← Marketecture: Get Smart. Fast.
Loading summary
Transcript117 lines
- [00:00]
A
A word from our sponsors. This is Ad Tech God and the 2026 IAB Annual Leadership Meeting is where the digital media market aligns, bringing together the leaders who shape investment, innovation and growth across the ecosystem. It's a rare opportunity to build high value relationships, surface new revenue paths across AI, commerce and measurement, and stay ahead of the shifts redefining how business gets deep done. Be in the room where the year ahead takes shape. Join us in Palm Springs February 1st to 3rd and register@iab.comalm use code ATG ALM26 to save $500. Again, go to iab.comalm and use code ATG ALM26to save yourself $500.
- [00:57]
B
This podcast is brought to you by Meloko if you're an app marketer concentrating most of your budget in Google and Meta, you're not alone. 88% of consumer app ad spend stays inside walled gardens. Meanwhile, there are billions of daily users spending their time elsewhere in gaming apps, dating apps, productivity tools, you name it. That's where Maloco comes in. Maloco's AI powered platform connects you to those users real reaching 2 billion daily users across 3 million apps ready to diversify. Visit Malocoads.com and start scaling with confidence again. Go to Malocoads.com to learn more. Hi, welcome to the Market podcast. This is Ari Poparo. I'm joined today by Eric Franchi and our guest is Erez Levin, who is a former Googler and he calls himself a media futurist. And what he's very on top of and is going to share with you is the issues around inventory, quality, ad quality. How do advertisers get the most out of programmatic and advertising and where their problems are in the ecosystem? Eric, do you think quality matters?
- [02:20]
A
Of course it matters.
- [02:21]
C
Yeah.
- [02:22]
A
What a loaded question. Yes, of course it matters. Yes is a very complicated and nuanced conversation depending on where you're coming at it from. But having read some of his stuff and spoken to Erez, he's a very thoughtful guy around this and he certainly brings a pov. If he's a media futurist, what are we?
- [02:45]
B
I am a media past historian. I'm a dinosaur. Just call me whenever you need to hear how people fucked up something in advertising ten years ago and I have the answer. What are you?
- [02:59]
A
I don't know. This desktop guy moniker seems to be sticking that Jim Payne bestowed on both of us the other day.
- [03:07]
B
That was very nice of him. Eric, do you want to describe the TikTok I sent you by Slack this week.
- [03:12]
A
Oh, goodness.
- [03:14]
B
We'll put that into the show notes.
- [03:16]
A
You didn't put it in the show notes. You sent me a slack of an Asian woman. Seems like a young Asian woman who went to Wawa for the first time speaking in her native language.
- [03:30]
B
It was Japanese.
- [03:33]
A
And just. It was like the greatest thing she had ever experienced. It was like going to Disney World. Just like the sheer wonder of a gas station convenience store. And she bought a sandwich and she thought it was good. She shared it with her dad. It was so cute. And apparently this young person is. They just have an account about all of the beautiful things about New Jersey.
- [03:57]
B
It's New Jersey.
- [03:57]
A
They discovered it.
- [03:58]
B
Yeah.
- [03:59]
A
It's incredible. I haven't spent more time on it, but it was, like I said, the most wholesome video I've ever seen about a gas station.
- [04:06]
B
This young lady is touring New Jersey and translating it for her Japanese audience to the background of the song ymca. And she speaks very quickly and it's a wonderful thing. If you've ever doubted the beauty of New Jersey, we will put it in the show notes.
- [04:23]
A
I did not think we were going here, by the way. We just lost a bunch of folks.
- [04:27]
B
Yeah, everyone's tuning out. Okay, so the Market Live agenda has been released. It's still. We're still adding some great speakers. So go to marketlive.com to see some of the highlights. We have Sophia Kolucci, who's the chief marketing officer for Molson Coors. We have Lance Armstrong, who's now an investor, the GO investor, Joanna o', Connell, who everyone knows, who is incredibly insightful. I'll be speaking as usual. And Eric and Sonya will be doing the startup showcase. What's the timing on getting the letting people apply for the startup showcases, like next week or so imminent?
