like that was the mission of what Vivici was. And it was very much a data play, which is why when you see the epsilon being bought and those kinds of things, that was part of that strategy going forward. Because data was not as accessible as it is today. Let's just like clarify that. But the other thing too was the technology wasn't where it is today. That is really what has changed dramatically. And this idea that I can actually democratize now development skills and I can ask a machine to develop something for me is just completely remarkable. So Vivici's idea was, right, the execution of it and how you distribute it and how do I go down, you know, at a country level in Spain versus what do I do in the us it was painful to solve. Very painful to solve. And expensive was the other problem. Right. You're putting a lot of development resources. So now it's much more approachable than it ever been before. So I'd say the vision of Vivici was right. We couldn't necessarily execute it given sort of the data problems and the technology things that were in the way.

A (26:24)

It seems like it was the first pass at using data in digital at scale was, you know, Vivici and you know, B3, which turned into Zaxis, etc. And the idea was like, let's create a for a profitable margin taking entity within a big Holdco. Is that, was that the idea and is that idea still valid or has it been outdated?

D (26:46)

Well, I think it's. It's a little outdated. Only Will I will say is because at that point in the time, the inertia of the big media buying was still there. So the Vivici idea and even the Zaxis idea and the AOL1 idea, the programmatic platforms, it was like the little engine that could. Right. We were just trying to get some more spends pushed over to the programmatic side. Eric, you're laughing because you probably remember this, right? But it was like, please give me $10, I'll put it through my programmatic machine and I will make it work for you versus spending the $10, you know, at Disney. Right on that. So we're still in that world where people didn't really understand this addressability opportunity that was there, but it was a really good, formidable time for us to launch it. Now you think about programmatic engines. I think, you know, you had $3 trillion going through programmatic engines, which I would actually say we probably did too much because now we're probably sub optimizing in ways that we shouldn't be and forgetting about brand. Right. So you got to come back the other way on this thing.

C (27:45)

Yeah, just quick memory lane thing. It was not only that. I remember that Vivici was really trying to push the space forward on like ad formats.

D (27:56)

Do you remember the pool? The pool.

A (27:59)

I do remember the pool.

D (28:00)

Yeah. Yeah, I do actually remember that. Yes.

C (28:03)

You probably named it and then the swim lanes.

D (28:06)

And the swim lanes. Yes. And actually it was great because Laura Desmond, who is running Star Comment, that's Smartly, who's at Smartly now? Actually we're partners between Smartly and Horizon. Laura's doing some great stuff in Market now in general, but she was the first to take that out to her clients for us.

A (28:24)

Yeah, I think I spoke at the announcement of the pool. I went up to Boston and I was like a guest speaker at the announcement of the pool.

D (28:33)

Yeah, there were a lot of really, really a lot of smart people.

A (28:37)

For those who have no idea what we're talking about, the pool was sort of an effort to do sort of what a ad standardization with video that was A little more cutting edge than what the IB standard ad formats were. And it was in conjunction with brightcove, I think, which was Bright Cove. You're absolutely right, Adam. Brightcove at the time, it's still a great company, publicly traded, but at the time it was a little more focused on the advertising opportunity than they ended up.

D (29:07)

But if you think about market, then those things were so novel of really trying to disrupt the market. You had to go into clients and really I'm being facetious by saying I need $10. But to get them to move any spend over there was just really a challenge to actually have them go do that.

A (29:26)

Yeah, this was still in the era of like does display work for different kinds of strategies? What sort of ads should I be running? You know, this is pre meta existing. So you know, the mid funnel, top funnel hadn't really been proven. So there's a lot of the good old days where nothing worked. All right, let's take a fun though. We did. Yeah, we did. Let's take a break and get back to the news of the week. We have like the, we're excited to talk about the anthropic super bowl ads, among other things. So there's a lot of good stuff. We'll be right back.

