← Marketecture: Get Smart. Fast.
Marketecture: Get Smart. Fast.
Episode 162: Eric Seufert on the SaaS-pocolypse, Meta’s Manus, and AppLovin’s Social Network
01:05:07
Loading summary
Transcript185 lines
- [00:01]
A
Are you still spending Mondays stitching together spreadsheets? Take back your Mondays with Bert Intelligence. Bert unifies data across, adopts programmatic and sales into one trusted platform. No more manual reconciliation, no more reporting marathons. Just clean data, reliable forecasting, and faster revenue decisions. Whether you're scaling a publisher business or building retail media, Bert gives your team the clarity to grow revenue with confidence. Visit Bert AI AdTechGod for a free trial and and mention AdTech God for up to 20% off your annual contract. That's B U R T AI ADT God. This podcast is brought to you by MediaOcean. Everyone knows the LumaScape now behold the MediaOceanScape. A singular graphic showing all the companies they've come together to form the MediaOcean of today. 28 acquisitions in 59 years. Yes, you heard that right. Founded in 1967, they are the OG of adtech. Next year, MediaOcean will be celebrating its 60th anniversary, and they're showing no signs of slowing down with the acquisition of Innovid and launch of a new AI orchestrator. To learn more, go to mediaocean.com escape that's mediaocean.com scap. Welcome to the Mark Deckscher Podcast. This is Ari Paparo. I'm here with Eric Franchi, and we're pretty excited to have friend of the POD Eric Suefert on. So Eric runs Mobile Dev Memo. He's very outspoken and intellectual about everything involving the walled gardens, the hedge gardens app, install marketplace, and sort of algorithmic ads. And there's a lot of topics that have been coming up lately. It started out with, like, this whole thing about AppLovin creating its own social network, which I found a little amusing. I was like, hey, let's get Eric's point of view. But then there's so much other stuff we needed to talk to him about, like this Manus tool from Facebook Meta that you probably aren't using today, but might be really important, and the sasspocalypse and other things like that. So, like, Eric, what's. How do you keep on top of all this as an investor? Like, I, I don't have anything on the line, but you're like, putting money out there and all. Everything's changing all the time.
- [02:32]
B
There's no substitute. You need to read a lot, follow smart people. And then, I mean, the way the job works is a lot of those smart people are founders. All of those smart people are raising money for their startups. And I happen to be part of, you know, just in the Middle of the. Of the workflow there. So I'm the beneficiary of being at the bleeding edge of, you know, some of this stuff that ultimately gets. Gets written up afterwards. But I'm so excited about Eric Sofer coming this pod that free. I still think about that presentation he did at Market Live too, where, like, everybody feels so dumb. He is just, like, he just operates on a different plane. And I learned so much from listening to him.
- [03:12]
A
Yeah. For those of you who weren't there, we invited him to do the keynote after lunch, and I gave him free reign. I was like, talk about whatever you want to talk about. And he brought out the calculus like there were integrals on the screen. And like all of our brethren and their ad tech tuxedos were, you know, taking notes, taking pictures, trying to keep up. So that was. That was very amusing. It was very interesting as well.
- [03:38]
B
The satisficer's regret. Was that.
- [03:40]
A
Was that the term, that satisficer's regret is the idea that if you. If, like, a company like Meta is too good at optimizing stuff, that you as an advertiser think that you didn't get as good a deal as possible because they took too much margin. I think that's the explanation.
- [03:59]
B
Yeah. Coffee.
- [04:00]
A
This is going to be a good one. All right, so two weeks from now, we have Mark Dexter Live, which is pretty exciting. Last week I was complaining that I had to do my deck. I did the first version. It kind of sucks. That's how it works. But maybe by next week it'll be pretty good. So we hadn't announced who, you know, we closed last to Market Live with what we call the Live Architecture interview, where I interviewed. And the first one is Brian o', Kelly. And he, for the first time ever, talked about the pivot of Scope three to AI. So that was pretty newsworthy and really interesting. And then at the Market Live in the fall, we had Terry Kawaja, who brought down the House. He just brought down the House, talking about AI and the future in America. And he's going to be back on stage as well, but not with me doing a really interesting presentation. So we had to have a good name to end the show. Day two. End the show. Eric, you want to take any guesses who we got?
- [04:52]
B
Give me a hint.
- [04:54]
A
It's someone we've wanted to have on the show before and have not had.
- [04:58]
B
Is it the big one?
- [05:00]
A
It is the big one. It is the person in the news, Jeff Green, the CEO of the trade desk, will be joining Me finally on stage for an in depth interview about all of the interesting topics. We're not going to talk about the stock price. I just am not interested in the stock price. We're talking about openpath, we're talking about Ventura. We're going to talk about what it's like running a public company that everyone cares about, whether transparency is the future. We're going to have interesting conversation. I'm very thankful that he agreed to join. It's the final session on day two, in addition to this amazing, amazing agenda. I looked at the agenda yesterday, I was like, oh my God, how do we get all these people? This is crazy.
- [05:40]
C
It's ridiculous. I know I say this every week,
- [05:43]
B
but you have the greatest cyclist of all time, Lance Armstrong.
- [05:48]
A
Yeah.
- [05:48]
B
You have Jeff Korean. It's amazing. Listeners. We've been trying to get Jeff Green on since day one of this pod I think is like three years plus or minus. This is, I mean, holy cow. This is the end of, end of the show. So Wednesday closing.
- [06:07]
A
Yeah. So it's on Tuesday and Wednesday, March 10th and 11th. We start with my keynote that day one ends with your startup showcase, more or less. We also have Joanna o', Connell, who's amazing, giving her a presentation at the end of day one. Day two starts with Jeremiah Oweng, who is an AI expert, talking about a lot of super interesting stuff and then the day is just packed with interesting content. And then we end with more or less end with Lance Armstrong, no hold barred conversation. He actually gave instruction to our team. He does not want to see the questions in advance. So that's always fun. And then I finish off with Jeff Green and it should be amazing. So I guess we should tell people you should buy tickets. Right? You should buy. If you haven't bought tickets. You should buy tickets if you haven't
- [06:53]
C
bought your tickets yet.
- [06:54]
B
Like, let this be your catalyst to buy the tickets. I mean this like if you're a market tech listener, you've been just like hearing us, you know, asking Jeff to come on like every few months.
- [07:04]
A
This is it. Yeah. Also I heard that the breakfast is being catered by Russian daughters, which is very exciting for a New Yorker. Russian daughters, it's a, you know, 100 year old bagel shop on the Lower east side. It's known for its bagels and locks. Yeah. So we got some interesting New York food to round it out, which is always important.
- [07:22]
B
Lower east side.
- [07:23]
A
Lower east side, Yep. It's very famous.
- [07:25]
B
I have no idea. I've never heard of it.
- [07:26]
A
Yeah, I know. I know. We'll have Staten Island Pizza, the next market after Life. All right, let's get into it with Eric Suefert and we'll talk about all the exciting stuff going on in the world of apps, walled gardens, Facebook, etc. Friend of the Pod, Eric Soufert. Thanks for coming.
- [07:44]
C
Hey, thanks for having me. It's good to be back.