- [05:09]
A
Yeah, over the course of the next week or so, we'll get the announcements out, the criteria, the form. Everybody can start submitting. We can't wait for this one again.
- [05:19]
B
Yeah, we had like 60 applicants last time for five spaces. I think we're increasing it to six spaces this time. So we look at the startups to think about who's the most interesting for our audience. And it's a great forum to get in front of hundreds, if not a thousand or more advertising people and talk about what you're working on. We've definitely heard from folks, they got business out of it, they got funding out of it, et cetera, et cetera. So it's all really great. So markitecturelive.com, buy tickets now. All right, let's get into it with Erez Levine, the media futurist, talking about quality. All right, we're here with Erez Levin. Thanks for being here.
- [05:59]
C
Thanks for having me.
- [06:01]
B
I pronounce your name Erez, like the land of Israel. Is it Erez or Erez?
- [06:05]
C
It's Erez.
- [06:06]
B
Yeah, Erez. Okay, sorry. We dove into a little bit of Ivrit here. So I wanted to have you on because you are very active on social, you're very active in general. You're a very passionate person about advertising. And why do you care so much? What's your beat? What. What is. What are you excited about? And why do you care so much?
- [06:31]
C
Why I care so much? I think there's probably two sort of angles to this. I am. I consider myself a marketer by nature, sort of. It's what I studied. It's what I've always loved. And so I think just from the business sense, as I look to what marketers do and how they spend their money, I think it's very ineffective, it's very wasteful, and it has been increasingly so over a long period of time. And then there's sort of the second dimension, which is, what are the implications of that waste? What's the implications of their dollars going towards lower quality, ineffective media instead of higher quality media? And that has a lot of sort of societal implications that I think are socially destructive. And if you think about the enshitification of the web and the consolidation of big tech monopoly power and the erosion of journalism and independent media, those are all sort of direct through lines from this. And so I think the better question than why do I care about quality is why doesn't everyone care about quality? Sure.
- [07:30]
B
Okay, so what is the state of quality in, let's call it open web display advertising?
- [07:37]
C
For now, I think about this as like a tale of two markets. I think we have a large part of the buy side that still cares about quantity, volume, sort of vanity metrics, cheap reach, things like that. But we have a sizable and growing part of the buy side market that cares more about quality. And they're getting more support from vendors that are either playing in both sides or ones that are increasingly focused just on sort of like finding differentiating between higher quality and lower quality media.
- [08:10]
B
Right. And what do you define as quality? Or maybe it's not up to you to define the quality. What's the marketplace defining its quality?
- [08:17]
C
Uh, it's really multidimensional, so there's not sort of one thing. Um, I think the sort of like main thesis is not all impressions are created equal. And I think, you know, there's, there's a lot of different angles to this. But ultimately quality is really defined through the lens of effectiveness over most of the time, the short and the long term, not just sort of the short term. What you can, you know, a click or 7 days or 30 days and what you can do attribution on most ads actually have that effect. Maybe they're going to drive some people that are in market to buy something immediately, but a lot of times it's actually just getting somebody to, you know, increase mental availability and brand awareness over the long term.
- [09:00]
B
Well, hold on, you're kind of like stealing my thunder because you know, the gotcha question was going to be like, isn't outcomes really matters, not quality? And now you're just saying you just define quality as outcomes. So like, maybe I don't care about quality, maybe I only care about outcomes and quality just comes out of the.
- [09:18]
C
Wash. Quality absolutely is only defined via outcomes. So they're not detached from each other. The issue that I have with outcomes specifically, and there's actually a lot of issues that I have with outcomes, the game non incremental sort of become a vanity metric. There's a polysimus term which it can mean so many different things. But I think the, maybe the core thing, even when an outcome is like the real thing, is that they are almost inherently looking at a single point of time. And so usually advertising has an impact in the short and the long term, right? The same, you're serving that to a hundred people, only five of them are in the market and they're going to make a purchase. But the other 95, it had an effect on them. And so you have to think about the quality of the media and be able to measure the impact in both the short and the long term across all of these people that you're going to reach.