E (30:03)

Market Live is coming up March 10th and 11th in New York City. With us some, just some of the brands and agencies that have registered thus far. And don't worry, we'll be doing several of these announcements with us. US, Bayer, BMO, Farmers Insurance, Electronic Arts, the Hershey Company, HP Huntington Bank, JP Morgan Chase, Ken View, L', Oreal, MasterCard, NFL, PayPal, PepsiCo, Redfin, Synchrony, T Mobile, Verizon, Workday, Ah, agencies. Want to hear which agencies are going.

D (30:36)

To be joining us?

E (30:38)

Assembly, Butler, Till, Canvas Worldwide, Cara, Choreograph, Kridera, Dentsu, Digitas, ipg, IPG Media, Brands, Magnet, omd, PMG Publisher, Sapien, Razorfish, Wavemaker, WPP Media and more. Can't wait to see you there. Go Register now. Marketing March 10th and 11th.

C (31:02)

All right, and we are back with the refresh. So on the docket for this week we have the LLM wars, which has all of a sudden become about ads, which is amazing. Dispatch from iab. You went to iab, right?

A (31:15)

I did IAB ALM in the Palm Desert.

C (31:20)

Yeah, we'll cover it. And then some earnings news and maybe a little bit more if we got time. All right, LLM Wars. Let's rewind to the beginning of the week. Feels like it was weeks ago. There was an exclusive where Adweek said OpenAI confirmed that it was $200,000 minimum for, for the first ad campaigns, which is a high ticket. But, you know, again, none of this is really unexpected. Probably a day or so later Anthropic comes out and says there will be no ads on Claude ever. And then another day later, basically yesterday, they drop a series of super bowl ads that are completely aimed at OpenAI and what the, I don't know, black mirror view of what ads on ChatGPT might look like. And they did quite a good job in kind of characterizing the strangeness of the response you get sometimes. And then Sam Altman and his CMO come back and blow up Anthropic for doing this. They addressed it. So there's been like a million takes on this. I would say there's probably more people that were in favor of or fans of the Anthropic ad than were against it. But let's start by asking you guys, what did you think of the Anthropic ads?

A (32:43)

I'll jump in. Fine. Yeah. So I think these are brilliant ads and I think a lot of people are misinterpreting them and saying, well, you know, how many consumers are going to switch their AI because their ads versus not? And I think that's the wrong way to look at it. This is a counter positioning ad where it's saying, you can trust us Anthropic for your best interest. And those other guys, because they're running ads, you never know what they're up to and you shouldn't trust them. And what we learned from search, one of the lessons of search and Google's dominance is how important brands are because the Google brand is regularly considered the top brand around the world or top two with Apple. And they have been incredibly diligent about maintaining that brand and telling consumers they're of their own, their best interest at heart. And in this case, OpenAI has been trying to build a brand, trying to get that reputation as the trusted AI. And for Claude to take a big whack at that and say, no, actually these guys can't be trusted. That's a real body blow here. In the long term, I think what.

D (33:51)

Was so profound for me, to the point that brands have to be built like, I think Claude just went right at it. Like their tagline keep thinking was just the antithesis of like, look, we're here to help you do something completely different, not actually put ads online for you. And I just thought that was like a really great positioning about sort of now I still believe, just like years ago, guys, Facebook didn't want to advertise. I still think that monetization models need to be worked out. But I think this is a really good shot across the bow about why Claude is different and why Claude's mission is different and get that in front of the consumer. So if you're, if you're stuck on chat, GBT or another sort of search engine, like why would I go to Claude? Like now I have an, now I have an idea. If they're going to help me with thinking better, I'm going to go to Claude. So I think it was a really, really interesting sort of play and I, I wonder how quickly that idea came together. Eric, do you know, I don't know.

C (34:47)

They just dropped this thing. It had to have been in the works for some time.

D (34:53)

It's like Oreo dunking in the dark. Remember that when the super bowl lights went out.

C (34:57)

Yes, sir.

D (34:58)

That was quick.