- [07:46]
A
There's just so much interesting stuff going on in the, in the, in app world, the walled garden world, the AI world. I really wanted to have you on to just like, give us the latest. And you, you also wrote this super interesting article about Apple's gatekeeping on AI data. That one seems pretty relevant, so I don't even know where to start. Let's start with SaaS Apocalypse. So we've all seen over the last couple of weeks this massive sort of fear gripping Wall street and the idea that AI can just replace software entirely and that stocks have taken 50% cuts to their valuation. Elon Musk, I just found this this morning before we were recording because I knew someone was saying this. There's just generally this feeling like that apps might not exist, I mean, obviously of games, but you know, if you are, if you have, you know, chatgpt or an agent openclaw, and you get me an Uber, you may not need the Uber app, in which case the app ecosystem may disappear. And Elon Musk said this on Joe Rogan. He said, we're not going to have a phone in the traditional sense. There won't be an operating system or apps. It'll just be a device that's an edge node for AI inference. So what's your take on this, Eric?
- [09:01]
C
Well, first of all, how are you? It's good to see you.
- [09:06]
A
We're diving right in, man. People are busy. They need the content.
- [09:10]
B
Have you had your coffee, Eric?
- [09:12]
A
I have.
- [09:13]
C
I've got a coffee. Yeah. So listen, I wrote about this two WWCs ago because Apple introduced Apple Intelligence or they announced it, right? And people were worried that I was going to kill apps because they introduced this idea of app intents, right? So you could sort of create a hook in your app and you could just ask Siri, hey, Siri, order me an Uber. Right? And there's just a lot of incentive around wanting to own a surface area that users interface with. Like, if you allow yourself to get disintermediated from the user, you lose a lot. Right? And that connects to my argument about independent agentic commerce being a mirage. But I also, you know, but, but that, that applies kind of like to any consumer facing surface area that you could imagine, including apps, right? And so one of the points I made in that piece is like, well, if everyone's just sort of running an API or an MCP it, then then they're commodities, right? Like what's, what's the real value that they bring to bear and how do they get paid? And you've got to find the sort of commercial incentive here. And so not only so if you just eliminate the surface area completely, you know, you get rid of things like ads, right? Advertising just doesn't exist. Or it takes on this totally new form of agents talking to each other, which again I just don't think is supported by any sort of commercial impetus or commercial incentive. I just don't think it's going to happen. Agents advertising or agents being advertised to is just not, I don't think it's ever going to be a format that exists. But, but you also, you lose, you lose any sort of ability to communicate with the user, right? If you're always, you know, mediated by, by just the actual surface area, then you lose any relationship at all with the, with the consumer. That's, that was the core of my argument with the agentic commerce as a mirage piece. But it applies to anything here, right? And then you have to ask yourself, well then what prevents the edge inference node, which by the way, it doesn't make any sense. You're going to be doing inference on the edge, most likely. Maybe you're talking about 50 now, but
- [11:13]
A
in our lifetimes, Elon loves to throw out like technical big note, like throw out some words he heard in a meeting and stitch it all together.
- [11:22]
C
Right? But this actually connects to another topic that we want to talk about, which is the Apple gatekeeping, the AI usage. Because you can't do inference on the edge, you have to do inference at a server. And that's where they're inserting the gatekeeping. But we'll come to that. So then what prevents the surface area from being disintermediated, right? What prevents somebody else from coming over the closer to the user surface area that disintermediates the old surface area, right? And so my sense is like you never truly get rid of apps. You still have to come up with the functionality that would hook into this app intent and you're probably going to do so in a way that privileges your own app, right? So it might make sense to have Siri order you an Uber. You have to know about Uber. How do you know about Uber? You're going to have the dominant service. How do you know about it? If there's no sort of way to touch base with the consumer, well then what are you advertising? Because you're going to have to get awareness in some way. Right? And so I think like there's always going to be some way to undercut that, that sort of framework. And I, I think the reality is like any, any sort of specifically consumer facing company, we're gonna, is gonna want to have their product in front of the user and they're not gonna wanna be just dumb pipes.
- [12:27]
A
Yeah, I agree with you. I definitely agree with you. Although if consumers adopt a sort of a universal agent, some companies will have to, they won't have enough power. You know, you can imagine like if chatgpt OpenAI offered this kind of Uber app and you could order an Uber, Uber might hold out, but then Lyft might say, hey, you know, order Lyft instead and participate. Right?
- [12:51]
C
Sure. But, but Apple doesn't want that. Apple doesn't want that. Apple wants to have an app Store. They make money from the App Store. They're not going to allow ChatGPT to, to have an App store. That, that undercuts them. They're not going to allow it. People have been trying to do that for years and years. No App stores in the App Store. It doesn't exist. There's always going to be a commercial, commercial incentive to prevent that. There's always going else closer to the user in line from you and it's going to be very difficult to leapfrog them. They're going to protect that position.
- [13:18]
A
Okay, what about Apple, what about Apple offering, you know, the ability to order an Uber from Siri? They already tried to do, they're trying to do that.
- [13:27]
C
They couldn't do it.
- [13:28]
A
Right. So I guess a different related question. Why haven't super apps taken off in the West? Why are they popular in China and not anywhere else? Have you thought about the reason that might be the case?
- [13:40]
C
Yeah, that's actually a really good question. Yeah, so like, so, so in, in the kind of, you know, the Asian context, you got these like OTT apps, right? And so they're bundled with kind of everything you use in your day to day life. It's payments, it's search, it's messaging and it's, you know, all sort of social functionality. I, I don't know, it's just that, you know, maybe you have to bring a sociologist on. I, I don't know why it does
- [14:06]
A
seem, it does seem like a consumer, like it does Seem like it might be a psychological thing that Westerners kind of want to separate their concerns in some way. I don't know if that's true, but it feels that way for me as a consumer.
- [14:20]
C
Well, I mean, it could be a function of just wanting to have, you know, that semblance of individuality.
- [14:25]
A
Right.
- [14:25]
C
I want to choose which payments service I use or I want to choose which, which messaging service I use. I don't want to, you know, I don't want to be one with the collective.
- [14:33]
A
It may, it may also be that it was so new, so many of these services that the first, first company to offer, say online payments was also the first company to offer a web portal and email and things like that.
- [14:45]
C
Yeah, yeah. But also Facebook tried it like Meta Stride, it didn't work.
- [14:49]
A
Eric Franchi, like, what's going on in the investment world with the SaaS apocalypse? Is this affecting private valuations and concerns about things like this?
- [14:57]
B
It's the closer you are to being a public company and having, you know, kind of like public comps. Absolutely right. Because that's how these deals get valued on the other side of the barbell, the early stage, you know, AI native seed, not, not seeing it at all.
- [15:18]
A
So let's, let's go a little out of order on the topics we're going to cover and talk about this Apple AI thing you wrote. This piece is the first time I'd ever heard about it that Apple has sort of a transparency framework for AI and that it's starting to actually affect app approvals and people are being declined for apps in the App Store for not giving sufficient privacy or disclosure. I sure, exactly. I haven't seen screenshots of it. Take us through this because probably most of the audience hasn't heard anything about this either.