- [10:08]
B
So in your worldview, if real human consumers spent their time on really, really low quality websites, but it generated purchases, that'd be fine.
- [10:19]
C
It depends on the brand and their objectives. So yes, that's sometimes fine. But a lot of times like brands also have to be able to, you know, like when I go to a store and I need toothpaste, I'm looking at that aisle, I'm looking at the shelf. And the three brands that I'm most familiar with are going to be probably the ones in my initial considerations. Same thing when I need a car, I'm not always going to Google it. I'm not always going to find that at that moment. And so having that sort of brand awareness, mental availability ahead of time and is what's always been probably the most impactful thing for, for at least big brands.
- [10:52]
B
So tell me how you got into this. You. I think you and I met when you were on the DV360 team. So you were, you were part of the Googleborg but you were very much focused on helping buyers. So it was not really like hardcore ad tech role. It was, it was like a, it was, I'm not sure exactly how to characterize it. How did that experience turn into this crusade on quality?
- [11:16]
C
Well, it actually started when I was on the sell side at goog. I was on the ADIX team, on the buy side of Addicts team. And so I was servicing all the DSPs, connecting Adix into the DSPs including DV360. And that's where I started to see sort of this disconnect like lower quality media being, you know, sort of capturing more and more share, you know, like weird video, small little video ads, muted ads, trying to capture, you know, more and more, not just dollars but also at high CPMs. And so as I kept like pulling back the layers and then switched to DV360, I was in charge of inventory as sort of a product specialist. And what I was trying to do is sort of like fill gaps as, as the landscape evolved, how do we make sure that buyers were getting the best effectiveness? And again, I studied marketing. Like I know the basics of it and I saw sort of those things getting violated. Like those things were not incrementally valuable but they satisfied some vanity metrics. And so I was able to be successful in a number of ways. It's sort of like plugging those gaps and getting quality sort of filled, whether it was measurement or buying tactics. But the problem just kept getting worse.
- [12:20]
B
And so yeah, let's talk about the video thing because that was the one that was really where you were pretty active and the IEB was active. So I guess take us through what was the problem with video declarations for a while and what did the IEB do and is the problem solved?
- [12:39]
C
The problem was that the definition for in stream was vague enough. It said that, you know, it has to be in the stream of content. And so that was like, okay, it was YouTube and things like YouTube but then all the small little video players that play in the bottom corner as long as they could play, yeah, it's outstream but as long as you could play a Video in between in that stream. Like content, you could call it in stream. And that's what everyone did. And so they were sort of capturing more and more budgets at high CPMs are basically in stream prices. I went to the IAB, that working group, I suggested a change to that definition. They changed it. Now it has to have the sound on to be considered in stream. And they did a second signal there in parallel with a time for transition. Unfortunately, there's still a lot of that running through the ecosystem. A lot of small muted ads calling themselves in stream that buyers are paying a premium for. And I think that's devaluing the real in stream out there, whether that's CTV or YouTube, because it just seems like there's an infinite supply of this one. There really isn't.
- [13:48]
B
Yeah. Is that so? Generally the markets move to devaluing what's sometimes called OLV online video versus ctv. And so there is some good online video that is in stream, but it's getting kind of thrown out with the bathwater. Is that, is that what you're saying?
- [14:06]
C
Yeah, I mean, I don't think there's that much good online video, to be honest. I mean it's not, it's not bad, it's not worthless. Like I'm, I'm like, I don't want to throw it out. The outstream has some value and every buyer, based on how they assess, like how important is the audio for your message getting across? Are you a big brand? Are you a small brand? Are you just trying to like get sort of lots of little like awareness over time. For Coca Cola that might be great. For another brand, not so much. And so you, it's more for every brand to figure out what's the value of quality, the dimensions of quality that matter to them and how do they value that for themselves.