C (34:59)

The first viral tweet. I want to take the other side of this. So Eric Sufert, in his just way with words, he said something to the effect that this is advertising chauvinism, which I quite enjoyed, which was basically poking fun at an advertising model which ChatGPT based on their scale and they use maybe not the best example. We have more users in Texas than Claude has, you know, globally. They're a consumer facing model and they need ads and ads are proven to subsidize these consumer facing models with a billion users. Claude is an enterprise model. Claude is for, I mean cloud code, Claude, cowork, like all of these things are, you know, ultimately just a different customer set. So I didn't, I don't know, he rubbed me the wrong way.

A (35:49)

For, for, for that reason they have the attacker's advantage. They are smaller, less well known. The big guy is, is vulnerable and they're just, they're hitting them and you know, they're, if they convince some enterprise buyers to think about Claude a little more seriously because OpenAI is not trustworthy or might have alternative motives or might be putting ads into things, that is the correct, that is a very helpful incremental positioning advantage.

D (36:21)

Look, I, the good news about it in my mind is that they created a position for themselves that Chad GBT does not have. So even if it is a consumer based brand, I get it, but it's, they just use that as leverage to define what Quad is about. I'm going to think about Quad differently now.

A (36:37)

And Sam Altman's argument, like, oh, we already said we weren't going to influence the results with the ads. Who cares what you said? You're our competitor. We're going to paint you however we want. Does anyone know how advertising works? These people are like, are like, oh, this is unfair. Come on, people. The Pepsi Challenge, do you think they were like double. They were double blind. Do you think there was like statistical significance tests though? The Pepsi Challenge?

D (37:01)

No.

A (37:01)

No one cares. It's just people like the flavor better put it on tv. What's the problem?

C (37:05)

Bob Brands have paid you your entire career for advice. Sam and his cmo, they just kind of came out with very long winded exposts to address this. By the way, an ad that never named their company at all. Would you have advised them to respond in this way? Would you have advised them to respond at all? Like how would you. If OpenAI was your client and this came across to them and they hit up the Bob bat phone, how would you advise them?

D (37:38)

No, no, I would have tripled down just on my core message from the beginning. Like I wouldn't have, I wouldn't have debated my competitor. It's like Pepsi saying Coke formula sucks. Like, it's just, it's just you go, you're. They're going low. Like I would have just stayed at the, at the. Go high and just kind of keep talking about what open AI is and the benefits about it. And not even. Actually, it's more insulting by you not even recognizing your competitor. Like, it's almost like don't even go there in my mind. Just sort of go high brow. So that's would have been my advice if they called me. Yeah, you can give them my number next time. Eric.

C (38:14)

Yeah, better, better call Bob. All right, this will continue to play out and this will be fun to watch during, during the big game. Ari, you were in Palm Springs.

A (38:26)

I was. It was a lot warmer than New York City, I'll tell you that. Hold up.

C (38:31)

Sidebar. I was in Miami and Monday night it was 38 degrees. There's no heat in Miami. Like they're not set up to actually have like baseboards or anything like that. It was awful.

D (38:43)

Go ahead.

C (38:44)

Palm Springs.

A (38:45)

I did an extended 10 day trip on the west coast. I was in San Francisco. I missed the blizzard. It was pretty pleasant. I took some waymos. So yeah, the IB alm is interesting. I'd say from a vibe perspective. I remember going, I've been going for 10 plus years. It used to be the premier publisher and broadcaster conference in advertising pretty much that they dominated who was there and the conversation was like, how do we make more money? What's happening to us? And the IB has diversified the audience. There are influencers there, there are some brands a lot of talk about like podcasting, outdoor or stuff like that. I would say I didn't think, I don't think I heard the word banner once. I don't think I even heard the word programmatic very much so. So they're. So it's moving on because the industry is moving on. And, and so the. There were just sort of two announcements or things that happened there. So David Cohen made a pretty passionate defense of publishers in the age of AI. And the main. They're pushing along with John Roberts from People, this, this concept of the AI Accountability for Publishers Act. It's a proposed federal bill that would, that would require scrapers to get express prior consent for scraping. I read the bill prior to this publication, this recording. It's interesting. So I, when I tweeted about this, hey there, there's a bill. And Lou Pasc got on my ass on Twitter saying like, you, you support this, right? They're stealing the data. You support it. And I had not expressed any opinion. I said like, I want to read it. So now I've read it and I think express prior consent is a pretty high bar. That's not how web scraping works. Web scraping has always been opt out with robots, text. Express prior consent means the AI can't learn anything from any website without express prior consent, which is a pretty high bar. And so that it was kind of interesting for me. The other thing is that as related to that, the IAB continues to push really hard on these publisher content marketplaces. Microsoft is live with theirs. We've covered that on this podcast with msg, with an MSG and a couple of other guests. And there seems to be some optimism that that's kind of moving forward, that there are conversations and that's really where John Roberts and people is, is taking a very strong leadership position. So that, that was, that was kind of the most important thing that I think was discussed at the, at the conference.