- [15:47]
C
Yeah. So I mean, I called the piece AI Tracking Transparency. That, that was just me shoehorning this idea into that, into that sort of existing framework. But, but that's not what they're calling it. That was just a clickbaity headline. So essentially what Apple did is in November they updated their developer guidelines and they said that anytime you're sending personal data to a third party AI service, you need to offer adjust in time consent, opt in screen. And so, you know, now if, you know, a lot of apps connect to third party AI APIs. Right. And for any number of reasons. Right. And so the issue here is, and you know, this is like part of the privacy narrative with Apple. The issue here is that Apple has a very nebulous definition of Personal data. They don't, they don't define it in a very concrete way, exhaustive way. And that's, that's just very characteristic of Apple. They never define anything in an exhaustive way. They want to have maximum latitude to apply, to selectively apply their, their developer guidelines. And so this happened in November, but nothing really came of it. I remember hearing about it in November and thinking, okay, this could be meaningful. But nothing came of it. And then all of a sudden they started rejecting app updates for not showing the pop up. And the thing about the pop up is you have to actually disclose. And, and this isn't actually, again, this is Apple being vague, but what Apple has said in the rejection notices, and you only find this on the Apple Developer forums or Reddit, right, which is where I saw it, that they say in the rejection notice that you have to actually disclose which service you're sending the data to. Now, now you might say, okay, well that's great for, for consumer privacy. And you know, sure, maybe it is. That's a sort of judgment. But like the issue here is like well, Apple has Apple intelligence, right? They were going to have Gemini powering Siri, but who knows, are we going to have to click on an consent opt in for Siri and every time we use it? Because presumably you have to, you have to click this consent every time this happens in an app. And so my sense is probably not. And then this does, is this is redolent of att where ATT had, or Apple had a very specific ATT opt in consent prompt for, you know, for apps, for any sort of like app developer. And then they had their own consent prompt for their own usage of that same data for ads targeting. Now they, they do targeting in a very specific way in a way that probably is more kind of privacy preserving than how other people do it. But not, not universally so. And in any case they access the same data and they called what they did personalized advertising and they made everyone else call that tracking as a follow up.
- [18:18]
A
So this is a system prompt that Apple has designed that you have an, some SDK, you have a API call that you have to make as an app developer. Or do you.
- [18:26]
C
No, no, no, no, you can customize it.
- [18:28]
A
You customize it. Okay.
- [18:29]
C
It's only redolent of ATT in that way.
- [18:31]
A
Got it.
- [18:32]
C
And that's actually an assumption. That's right. We don't know. Maybe they, maybe they do the same kind of consent prompt. But here's the, here's the issue. So my sense is they probably won't because how are you going to do that with Siri? Right. It's voice activated. My sense is, what they're going to say is, well, we don't have to do that because it's first party, because we host what they've said and they said this in the Gemini announcement, in the partnership announcement and I covered this a couple of months ago when they actually sort of announced this partnership, but they're hosting this fine tuned version of Gemini for serving Apple intelligence on what they've called private cloud compute. So private cloud compute is this, this sort of like privacy framework that they, that they revealed when they revealed Apple intelligence and what that is like air gapped servers dedicated to doing this kind of processing that have a bunch of different, you know, sort of privacy protections, but they're essentially run by Apple. Right. So, well that's first party. We're not sending it to a third party AI service even though it's operated by Google Gemini. It's, it's, it's first party. We own the servers, we, it's air, you know, it's air. These are air gap servers. There's, there's, there's specific protections put in place and you can go read the white paper for why they can't be hacked. And so. Well, that's also redolent of att. Right. And so my sense is right now they're paying Gemini not that much in the context of Apple's, you know, earnings. But they're billion dollars. Yeah, a billion dollars. But, but the thing is they're not going to want to do that forever. I think they're going to want to make Money on exposing AI models to iOS users and iPhone users and they're going to want to make money probably in the same way that they make with search engines. Right. So they take a re. And so my sense is what they're going to do is, okay, if you want to not show that not or not have that pop up be be sort of positioned between this third party app developer client of yours, AI vendor and your services for which they pay, then you're going to have to be hosted in some sort of partnership program on pcc. And if you're on pcc, then you don't have to show the pop up because that's owned by us, but you got to pay us some share of the revenue that you're getting from these app developers.
- [20:31]
A
Yeah, that makes, that makes total sense. It also makes sense that Apple developer EcoSystem would have AI calls in it for simple AI use. Cases and especially with Apple in its current state not building foundational models, they could just rent seek on all the people who are creating foundational models.
- [20:50]
C
But they do, they do, they do have a foundational model. So they've got the foundational model framework which hosts on the device. Like I think it's 8 billion parameters. So it's small, but it's a multimodal LLM. And they also have the core framework, core ML framework that allows you to sort of upload your own models with your app that are, you know, sort of like they'd be very small. Right. But the problem, and again like the problem with Musk's idea that all the inference will happen on the edge. I mean the problem with the foundational model framework and even core ML framework is that the models that could run on your device are going to be very small by definition. So they're going to be like single purpose because you don't have enough memory. It's not necessarily the storage space, it's the memory to run to do inference on these models.
- [21:27]
A
Right, yeah, I was confusing frontier models and foundational models. So just a final question on this topic, like what are examples? I know personal data is not defined, but in these sort of rejections, have you gleaned what is being considered personal data? Is it any query because you don't know if the query is personal or is it, you know, name, email, etc.
- [21:49]
C
Well, it's any sort of like the canonical PII. But what they say is like freeform, freeform text.
- [21:58]
A
Could, could, it could be, right?
- [21:59]
C
Not be, but I mean it could be. And so they might say like, well look, if you vetted it for that and if you haven't, then you've got to flag it.
- [22:05]
A
Super interesting.
- [22:05]
C
I think what Apple's going to want to do is they're going to want to sort of put their thumb on the scale here in defining it to sort of probably force everyone to have to show the prompt.
- [22:14]
A
Yeah, right. It makes sense that more and more things would be. And ad targeting could be depending on the data that's being collected. You could have an SDK that violates this potentially by sending data for targeting.
- [22:27]
C
Yeah, it could be. Well that's already covered by att. Right. So if you're using any sort of third party ad service, even if it's not SDK based, you're doing it server, Server. You've got to show the ATT prompt. Essentially everyone had to show the ATT prompt. That's what the thing is. They say like, there's all these car bouts. But the reality is if you're running a business from your app, you're going to have to show the ATT prompt. And so it ended up being that the ATT prompt was shown to almost 100% of users.
- [22:50]
A
There's a total side quest here which is not really on the agenda of things we want to talk about. But you know, when I, and I'm trolling on Reddit and on Twitter and other places like that, there seems to be this massive delay in all of the app stores because so many people are vibe coding and submitting apps. I've seen complaints everywhere from Shopify apps to X apps to. I mean, it's just like they can't keep up because there's so many folks in their basement submitting apps.
- [23:16]
C
No, that's true. There was a chart that was going on Twitter and it showed this massive inflection or this like very, you know, sort of significant inflection of app submissions that coincided with some, you know, release probably of the cloud code cli. But yeah, I mean, it's true.