- [14:39]
B
So this change though had real world impact. I mean if you think about the companies known for web based video, they haven't really done very well. You have Kinetics and JW had to merge. You had Teeds and Outbrain had to merge and you have Tremor, which is still out there, but not a skyrocketing stock. Are you. Is that an accurate assessment as far as, as far as your point of view on this?
- [15:05]
C
It's related and there's nuances in each of these. But some of these companies, like the legacy outstream companies, I made the case to them forever that they were the ones getting the most screwed. Right. Like quality publishers that had in Stream that all of a sudden had this small Stuff calling itself in stream too were getting devalued but so were the ones that were honestly calling themselves outstream and then had somebody that basically had the same exact format, maybe even worse, calling itself in stream and stealing budgets that should have gone to outstream. And so yeah, this, that's all related.
- [15:35]
B
Do you think outstream video is a viable format moving forward?
- [15:38]
C
Absolutely, at the right price. So I think there are like not all out stream is created equal. I'm very adamant about interstitials. You know, like you get wi fi at the airport or you know, games is a weird space. But I do think that there should be a more explicit trade off of some, you know, somebody has to pay for the content somehow and so sometimes it's going to be watching an ad, sometimes it's going to be paying a subscription or a micro payment or something else. And so I do think those are especially viable.
- [16:08]
A
So okay, so maybe you answered my question which was specific to mobile games and this is just top of mind because we'll talk about this in the news in a few liftoff found their S1 applovin continues to be just so dominant and a valuable company largely on the basis of interstitials prior to games. So from your perspective, because you think about this stuff, what's quality, what's not, think that's an acceptable use case for that type of format.
- [16:38]
C
I do try to break apart a little bit. There are certain use cases that make a lot of sense for a focus on outcomes where quality isn't as important. If you are an app game and you're just trying to churn through like getting somebody to download their markets. Right. Beauty brands is another one that's really interesting. There's a lot of people on TikTok that just see an interesting thing and they're just going to buy it. And so there are plenty of use cases. Marketing is like there's no universal truths here but I think for the majority of marketing and especially on the enterprise marketer side, a focus on quality is more important.
- [17:16]
B
Yeah, I mean if, if the web could magically put required video watching in front of all of its articles and people actually listen to it, they would do it would do really well. The problem is that so much web traffic is, is one and done and consumers won't sit for it. So there's gotta be some sort of compromise UX there.
- [17:36]
C
I agree and I think for quality content, especially in the AI era that we're entering like only things that are truly unique and valuable. Like if you Just have a turkey chili recipe like no one is going and watch an ad or pay you for that. But if you have something unique, you build an audience. People will stay, people will pay some amount. Right. To. In order to get that.
- [17:54]
B
The turkey chili thing is interesting because I'm actually gonna have that for lunch. But actually it's more interesting because people are willing to scroll through thousands of words about their grandma's recipe to finally get to the stupid recipe and get exposed to 10 annoying ads. But they wouldn't sit through.
- [18:11]
C
They're ignoring that, right? They're ignoring those ads for the most part.
- [18:14]
B
They're ignoring them. Yes. Those ads are not willing.
- [18:18]
A
Like I think ERA said that said the right thing. And again, we'll talk about this on news. Yeah, Sites like that are probably not long for this world.
- [18:27]
B
Yeah, I get that. All right, let's. What are, as we sit here right now in, in 2026, what are the top quality problems in digital advertising?
- [18:38]
C
I think most of the things that we hear about, like most of the issues in the industry are all quality problems, like the SPO and curation and MFA and video declaration. Like, all these things stem from a lack of focus on quality. Those stem from something I think, that goes even sort of like upstream of that, which is marketers caring too much about the short term, not looking at incremental outcomes, too much focus on vanity metrics and binary metrics and things like that and assuming probabilistic models are deterministic. And so I think it's. Almost all of these things are very related.