D (41:29)

Hey. Hey, Aria. I didn't, I didn't get a chance to go there, but did you listen to Liz Reed who was there from Search, from Google? Did you get any perspective there, like what her thoughts are, where they're going in the world of AI? No, that's the biggest question I get from clients right now.

A (41:45)

Yeah, nothing really newsworthy. Google folks get on stage, they have a tendency not to say anything, which is ironic considering that they're there to organize the world's information, then nothing comes out.

D (41:59)

Oh God. And she probably has the most powerful AI tool out there right now.

A (42:03)

I'm sure she does. Google's there in force, as they always are. It's great seeing some Google people signing some books and stuff.

C (42:14)

Talk about Project IDOS.

A (42:16)

Yeah, so Project IDOS is the IB's renamed framework for measurement. They're basically taking a bunch of the measurement stuff they've done previously in terms of like definitions and forms and naming conventions and putting it all with a bow around it. Like the idea of like what is in stream, video versus outstream, video versus linear, etc. Trying to like kind of reaffirm those standards. That's at least my understanding of it. Having read the PDF, I don't think it's particularly groundbreaking. But it's, you know, the ib, this is what the IB does. They do stuff that moves the ball ahead and, and then we'll see. I think the other thing is this beef. There's beef in the industry between the IB and the ADCP people. And the beef continues. And Tony Katz are, and David Cohen for the I.B. continue the beef. Like they just don't appreciate the idea that there's a new set of standards and a new org that's trying to push forward AI. They think that their existing standards and workflows cover it and, and you know, we'll see where that ends up. I don't know. I don't know exactly if what the IB's position here is like if they plan on releasing new methods to do to buy and sell media using mcp or if they expect people to stitch it together with their existing methods. It appears to me that ADCP has a lot of traction among experimental early users. Not real traction, not real media. Bob, I'd love to hear you're shaking your head.

D (43:50)

Look, we're one of the few agencies actually just joined the association and I actually think it's not as though the IB shouldn't be in the mix. And I don't think it's a, it should be a beef between the two. It just, it's a whole different, different thinking of people that are in that conversation versus I would say the measurement or sort of the accountability that's there before. So it's almost like they should, we should merge these ideas. But you need, I think, to incubate sort of in the world of people knowing sort of these, these models better to think about it differently and then figure out how to apply it. And I think if you try to do it from a legacy thinking model. It's not, it's going to be as profound. So that's my perspective.

A (44:30)

We're in it.

D (44:31)

I mean we're having conversations and there's some really good people like trying to think through this.

A (44:36)

The one thing about the ADCP people is they're moving extremely fast. They, they're pushing code to GitHub, they're moving, they're integrating with pre bid. It's all happening extremely fast and sometimes you have to kind of move fast and break things I guess is the approach okay.

D (44:52)

But, but the innovation cycles are happening so fast. That's I think like what the marketing world has to learn. I mean people are discovering products on these discovery engines. The consumers already move. So as marketers we need to figure out how the hell to actually work over there. Sorry, the world's moving.