- [23:34]
A
All right, I really want to talk about Manus. So Meta acquired this company called Manus and It was like $2 billion and it was business workflow tools. And I think the first take a lot of people had was it was an aqua hire for just like the AI power tools, not necessari, a functional acquisition. That was at least what I thought. And now they've already kind of launched Manus for the Facebook ads platform as, as kind of an AI solution. And it appears they have quite a bit of, quite a bit of ambition around this. So why don't you take us through this again? For a lot of our audience, they're not deep in Facebook Meta ads. So what's going on here?
- [24:18]
C
Well, that wasn't my first take. So this happened right before New Year's and I posted about, I mean, I didn't. I, I just, you know, tweeted about it. But I wrote that, you know, Wall Street Report, Wall Street Journal reports Meta paid 2 billion for Manus. Da, da, da. And then I wrote it seems likely that the Mana's team could be deployed against Meta's Ads Manager tools or its Advantage plus suite. This could also extend Amanda's newly announced Business AI and Meta AI Business Assistant. If the Mana's team is primarily tasked with developing agentic capabilities for Meta's advertising tools, which makes sense given the company's core focus on agent development using third party models, then it'd be a Strong validation of the value inherent and continued in continued campaign optimization automation. So that was kind of my sense, like my sense was that that's, that's why they were acquiring them. So Manus, just for some background, they didn't build models, they just built agents. Right? They built agents. This is a core focus for, for Meta. I mean, just, just is strengthening these automation tools. Right. And so I mean, I think it's, it's like, it's central to the company's thesis at this moment. So I think there's two really interesting ways this will be applied. Right. One is how they did it, which was kind of, you just, just sort of consistent with what I thought that they were going to do is just applying it to that, the ads manager layer, which just helps people spin up ads, ad campaigns. Right. And that was already mostly automated with Advantage plus, but there were other, there's, you know, knobs that they could automate the turning of and that seems to be what they've done. I think the real value unlock here is applying this stuff onto the publisher properties. And so Meta announced this business AI at, actually at Adweek Week, which I think was, is really interesting. It has been, it has been totally ignored. Right. I had met as VP of, of North America, North American Business on the podcast to talk about this, but I haven't seen it covered anywhere. I, I, I, I, I think it's really meaningful though. But I think they're gonna sort of like very slowly pursue the strategy because I, I do think it's, it's, it's kind of like a fundamental change in the way I think people think about like their relationship with Meta. But what business AI is, it allows a publisher either on the web or in their app to have like a chatbot that kind of picks the up or takes the baton from the ad handover and walks the consumer through the purchasing journey.
- [26:31]
A
Right.
- [26:31]
C
And like, why is that meaningful? Because you can grab a lot of the ad information specifically. I mean, particularly if you're going from the web, you can grab a lot of that from the utm and so you know exactly what they clicked on. You know, like a lot about their, their, the ad that brought them there, there. And then you can sort of on that basis you can personalize the, the, the journey for them. So you can, you can sort of, you can assist with discovery, you can sort of walk them into the product that they just clicked on an ad for or like variants of that. Right. But you know, there's a lot more power that you could bring to that. This is just like a sort of very early application of this idea. And my sense is if you're going to get a lot of, you know, especially like E Com, DTC retailers who would say, look, am I better at personalizing my own sort of product journey than Meta? Probably not. Look how good they are at serving ads, right. And that like some people would reject that out of hand. But a lot of companies, the same ones that sort of like go all in on Meta ads, which is like a lot of D2C, they only buy on Met ads, They'd be happy to turn that over to Meta and they'd probably get a lot better results. Right. So my sense is like this is a way to sort of extend that.
- [27:33]
A
Yeah, this is like next gen lpo. Right. It's like first of all, in the era of AI, there's no excuse not to have have landing pages for every single offer. Every single ad you run really should have a custom landing page. But if you throw in Meta's intelligence and AI into that, that makes a ton of sense. How does this contrast with what Google announced? So Google, at the beginning of the year when they announced ucp, one of the other things they announced was sort of Gemini on your website, but it felt a little more like a SaaS tool. It didn't feel as integrated as what you're describing.
- [28:04]
C
Well, I mean, I think the early versions are very identical. I think what they described with the. They had a very kind of weird name for it.
- [28:12]
A
Yeah, they did have a weird name. Yeah.
- [28:14]
C
I don't remember what it's called.
- [28:15]
A
Gemini on your site or Gemini Enterprise or something. It was really bad name, it was very googly.
- [28:21]
C
But yeah, that sounds identical. But my sense is these things are going to evolve into like very seamless integrations. Right. So where you wouldn't even really understand that you're communicating with like a chatbot because I mean, that's kind of clunky anyway. I think the chat bots is kind of like the easiest way to manifest a lot of this stuff. But the reality is it's going to be seamless, it's going to be invisible.
- [28:38]
A
Google's been sort of reticent to use its sort of rich consumer profiles for this kind of thing. You know, search is very context based obviously. Right. Whereas Meta would seemingly not have much of a problem at all about using, you know, everything about you to improve the shopping.
- [28:55]
C
No, I don't think Meta has much compunction about that. I think, I think they, you know, because the reality is I mean like if you look at their kind of outward facing comms, and I think this is true, by the way, but it is their sort of party line. We're here to serve a small business. Right. And the small business could really benefit from having just a lot more personalization effort applied to not just the onboarding, but everything. Like what if Meta was sending your follow up emails? What if Meta was deciding your pricing? What if Meadow is deciding your catalog management? What if Meadow is managing a CRM for you on that basis? If all those things, there is a very large group of small businesses that would benefit from Meta running that aspect of their business for them.
- [29:33]
A
Well, Eric Franchi, like, we also have a note in the news section about Claude just killed my startup. It feels like Meta just killed my startup. Like how, how can independent ad tech martech compete with these kind of things?
- [29:49]
B
How much time do we have? No, just kidding. Yeah, I think, I don't think independent ad tech MarTech is necessarily focused on competing with or trying to beat Meta at all. Meta is probably unbeatable at Meta's game, which is focusing the SMBs and using their unique data and using their footprint. It's just a matter of can you capture share outside of it?
- [30:12]
A
Right, right. Yeah. I just wonder with like, well, I guess back to Eric. Eric S. So is there any indication that Manus will be. Will help Meta move into other competitive ad platforms? Can you use Manus to optimize your Pinterest spend or your SNAP spend?
- [30:28]
C
Oh, well, yeah, I mean they're leaving it running, I guess as a, as a kind of an independent entity. So maybe I don't think they would allow that. I don't. Unless they were just using that as a way to harvest data.
- [30:39]
A
Yeah, yeah, I'm sure SNAP wouldn't be thrilled with man is logging into their account for them.
- [30:44]
C
Yeah, exactly, exactly.
- [30:46]
B
And that's kind of the point, right? Like one of the things that I'm starting to get obsessed with is there is a class of company and there's no acronym for this yet that is, you know, sitting on top of all of these layers of channels. Right. So it's like social and search and CTV and display and so on and so forth and providing like that layer on top of it in a Manus like way. I think there's a real opportunity there.
- [31:09]
C
Well, but just to, just to answer your question though, you asked of friendship, but I'll answer myself. I think you grow the pie like you're not gonna like. To Eric's point, you're not gonna beat Meta at Meta's game.