- [19:12]
B
But those are all symptoms. What are the causes of the inventory problems? Like, like a year ago we were all, we couldn't stop talking about MFA, MFA, MFA, MFO. And it feels like MFA is not everyone's top concern right now. Maybe it is. What, what are your customers or your clients as you go around consulting and talking to people, talking about.
- [19:34]
C
I, I think the, the challenge is, is obviously like quality is not a, there's no single, like silver bullet to solve this thing. And so you can solve like one element of mfa, which is adds to content ratio. But then there, and there's, there's, there's always going to be some new creative things coming up. You're talking like root causes. The root cause is the fact that marketers don't generally care about, about quality. And so lots of lower quality stuff is what gets packaged up and sold, especially sort of in this open auction world. But figuring out how to, to address quality and think about sort of all the dimensions of it, that's, that's the big challenge.
- [20:12]
B
And what are the tools that the marketer should be using? I mean, verification is there, but then verification is mostly a content quality issue. Right. That doesn't affect RTB quality and, you know, data quality and things like that. So what's the toolkit you recommend?
- [20:26]
C
Verification is part of it. I think quality needs to sort of play a role in each of the sort of like, you know, parts of the media buying cycle. I'll give a little bit of a plug here and a tease on something that I'm working on that I'm really excited about. So I'm working with sim, which is the trade body, the Coalition for Innovative Media Measurement. We're working on an initiative that'll sort of like try to help the industry. Quality is a big buzzword and no one really knows what it means. So helping the industry try to understand and navigate this space, understand different ways to look at quality, to understand quality. What are the different dimensions? How does media quality play in with audience quality, with creative quality? And then how do those different touch points, like what are the different solutions in market from planning and, you know, activation and optimization to media measurement, which I think is in the verification side, which is great, but then also it needs to get into the effectiveness measurement place. Right. So mmms and experiments, not so much mta. There are use cases, I think, where multi touch attribution can still do a decent job, but it's much smaller than it used to be. And so because to the point earlier, we don't, we can't define quality without sort of knowing that it actually correlates to the outcomes. But those outcomes are mostly determined with MMS and experiments, not with multi touch attribution.
- [21:49]
B
Right. And we're going to talk later in the new section about agentic and ADCP and stuff like that. But as a first pass, if you're a buyer and you use AI to effectively push your media plan over to the sell side to execute, aren't you losing a lot of control over quality?
- [22:08]
C
That's where my. One, what I'm skeptical about with agentic is like, unless quality is really baked in and quality has to sort of going back to the earlier point, you have to think about it for effectiveness in the short and the long term it's possible to do. I just think it's going to be so much easier to just sort of like ignore quality. Focus only on something that you can measure in those first 30 days. Like that's what the system is optimizing to and so unless you have this two like dual time horizon baked in where quality you're both optimizing two short term outcomes but also thinking about the proxies to long term at the same time, then I'm really not optimistic. It's possible. I just think it's going to more likely to be abused.
- [22:49]
A
Yeah, yeah. This is an interesting conversation. It's like everything is moving so fast. On the other side of everything that's being built with Agentic is all of these next generation contextual brand suitability products and ways to classify pages and ideally embed them into an agentix sell side or buy side platform. So it feels like we're speeding down a road that can address everything and help everything. Or more likely there's going to be some minds that we all step on on the way, just like with the history of everything. But it feels like all of this stuff can be largely addressed with AI, which makes me very optimistic.
- [23:34]
B
Yeah, incentives are the challenge, right? They can be. But now you're telling your AI prompt like hey, and here's my include list and here's what you shouldn't do. And here, make sure it has a verviv. This podcast is supported by Walmart Connect. Hello marketers. At Walmart Connect, you'll find things no other retail media has. Not only do they have the most shoppers buying all kinds of things like cleaning supplies, video games and cookware, they get how shoppers tick and why they click. And they can help you connect with them at any point in their journey from discovery to purchase. Because your bottom line isn't just about views or engagement. It's about the most valuable truth in advertising, that your ads work. So spark innovation, let's spark growth, let's spark sales. With Walmart Connect verification score of X on this, whatever, it gets pretty complicated. All right, we have a lot of news, so let's take a break here and get into the, you know, the overwhelming amount of AI news as well as some fundraising and then a potential ipo, which is exciting. So we'll be right back with the news of the week.