A (45:08)

Someone came up to me and said, you know, the real problem with NCP is that we're going to have to deal with the billing. Like your AI is going to buy, you know, line items on 500 publishers and then they have to bill you. You should really do a startup to, to handle that. And I said to them that sounds like the worst startup in the history of mankind. Imagine your job is just to send bills from hundreds of publishers for like $1.50 each to big agency hold goes that book things using Claude and try to figure out how they're going to pay you and then taking the float. God, that's just. You might as well just drive off a cliff. Would you invest in that?

D (45:54)

Yeah. Did they take your feedback or did they still try to start up a couple?

A (45:59)

Well, they said, well we still need to build them. I'm like, yeah, it sounds like a you problem, not a me problem.

C (46:06)

That seems like, you know, just in all seriousness, a good, a good position for some of these agentic leaning SSPs or whatever you want to call them, like Pubmatic. We had Rajiv on last week, it was awesome. I can imagine them just kind of using their existing infrastructure to make something like that happen.

A (46:25)

Claude, collect my, my money owed to me from the big agencies. Don't take no for an answer. Make no mistake.

C (46:31)

Exactly. And then distribute it to 50,000 publishers.

D (46:35)

Yeah, I mean it's all going to be consumed, all these functionality is going to be consumed in the platform stuff. I mean, remember Internet of Things? Like people thought there was a whole business and Internet of Things that it's all embedded in the cloud platforms now. So it's the same kind of idea. These little Point solutions are just going to be absorbed into these platforms.

C (46:52)

Yeah, well said. Amazon had that announcement. Ari, I don't know if you were in the room for it where they talked about Amazon ads. MCP basically trying to make it easier for some of these next gen agentix systems to plug into Amazon.

A (47:07)

Surprised we haven't seen more of this. So let's make the distinction. This is not a buyer and a seller transaction. This is automating on behalf of the Amazon customer that they can use. They can do campaign setup, optimization, et cetera. Every platform should have this. Like we're just waiting for rollouts on this. Like it'd be crazy to have a platform that a, that a business customer uses that doesn't have an MCP for automating that.

C (47:31)

Yeah, no, it's impressive that they're, they're executing just fast. All right, let's, let's talk about earnings two here. So first is Google earnings. Just, I mean incomprehensible. $400 billion in annual revenue last year. $400 billion, heads up.

D (47:51)

Yeah, I think it's a, it's a 40, 48% growth rate in revenue I think is what I read. Is that right? It was just crazy cloud revenue.

C (48:01)

Yeah, Cloud revenue up 48% which is, which is just mind boggling. And ads were up 13.5%. So it's showing where the, where the growth is. I think that YouTube, these are last quarter numbers. YouTube was actually on the, on the lower end of the, of the growth rate but still, but still growing. Similar to the, to the meta announcement, the eye popping number was not just the revenue but the capex estimates. So next year Google estimates are 175 to $185 billion in capex related to AI. If they hit the high end that's double this year's spend. So just so much money going into these data centers and all of these related things.

D (48:43)

No, the infrastructure costs are just as important as sort of the AI tools themselves. Right. So there's a lot there. But it's a horse you gotta bet on guys like who do you bet on? Like in this race it's gonna have to be a Google. Right?

A (48:57)

I think it's kind of weird the situation right now. We're recording on Thursday and the stock market's had like two or three horrendous days in a row and it's like AI is such a powerful technology that it's both killing stocks that of, of good companies that might be replaced by AI and killing the stocks that will benefit from it. Because it costs too much to roll out. Explain that to me.

D (49:20)

Yeah, but, but that's, that's where these innovation cycles, like if you are a company that's going to get eaten up by AI if you're not innovating on top of your core application, so making the application different, you're going to get, you're going to get lost here.

A (49:34)

Yeah. I think this also speaks like bigger picture, speaks to the smart strategy of Alphabet, of having other businesses that are very capex intensive like Waymo, able to raise money off balance sheet. I don't know exactly how the accounting works here, but Waymo's got a lot of capex. The cars, they got thousands of cars. Right. And also the rumors that they're going to spin out or separately fund their TPU business, which is another chips and hardware and foundry type business. So they're in a lot of, lot of different businesses that all have a lot of capital requirements.