- [31:19]
B
Right.
- [31:19]
C
Meta owns Social and we'll talk about that in a second, I think. But like you could. The CTV is pretty explosive right now. Now, like, there's ways to grow this pie. I mean, there's like a lot of really interest stuff happening with like digital out of home.
- [31:29]
B
Right.
- [31:29]
C
I mean, like, there's ways to actually attack those other surface areas that Met is probably not interested in. I know there's a lot of. I mean, I think you said this, Ari, that you correct me if I'm wrong, but I believe you said this. You think Meta is going to expand in ctv? Maybe. I feel like they just don't see much of an opportunity there.
- [31:43]
A
Right.
- [31:43]
C
I think it just doesn't matter to them from like a, like a, a revenue potential standpoint. But like, why, you know, the, these kind of, the same kind of like Meta approach could be applied to those other surface areas in really interesting ways that drive a lot of. And I think that's where I'd be focused if I was running like a startup at this point.
- [31:59]
A
Yeah, I think the Meta CTV thing is, is an interesting flyer. I mean, we saw Pinterest, bought tv. Scientific TV is a big business and Meta could probably do extremely well there. But I don't know how big it is compared to their other opportunities, which are enormous. So social network. So our friends over at Applovin, some of you may recall they made a bid to buy TikTok back, you know, all of six months ago. Feels like years that didn't work out. TikTok's now owned by a random assortment of people. And this week or last week, an interview with a. In Chinese with a Applovin exec talked about their desire or maybe plan to start a social network of their own. Because, you know, why not, why not start with ads and then add. Add content and social connections on top of that because, you know, it'll monetize well and they'll be able to drive some installs. So that's about as much as I know about this topic. What are you hearing, Eric?
- [33:04]
C
Yeah, they had, they had sort of publicly disclosed their intentions before that interview, which is, you know, why the person was allowed to say it. Yeah, I, I think so. First of all, like, they tried this to an extent. So they made a pretty substantial investment to a company called Florida Flip, which was founded by an early Applovin employee. I want to say he was a co founder, but I think he joined the company when he was like 17 or something. So I don't know if he was an official co founder but he was very young and also early and that didn't work out at the company shut down. But, but so they tried. It was, but it was like a social shopping experience. It was like a social e comm company. But there was a strong social component to it. Yeah, I think, look, I mean what does it take to launch a scale social media platform? Well, just look how TikTok did it. They spent a lot of money. TikTok spent a billion dollars in 2018 alone in US user acquisition. They were the number one advertiser on, on Snapchat. They were number one advertiser on Facebook. So you could do it. And Applovin is a cash generating machine. They generate an incredible amount of free cash flow. They have very little debt load. I mean they could finance that and they also have a scaled ad network that they could tap to to drive installation calls. I think like that's kind of like a long term proposition though. I don't, I mean I don't know when that starts reaping rewards and maybe it doesn't need to see any sort of like dramatic you know, early adoption. It could just be like a, this sort of like continuously growing thing. But I mean I think if, look, if you're Applovin, I mean I made the case a while back that Applovin kind of represents a sort of new class of walled garden. They're not a walled garden in the sense that they own the inventory but they essentially do because they monetize it the best, because they have the best date data. Right. And, and because they have this totally integrated supply and demand side. Right. So now they could truly become a walled garden and that, that would also like in, in, in, in enlarge the data footprint. And so my sense is like it, if you, if you look at that from a strategic standpoint it makes a lot of sense if they're willing to finance it. And I think they probably could to the same Tune that, that TikTok did in terms of growth. But you know like social media is really difficult. It's really ossified. Right. Like a lot of companies have tried it. Like the last company that probably came close. Although I will say this, I was in an Uber two days ago and which is like worrying. The driver was on Clubhouse. I thought that was dead. Apparently some people are still using it. But the company that probably came closest to sort of being perceived to have the opportunity to break out was be real and then that was sort of acquired in a fire sale by the French company Voodoo. So it's just really difficult because it requires a lot of money. Well, that's one thing that Applovin has a lot of.
- [35:37]
A
Yeah, a couple points here. So a clubhouse for those who don't remember, was an audio only social network that got popular for like a week during the pandemic. TikTok. I'm sorry? Applovin sold its owned and operated properties recently. Right. I can't remember exactly when. So it's a little strange that they had some owned and operated and now they're going back into it. I also wonder, as a social aspect, as a publicly traded company, if they start putting billion dollar a year user acquisition onto their books, how people are going to react to that? Obviously, there's upside. If it was successful, their valuation would be an order of magnitude higher. But it seems a little. It seems tough as a public company to take on a daunting, hugely expensive opportunity like this from scratch.
- [36:25]
C
Well, they did 1.8. No, they did. So they did. 2.4 billion in free cash flow net of debt service in Q3.
- [36:37]
A
Right.
- [36:38]
C
In a quarter.
- [36:39]
A
That's a lot.
- [36:40]
C
Right? So, I mean, it's not like they don't have money. I mean, they've got enough. They've got plenty of money. And look, I mean, you have to sell this. I mean, look, CEO of AppLovin is very convincing guy.
- [36:48]
A
Right.
- [36:48]
C
And you sell this just like Zuck sold, sir? The metaverse. I mean, sold the metaverse. If you could sell that, that, you know, a charming, you know, person could sell anything.
- [36:56]
A
Well, his stock suffered quite a bit on that one. Have you had Adam on your podcast?
- [37:00]
C
No.
- [37:00]
A
No, me neither.
- [37:01]
C
Adam.
- [37:02]
A
Either one. Eric, or mine. Come on. Or both. We'll both. We'll both do it. Okay. Well, we. There's so many topics, we could probably talk for another half hour. We're going to switch to the news section. We have a lot of news that's really relevant here, including a lot of funding, the trade desk earnings, Catalina acquisitions. You know, there's a ton to talk about. So we will take a break and we'll come right back.
- [37:28]
C
Close your eyes. Exhale. Feel your body relax and let go of whatever you're carrying today. Well, I'm letting go of the worry that I wouldn't get my new contacts in time for this class. I got them delivered free from 1-800-contacts.
- [37:42]
A
Oh my gosh, they're so fast and proud. Breathe. Oh, sorry.
- [37:46]
C
I almost couldn't breathe when I saw the discount they gave Me on my first order. Oh, sorry.
- [37:51]
A
Namaste.
- [37:52]
C
Visit 1-800-contacts.com today to save on your first order. 1-800-contacts.
- [37:59]
B
All right, everybody. And we are back with the refresh, the news of the week bunch of stuff to go through this week. Let's take a break from AI and apps and talk about consumer data. So Infilon, for those not unfamiliar with Infilon, it is a kind of like a private scaled private ad tech company that has acquired assets along the way, most notably MediaMath out of bankruptcy. They made a big acquisition this week. So Infilion acquired Catalina. So Catalina, if you are not familiar, it's like deterministic purchase intelligence Data. They're in 70 retail locations, they track 130 million households, 600 billion annual spending like literally at the cash register. So it's a really interesting asset with purchase data. Why is this interesting for Infilion? I'll start and I'd love to hear you guys pile on. So it's thing one for any DSP or programmatic company, having unique inventory or having unique data is becoming increasingly important. This is unique data, at least at this point. It's going to be kind of widely available. I would assume at some point it becomes specific and exclusive to Infilion. And then number two, this is in the release, Infilion has a ambition to kind of become a power of RMNs, basically kind of build your RMN on infilion. This is unique data that they can use to infuse that type of offering. So it's cool from two standpoints, right? The unique data angle and kind of empowering the next gen of RMNs.