- [24:42]
A
All right, we're back everybody, with the refresh. We got a bunch of fun stuff this week. Let's talk about AI. So CES was last week. We had a good recap on all the agentic announcements last week with respect to ad tech. This week, strangely starting on Sunday. Sunday, Yeah.
- [25:00]
B
I want a day off, man. I'm watching football. I do not want to be tweeting on Sunday.
- [25:05]
A
Nobody has explained the rationale for Having the NRF starting National Retail Federation starting on Sunday, but it started on Sunday and mixed in with maybe the CES wraps and the beginning of NRF was a bomb that Google dropped with their AI announcements. So there's three things that I think are interesting with respect to what they announced and have. They're commerce related, but certainly have broad applications for advertising. So the first was the launch of UCP. So we had ADCP, which was a protocol for agentic advertising. UCP is a protocol for checkout in agentic LLM experiences. 20 something companies, including some of the biggest besides Amazon, are participants in this. And basically when you're searching a product and you're in AI mode, you can very easily have a seamless checkout experience. So the vision for this is like pretty awesome. And then there's a bunch of downstream things that can happen. But first let's get takes on ucp. Is this in any way not a good thing?
- [26:11]
B
No, this is a great thing and it follows on a bunch of announcements that OpenAI made with PayPal and Walmart and some other folks. So the model here is that the consumer is looking for products. You have a product catalog. So Google has a big advantage here because they have, they already have every, every SKU in the world in their database with all kinds of metadata about it. So you're finding products. And then what this protocol potentially allows is that when the user clicks, I want to buy it. You give a lot of control, surprising amount of control to the retailer who can make offers, who can collect credit cards, who can, who can kind of break free of the Google monolith here and own the customer to a great extent and potentially have a agentic agent that's making decisions about what offers to give that customer in real time. So it empowers the retailer and the payment processor to do smart, interesting things and to not be cut off from, from that experience.
- [27:10]
A
Yes, exactly. And those are the two other products that I highlighted that I thought were very interesting. The first was direct offers. Basically advertisers off define offers that Google AI can use to increase conversions. So you're handing over Google some of the power here, which I think is controversial. I tell you, you know, having debates around this on X, but basically it's like, you know, if the AI judges this is a person who will convert with 15% off, 20% off, free shipping, whatever it is, they inject it and it could lead to, you know, obviously like, you know, higher conversion rates seems like a no brainer. The other thing that you mentioned Which I thought was very interesting is they have two things. It's in the direct offers link, which we'll put in the new newsletter is a brand agent. They also released a business agent. I think it's one of the same. Basically it's a little bit of a customer service chatbot that will allow a brand to have a conversation with a user again as part of this whole experience to ultimately answer their questions, put them in the funnel and then you know, obviously drive.
- [28:16]
B
But this is on Google, so this is important part is like they're giving the brand some power over the user experience in Google Gemini, not on their own website. So that's really exciting. So somebody is looking at a sweater from J. Crew potentially J. Crew can have some influence on what conversations happen about that sweater. On the Google experience, which is I think pretty smart model. And Google really showing some flexibility here, not totally trying to dominate every aspect of the experience.
- [28:51]
A
Yeah, I think that's right. And then it also I think tips the hand of what everybody's feeling, which is, you know, website shopping might end up being a thing of the past or perhaps a lower percentage of how commerce gets done today versus people just staying within AI environments and being able to have everything done there. Paul Silver from miq, I don't want to skip too far. He thinks this is incredible potential catalyst for advertising by the way, which is if commerce goes agentic advertising and upper funnel advertising like CTV out of home become that much more influential in basically all those things that it should do from a brand perspective to help, you know, increase influence. So it's pretty fascinating what we're seeing here.