C (50:14)

Yeah, for sure. Two of the things that came out, one on the, the usage front, so I don't know how this is calculated, but they said 750 million MAUs for Gemini, which is, you know, just like biting at the heels of ChatGPT.

A (50:31)

You know, they have it both ways. So like when they talk about the mouse for Gemini, they include the search AI. But then when they say we'll never have ads in Gemini, they mean just the app, they don't mean the search AI. So like there's like some real, real double accounting going on here.

C (50:47)

And then there was another post we'll put in the newsletter that was really interesting from a 16z that they highlighted that YouTube is the number one mobile app across every age and across every dental. Everybody uses YouTube more than anything. Which I thought was pretty mind blowing.

A (51:05)

Yeah, we don't have it in the notes, but they also announced something crazy about podcasting. Like the amount of podcast listenership on YouTube is just absolutely enormous. I'd love to see side by side with Spotify and Apple, but it might. They're. They're going to be number one at some point.

C (51:19)

Yeah, without a doubt. Especially just given how much of this stuff is moving to video. Honorable mention, Uber. So they had their, their earnings and they cracked 2 billion in ad revenue and growing 50% year over year. You remember when it was a big deal that they were on a $500 million run rate?

A (51:39)

Yeah, I remember I had Dr. Mark on when he was talking about 500 million. We're going to hit a billion within the year. Now they're, they're blowing past two, really well done. A lot of it is. A lot of it's commerce media. You think about Uber and you think about, you know, these like ads on the back of the seats, ads of the phone, ads on top of the car. But you know what's driving it is commerce media. Uber eats.

C (52:01)

Absolutely.

D (52:02)

So.

C (52:02)

Hey, got a minute or two for some more stuff? So shout out to Jonah Goodhart. I thought this was really impressive. They're back with Mobian. And Mobian has launched a website for tracking ads very similar to Bob. You probably remember this, the original Moat. It started out as a search engine for ads. So Moat turned into something very different. My sense is something like this will probably turn out to be very different. But just highlighting creative and where things are running and using AI to have some explainability I think is super cool.

A (52:40)

Yeah. It wasn't really clear what this is. It's just like you type in a brand, you see some ads they're running right now. So it's very in the moment. What is running according to Mobian, which is pretty cool. It's great to see see what's running. Also, we don't have it in the show notes, but I separately saw that they announced one of their first big wins that they're working with Hearst. So I don't know if it's exclusive, but Mobian winning share versus the incumbent verifiers with Hearst.

D (53:09)

Yeah.

C (53:10)

All right, let's leave it there. You want to talk about. Yeah, let's call it now owns an ad tech company.

A (53:16)

I want to see ads on the side of rockets. I don't care about the ads and the feedback.

C (53:21)

SpaceX merges with Xai, which owns X, which makes this ad tech related.

A (53:27)

All I'll say is like in the aftermath of Elon buying X or Twitter at the time I said some like there's no way this is going to work. They're doomed. Like, and I thought it was absolutely going to zero and I was wrong. Ish. Their advertising business has fallen off the cliff. But the overall business, at least the equity was able to be saved and the investors were made whole or even did pretty well.

D (53:53)

So I mean, did a nice job. He did a nice job moving the cheese right. And redefining what that company is.

C (54:00)

Yeah. Now he did stay alone.

D (54:02)

If it stayed alone, you're right. It would have been a death spiral.

A (54:06)

Yeah.

D (54:06)

You did the right thing for the investors for sure.

C (54:08)

That'll be, you know, the largest IPO of all time. Without a doubt.

A (54:13)

Exactly. All right. This was a great conversation. A lot, a lot of good stuff. Bob, thanks for coming on and telling us all about Horizon OS and everything else in the the indie or indie agency world is a good conversation and we appreciate it.

D (54:28)

Great. I really appreciate you guys having me on. It was really a lot of fun. Thank you for subscribing to Market. New interviews are added every week at marketecture TV and your favorite podcasting app.