- [39:39]
A
Yeah, I have some experience with Catalina. When I worked at Nielsen a long time ago they had a joint venture called Nielsen Catalina. You know, Catalina was known at the time for having loyalty card data. So they had, they had quite a bit. It's opt in but it's very substantial amount of shoppers in various retailers and they get SKU level data and then they link that back to profiles. It's a really valuable data set for me. The, the interesting thing here is how did Infilion win this deal? Because you could make the argument of Catalina's synergy with a lot of other companies, some of which have very deep pockets. The trade desk, you know, the trade desk will get to their earnings, but it was soft on cpg. You can, Yahoo, with its DSP has, you know, some data but a little bit of a deficit. There's so many other companies that could have bid on this and I'd love to know the tick tock on that. What.
- [40:39]
B
Guess what we're going to have an answer for because Rob Emrich, the exec chair of Infilon, who's the guy behind all these deals, he's coming on the pod in a few weeks.
- [40:49]
A
I, I guess I knew that. Right. I should have known that.
- [40:53]
B
You need your coffee. I was waiting for the punchline to come.
- [40:57]
A
Yeah, yeah, exactly.
- [40:59]
C
Yeah.
- [40:59]
A
So we'll, we'll have Rob on.
- [41:00]
B
I think it's a super interesting deal and increasingly a new kind of, kind of neat company. Eric, what'd you take on this?
- [41:05]
C
Yeah, can I, can I ask with the, the Nielsen jv, what, what was like, I mean I can, I can kind of conceive of that data. But like, can you, can you walk us through like a, like one example of a use case where, where that was used by like a marketing company?
- [41:18]
B
Yeah.
- [41:19]
A
So Nielsen has a different shopper panel that is opt in where users actually scan their, their receipts and they've had that since like the 1980s. And so the idea was you take Nielsen's data and Catalina's data, put it all together and then, and then it was like step three, profits was more or less. You know, they each, they, the primary business was selling back to CPG companies about market share and you know, volume and things like that. But they wanted to use it for consumer profiling and targeting and television advertising and all that other stuff pretty ambitiously. And my understanding is that it didn't go that well actually that there were, it was, this was, you know, probably early in the days of big data when it wasn't, you know, things like lookalike modeling weren't as common and they had some scale issues because like even if you have, let's say Catalina has 20, 30 million dollars, 30 million households in the U.S. and Nielsen has 130. Yeah, it didn't at that time, but yeah, so like the overlap wasn't that great. And I don't think it was very successful. I think they unwound that deal. So now Catalina was sort of an orphan at some point and that's how this ended up happening. That's a very short version.
- [42:30]
C
There's a lot of interesting research being undertaken right now about, in like the topic of like digital twins. But a lot of that is, is being directed at like the marketing use case. Right. So it's like I, I could, I could see that data being used to construct like purchase proclivity adjacencies for like some specific profile. Right. So it's like, okay, like, look at all this data. And then if I'm a user who's bought this thing, like, what else should I have advertised to me, you know, based on like historical purchase patterns? That seems like a really interesting application of that data set.
- [43:06]
A
Yeah, there's that. What's the name of the other company that does the receipts for supermarkets where they use that data? So you buy Coke on their seat. It says buy Pringles. I can't remember the name of it. I think they're also a Nielsen partner. I'm just kind of totally blanking on the name. Anyway, readers hit us on social and tell us the name because it's a very well known company.
- [43:27]
B
All right, all right, let's move on. You want to talk about ttd?
- [43:31]
A
I think we have to, sure.
- [43:33]
B
So TTD had earnings yesterday. The company reported a good quarter. Sales were up 14.3% year on year to 846.8 million. On the other hand, next quarter's revenue guidance of 678 million was below analyst estimates. The stock went down. I feel like talking about the stock.
- [43:57]
A
I hate talking about the stock. I hate talking about the stock.
- [44:00]
B
Kind of get gets us nowhere. You mentioned you dug into the transcript a little bit. What did you, what did you get out of it outside of the, the softness on cpg?
- [44:11]
A
Well, I didn't dig into the transcript, but I kind of. Various people sent me stuff. This is much easier than doing my own work. The, the thing that had. There's. There's two stories that people are. Or three stories, I guess. One is take rate increased. So the take rate, meaning gross bookings divided by their revenue, the 20s. And I was in the 21s. And my reaction to that is, well, any dollar that now goes through OpenPath should increase take rate. Because in the old world you're paying 10, 15% to an SSP. In the OpenPath world, you're paying 5 or 10% to the trade desk. So you expect take rates to go up even though it looks bad. And people love to hate on that. That is explicitly the strategy of openpath. And the, the promise to customers is that it should cost less in total. Right. Because you're consolidating margin. But people definitely. I had some DMs about, hey, do you see that take rate? I'm concerned about that. And it also, unfortunately for the PR department, the trade has kind of feeds into the other conversation going on about open path and hidden fees and transparency where it was announced that like Dentsu and wpp, I think Adweek said, quote, quietly exited openpath, unquote, for transparency reasons. So there's a lot of talk and pressure against the narrative that Trade Desk is trying to push, which is transparency and a complete ecosystem.
- [45:41]
B
Got it.
- [45:42]
C
Okay.
- [45:42]
B
So take rate. Going up was number one. You said there was another thing.
- [45:46]
A
Well, the other thing was the open path kind of. Kind of thing. And then the third thing was this stranger announcement about Ventura. I don't know if we want to cover that. The television operating system. The company put out an announcement about what they called the Ventura ecosystem, which I interpreted to be a bit of a reversal or retreat on the idea of selling a complete operating system. And it seems like they're more pitching to oss to use their SDK and use the sort of a layer of advertising instead of of a complete OS switch, which obviously is a lot easier.
- [46:23]
B
Yeah, yeah. And just get access to demand. It seems pretty straightforward and plays into their strengths. Back up on the. On the quietly exiting open path. So there was an article in Advertising Week or Ad Week specifically naming Dentsu and WPP with, you know, revenue going up and take rate going up. I wonder how much impact that's had on Q4 or if this was a Q1 thing and we might actually see some of the impact of impact in Q1.
- [46:51]
A
Yeah, I mean, it's a good question about if OpenPath is accretive on take rate and on margin, then you lose share in openpath and you lose margin. Right. It makes sense. I put a tweet that got a lot of reaction about this, which was that I think we should be skeptical about this take from Adweek's headline, which is that they left because of hidden fees and transparency. I pointed out that it's very possible that Dentsu and WPP have preferential financial arrangements with various SS SSPs, sometimes called supply path optimization agreements. And they may be getting rebates or kickbacks on the amount they spend on SSPs, in which case they wouldn't like Openpath because they would not get those kickbacks. And then they might spin that as a transparency concern. But then, you know, Adam Heimlich pointed out that there is a lack of transparency. He mentioned that they don't specifically give guidance on OpenPath in the raw log files they provide, which is interesting. And him, as a partner of theirs, would probably have a lot of knowledge about that.