- [29:40]
B
Yeah, so I had some follow up conversations about this because I put my takes in my newsletter and there's two interesting things to think about, two sort of rivers of thought here. The first one is that product discovery still firmly is in Google's hands here. This protocol doesn't really help you with getting your product in front of the consumer in the first place. It's after the consumer has shown some interest that you can potentially check them out. And discovery is where the money is. So Google's obviously keeping that to themselves. Where you will pay to potentially as part of Google shopping to have your product more, more prominent. And the second thing is this question about when there is some transfer of kind of identity between the distribution partners, meaning the LLM and the retailer. If you think about a simple use case like retargeting on the web experience, as soon as the user puts something in a shopping cart, you can start retargeting to them. Assuming you have a cookie, where is that line where the Google or the OpenAI starts telling you as the retailer who this person is is after conversion, before conversion, because that determines a lot of what you can do from a marketing perspective.
- [30:55]
C
One thing I. One thing I'll just add on. Yeah, just on both of these. The way I think about a lot of this stuff within AI and agentic, especially as we get into ads, but even before ads, it's like, where is this adding value to the consumer? And a lot of times I think AI agentic, it's shortening the shopping discovery process, like making it a lot easier. But then if you're thinking about where does it add where do ads help the business as well while still helping the customer, giving them the best answers? That's where you start to really narrow down. So I do think discounts are like probably the best use case where advertisers can sort of get ahead of the organic answer. That might be the best one because they're saying this is actually better because we're giving you a discount. But otherwise I don't see as many use cases, not in the near term for sort of satisfying both advertisers and consumers in these AI spaces.
- [31:50]
B
Yeah, you have comparison shopping, which should be really super powered by AI, but we haven't seen it yet. You know, obviously probably everyone who's listening to this has done some shopping through an AI and it's pretty good for like high involvement, like B2B kind of products. But telling me what sweater to buy is not there yet. And that's a big area of investment.
- [32:12]
A
Agreed. Shall we move on?
- [32:14]
B
Yeah, let's keep going.
- [32:16]
A
Maybe this might be even bigger than these announcements. So Apple announced that they're going to be working with Google Gemini to power Siri. So did Google just win AI? I asked this because at least until we come up with another form factor, mobile is everything, right? Like, let's just like put it there. A good percentage of the world, I think, uses Siri or the Siri equivalent, hey, Google. To find information, you know, do what they do. So Google Gemini is going to power all of it. There's 5 billion mobile users. Basically all of them either use an Android or iOS. 3.5 billion Android, 1.5 billion iOS, all powered by Gemini. Like, this is big. The question is how big?
- [33:07]
B
Yeah, I think there's a lot of questions here. First, it's a big accomplishment from the Google team. It's a big loss for OpenAI. I think OpenAI is losing across the board right now. So there are some, I got some insights listening to TBPN and Ben Thompson on that show. One thing, a couple of things to note first is Siri penetration in terms of like active users among iPhone users is really low. It's like a quarter. Right. So that's open white space that Apple has ceded because their product's so bad. So this is potentially a big onboarder for ordinary people to AI who are not currently using it every day and very actively if Siri actually gets good. The second thing that I thought was super interesting is like, yeah, it's a billion dollars a year that Apple's paying Google. You know, that's nothing dropping the bucket for these companies. What gets super interesting is when I forget whose insight this was, not mine is when does this switch where Google starts paying Apple. Right. You know, because Android, Android's not free. Android's less than free. It's, you know, Google pays the, develop the handset manufacturers and that's where this is headed. All right. It's just not there today. And also it helps with the government antitrust folks if it happens. Apple pays for it like a software product. But long term Google pays Apple for this thing.
- [34:36]
A
Yeah. Just like they did with the default search.
- [34:38]
C
Yeah. Until Google is monetizing Gemini from consumers who will be the primary users of this. Like they're not going to be able to pay Apple for this. And so it's really interesting. And this, this, the, basically the, the nightmare scenario that after the monopoly ruling didn't have any sort of teeth and let Google just basically try to replicate the same dynamics in the AI era. Yeah.