- [47:57]
B
Yeah, that makes sense. And I think, you know, kind of transparency and rebate aside, the other thing that might be happening here is just this shift towards sell side Decisioning and some of these like unique products. Speaking of Adam Heimlich, like Chalice has that interesting integration and partnership with index that made the news because Dentsu was using that to power some of their blockchain buying. So it could be a bunch of seismic shifts happening.
- [48:23]
A
Yeah, if you're, if you're using a curated deal ID then you're not using open path by definition.
- [48:29]
C
Can I ask two questions? This is not a company I, I follow super closely. Certainly not quarter to quarter but, but like what, what did you make of the. So first of all the stock being down is, it's not really indicative of anything. I mean Nvidia just annihilated earnings last night and their stock's down 5% today. But it's just like this is, this is a moment where, where everyone is scared and, and they don't really know how to interpret good news. Although the trade desk earnings were kind of weak or the guidance was weak. But what do you make of the executive turnover that feels like that was maybe not a positive signal? I mean there's been quite a bit of that recently, right?
- [49:04]
A
Yeah, yeah. The two different CFOs in the last six months have left. It's definitely something people are concerned about as just this general feeling like they're not putting their best foot forward into the PR world, both for the Wall street investor relations and the industry as a large. Which is why, you know, I'm so excited to have Jeff Green on stage at Market Extra Live where I'm going to have to ask him, you know, what's going on, how are we, what is happening here? Hopefully I'll get some good answers out of that.
- [49:36]
C
Can I ask one more question? Do you think is so like. So Jeff Green called out in the earnings call that you know, Amazon's not that serious of a competitive threat and Google was far greater over competitive threat. Is that comm strategy working? Like, I just feel like maybe people don't believe it.
- [49:53]
A
People do not believe it and a lot of analysts have been harping on that. I think it's probably true ish in that I think the trade desk is really focused on the sort of middle to upper funnel marketers, the TV budgets, what you might call the traditional 500 global advertisers. I think that's just like the vast majority, majority of their revenue and, and the Amazon's coming at it from the other direction from the smallest guys upwards. And Amazon, other than Fire TV is probably not as strong in ctv. So I could see the argument that they're not losing head to head against Amazon as much. I think it's also interesting on a. I've always liked to look at market share in the DSP market like, you know, by gross spend, where is the money going? And it used to be that it was DV360 and then a long way back was trade desk and then Amazon was like zero. And now from my sort of first of all, the tradesk gave a number. I think they said 13 billion in media spend in the last 12 months. I think that's right around where the other two guys are. I think it's like a three way tie at this point for market share. More or less with some rounding. They're all in the 10 to 10 to 13 range. Let's say maybe updated DV360 data maybe is different because of YouTube, but it's around the same. And then you have this giant gap to number four which is probably Yahoo DSP at a billion or two. So it's a horse race.
- [51:24]
B
No, just to talk about the Amazon point, that was a few quarters ago where Jeff said that and then the analyst community jumped all over it. And then subsequently where the criticism has continued is Amazon just going after market share via lowering their take rate. Right. So to your point, all right, if they started at zero, I don't know when they were at zero, you know, and then now they're at at parity. Taking that amount of market share means that it's a competitor.
- [51:55]
A
Yeah, it is. I think though, if I would guess that Amazon's taking a lot of share from other parties like you know, retargeting companies and you know, small retail media companies and they're just, they're just marching through like, like, you know, the Huns coming over the Urals. So it's not, it's definitely something that Wall Street's very concerned about, certainly.
- [52:17]
C
But he did, he did say it, he said it this quarter too. I mean he said, I would say. Let me explain why people want to talk about Amazon a lot as it relates to competition. I've said before I think Google is a far better competitor than Amazon is today. Which is kind of like saying, saying the sort of like the heyday of the competitors to us is over.
- [52:35]
A
Right.
- [52:38]
B
All right, let's move on and look forward to seeing him on stage with you. My goodness, talk about AI. So two interesting fundraising announcements this week. So the first was profound, which is a Geo AEO company is now a unicorn. There is 96 million in series C at a billion dollar valuation. For AOGEO, this is pretty impressive.
- [53:04]
A
Yeah, it is, considering I've barely heard of this company. We've talked to that profile like 20 times.
- [53:11]
C
What are you talking about?
- [53:12]
A
All the Geo AIO companies run together. The one thing I will say is that last week we had Nate Elliott on and he talks to marketing folks and AI folks all the time. And the one thing stuck with me from that conversation was he said every CMO is obsessed with this subject. How do they, how do their brands appear in AI? And that's a, that's a buy signal if, if there's ever been one.
- [53:40]
B
Yeah, it absolutely is. We, we'll, we'll link to the posts. Basically, the post was just a, you know, an X thread by the, by the founders, but they say that 10% of the Fortune 500 are using the platform. They're rolling out agents to be able to basically take some of the insights from the answer engine, turn it into usable content and do so automatically. So it's pretty impressive. It's also like a really, really hot category. I mean, like, I'm continuing to see deal flow for this category. We're continuing to see fundraising for this category. It's super important. And then just the question is, you know, is there going to be one winner or a few?
- [54:20]
A
Is there a switch in costs? Is really my question. Like, if I sign up with Profound for a year, I get my insights and then, then the next year I'm like, oh, Bluefish looks pretty good. It feels like there's no switching costs except maybe historical data. Also, all these companies have switched. They're not optimizing anything anymore. They're basically analyzing their, their analytics insights companies.
- [54:40]
B
Yeah, and then, and then now, now action. Right. So actually taking these insights, turning them into content, having that content, you know, better fuel the, the answers that people get. You know, a bit of a closed loop. This a category you're playing, you're paying a lot of attention to arc.
- [54:56]
C
No, I, I just wonder what happens, you know, so if you want to appear in, in chatbot results, Chat GPT has good news for you. You can buy ads from them. I, I, I, I, I, like, I, I wonder how that kind of undermines the, the, the efficacy of, of these kind of approaches. But like, no, I agree. Like, I, you know, I know that it is a massive, massive initiative for like a lot of skilled companies. They want to appear there. My sense is though, like the chatbots themselves have incentive to get you to pay for that.
- [55:31]
A
Yeah, yeah. I also, I really wonder about this category. It Feels as though, like there's so much uncertainty that the CMOs just need a friend, they need a corporate friend to be like, we're going to handle your AI stuff and we don't know exactly what that is. It's changing every day. We'll just keep building shit for you. We'll become the, the HubSpot or the salesforce for AI. And don't worry too much about what category it is or what budget it is. You need a enterprise vendor to be side by side with you.
- [56:02]
B
Why are you questioning that? Sounds like a pretty damn good business to be in.
- [56:05]
A
Yeah, no, I think that's what it is. I think the actual features aren't as important as that. They have like, oh, we raised at a billion dollar valuation. We have the enterprise salespeople you used to work with at Company X. X, it's. Yeah, yeah, get it.
- [56:21]
B
Okay. There's this other one, Koa Koa is being called the adsense for AI. I don't know if they're calling themselves that. Somebody's calling them that. That's. That's pretty big.
- [56:31]
A
Their investor called them that in his, in his podcast.
- [56:35]
B
So they raised they, they raised a. A 20 million series A by, by Theory Ventures. Here's.
- [56:41]
C
So what do they do?
- [56:42]
B
They help chatbots and I guess long tail LLMs monetize. Right? So you know what, the picture that eric painted for ChatGPT, like, can you build a network? Can you build a business to power this? The thing that I'm trying to understand is what is the real opportunity in the long tail of LLMs, the long tail of chatbots. I think it's interesting in that there's vertical ones being kind of coming to market. I think there's a lot of utility in these maybe niche and vertical LLMs. But I just remember this, this anecdote that I heard from the open ads guys. So Stephen and Michael, most folks are familiar. They're like awesome, brilliant and super prolific. This was their original idea and they very publicly and visibly pivoted because their sense, and this could be a timing thing, is that some of the vertical and niche LLMs are not good for advertisers. They're like AI girlfriends. They're like weird kind of niche chats. And it's not necessarily a category that is at one scalable or at two kind of kind of advertising friendly. But this thing is moving fast and maybe now the market's ready for it. I don't know.
- [57:56]
A
Do you have a sense I'm chomping at the bit to get in on this one. So I already wrote, I was so worked up, I already wrote a full newsletter coming out on Monday about this subject so you could hear my thoughts there. This is a terrible business plan. It's a terrible business plan, right? Because basically it only works if there's a fragmented long tail of AI agents. And most of the fragmented long tail of AI agents have no context. Like AI Girlfriend is not a context that, where AI ads make sense. Applovin will be able to monetize your AI girlfriend better than any AI company could. And then if you're like in the vertical space, monetize your AI girlfriend. That sounds really bad, but the, but then if, like if you're in the vertical space, if there's some emerging leader that's like AI for travel or AI for shopping, they're going to run their own ad platform. That's what happened in apps like Pinterest, Snap, they all have their own ad systems. As soon as you get scale, you want your own ad systems. And if there really was an opportunity for this long tail of AI companies, then the critical mass and scale on the demand side would be the winner. And that would be Google or OpenAI. Right? There's no way to square this to make this a big business. With that said, I could almost guarantee they're going to make their money back on this investment because someone will buy them just, you know, for a couple hundred million dollars in a couple years or maybe even a bill. You know, I, I, I'm not going to suggest that KOA itself will not be successful. It's just a bad business plan.
- [59:25]
B
Yeah, Eric, what you got?
- [59:27]
C
I'm going to drop everything I'm doing and build the company that monetizes AI girlfriends.
- [59:31]
A
That's, that's a brilliant business proposition. Turn your AI girlfriend on the street, earn a little extra dollars called Trick AI or something like that. I think we might have to cut that producer Mike.
- [59:52]
B
Anyway, let's, let's move on, let's move on to, let's move on to Family store Walmart. Bring it back to the content that is appropriate for the audience. So Walmart's ads business is right ripping. So Walmart Connect grew 41% in the US as reported in Q4. Earnings, I figure is up from 33% in Q3. So this thing is going, I mean, it's still a ways to go versus Amazon. I think it's like a $6 billion business versus a $60 billion business. But they're moving in the right direction. I'm super impressed. And obviously Vizio's a real tailwind for this.
- [60:32]
A
Yeah, 41%. That's a really impressive number.
- [60:35]
B
Stop tasting market.
- [60:37]
A
Someone asked me this question the other day and I really don't have a great answer which is how much of retail media is growing the pie versus share shift. Share shift meaning like you know, you, you have coupons and Sunday circulars and end caps in the store and all that stuff. Moving all that to digital, you know, is great and it's good for the retailers. But are we growing the pie is like the entire advertising market for these products going up and taking away from let's say TV or, or more products being sold. And I have sort of a hard time answering that question.
- [61:09]
C
I mean I think that you think in theory that it should be because these are just much more efficient ways to, to serve ads to people.
- [61:19]
A
Right.
- [61:19]
C
I mean they should perform better. They should produce you know, more results and more sort of like economic output. So you'd assume that, that, that show you just this analog ad. But, but, but yeah, I mean that was, that's always been the argument for rmn. It's like well that's taking like events budget and that's taking like in store budget and that's just shifting it to digital. It's not really growing the budget. But my, my sense would be that I think in Walmart's case it must be. Some of that must just be compounding, right? Because like well I was spending this money on in store discovery before but now I'm getting like 30% better outcomes or 30% more revenue that's like attributable through you know, some model that I have. Assuming I have a model that, that and so I'm reinvesting that. My sense would be like some of that growth is just a function of, of like digital is a much more efficient path to, to sort of driving commerce than these, these analog approaches.
- [62:07]
B
Exactly. And at the rate that it's growing, it's growing faster than either other segments of digital. So like one of the theories that I have is that it's pulling budget from just like standard display where it's less accountable, less, you know, less efficient, less performant because it's just growing at an outsized pace. And I think that's validated with like all of the momentum around like off site retail media. It's just becoming a better segment of programmatic from a performance standpoint.
- [62:36]
A
It's pretty hard to tease it back because the consumers are shopping more online. So there's consumer behavior that's switching as well. And some of these dollars would have been spent on in different trade budget ways. So it's a little tricky. But I think Walmart, given their scale and also their Vizio acquisition, has the pole position here. I mean, secondary to Amazon to, to really grow the pie.
- [63:00]
B
Absolutely. Want to call it there?
- [63:02]
A
Yeah, let's call it there. This was an amazing conversation. I definitely. Eric, whenever you're on, I learn a lot. This was particularly interesting. A lot of topics that are very relevant and cutting edge. So where can we find you? We should, where should our listeners find you? They should subscribe to your newsletter.
- [63:17]
C
Yeah. So I just. Something that I'm kind of focused on right now is this, this podcast series that I'm working on called the Prosperous Society. Right. So I wanted to have, have. I wanted to, I wanted to sort of put forth a counter argument to all the AI doom and gloom. And I'd been working on this long form podcast for a couple months. It was up to five hours of runtime and I decided as a result of the Citrini blog post that I was going to cut it up into episodes. I released the first episode this week and it's, it's done really well and the responses has been, has, has, has been really positive and I'm really excited to continue releasing episodes in this series. I think it's gonna six, six episodes. But I'm just sort of like making the case for why the application of AI to the digital economy is going to be a boon for consumers and boon for society at large. That's kind of what I'm focused on right now. That's just a free podcast. You can find that on mobile Devoid.
- [64:14]
A
What's it called?
- [64:15]
C
The series is called the Prosperous Society, but it's just under the Mobile Dev Memo podcast. So if you search for a mobile dev memo on Spotify or wherever you consume podcast content, you'll find it. But other than that, you can always find me on X. You can find probably too much of me on X and Mobile Dev memo dot com.
- [64:32]
A
All right, well this was a great conversation. We'll be back next week.
- [64:37]
B
Thank you for subscribing to marketecture.
- [64:39]
A
New interviews are added every week at marketecture TV and your favorite podcasting app.