- [35:00]
B
And it's already immediately after the monopoly finding there were follow up hearings about Google trying to push YouTube and other services on the handset manufacturers. So they're, they're a little bit cautious but they're not cowed by any means. And just the story, I think the vibe in Silicon Valley right now for everything I read and talk to people about is just the implosion of OpenAI there. You know, Claude took all of the professional excitement away from them. Consumer side, there's some stats, I don't think we have them in the notes that their penetration is declining pretty rapidly versus Gemini and just the feeling like their models are not, are not there and are slower and are not not up to speed for the cutting edge doesn't look so great for them right now.
- [35:46]
A
It doesn't, it doesn't. And that said, we've seen just back and forth, you know, so, so long and everything can change with one, you know, incredible ship. The thing that's on my mind is like, what is the future of like consumer interaction? Is it going to be with hey Google and hey Siri? Hopefully I didn't just like trigger some, something on my devices or is it going to be open up my chatgpt, Claude, so on and so forth. That's the, that's the real question here.
- [36:14]
B
And that's where Apple. Yeah, that's where Apple by owning its own chips potentially has a big advantage because baking those start words into the chips to be as fast as possible and to get the results as fast as possible matters enormously towards consumer adoption.
- [36:30]
A
Agreed. Final point on Google, just a quick one. There's a article in Prescosette that grabs some chartbeat data from last year. We should link to it. It's super interesting. There's a bunch of stuff here, but basically validating that traffic to publishers globally was down a third last year to you know, via Google search and you know, forecast it to continue to drop. Interestingly it's much higher in the U.S. so U.S. down 38%, Europe down 17%. So if Europe just catches up, you know, you see the, the acceleration there.
- [37:07]
C
It's, I mean we know what it is. I mean going back to, I can, I, I can always tie back to quality. Yes, quality publishers have unique content that people want to go and read. They don't want just the summary that you know, AI overviews will give them. They want to go read a long form article. They want to go not just get the recipe but hear about this, you know, creator. It's a vegan person and they're talking about their kids and how they love it. Like that stuff will exist and survive and people will value it. But the, the sort of commoditized content that first, you know, Google can steal or AI can take without paying for it, that is going to be summed up and that's gonna, there's gonna be a huge market for that. So that's gonna be less traffic for these publishers. But the one, the remaining traffic that's loyal, that really wants to go to their sites and values the full content, can find ways to monetize them, I think better.
- [37:58]
A
Yeah, absolutely. In this article and again, very much worth reading, this is where it points to some of the categories that might be ripe for real disruption. That is the recipes versus others. Anyway, let's move on. Bach dropped an article. Brian O', Kelly Scope 3 founder with an interesting perspective that got everybody all fired up. Basically Making a analogy that agentic AI is if this was finance about portfolio allocation, whereas RTB is for the quants and the day traders. And it's a great article, you should read it. Everybody seems to have read it. Everybody's got a perspective on it. But his POV is that RTB has limited use cases. It's big market obviously, but for some of the non digital advertising investments and even some parts of digital that's not RTB based, RTB has either limitations or can't touch it. And there's a much larger market opportunity there that can be addressed with agentic and kind of likens it to portfolio theory which is using agentic. If you can expand the set of investments that you have, you can optimize more when your goal happens to be outcomes or quality or whatever it is. I thought it was very interesting, very thought provoking, got people fired up. I don't know if you guys have a, have a take on it.
- [39:25]
B
Yeah, I do. I, I think people are sort of in some cases intentionally misinterpreting what he's saying. Right. But effectively the whole point of ADCP is to say what you want and then to have the publisher deliver what you want. Either the publisher or an ad network do that. And, and the benefit is it potentially can do that at scale automatically and for potentially smaller buys which were always burdensome to do it like that, you know, do direct buys with small publishers. So that's the benefit to potentially an agentic media buying. And, and so he's trying to up level it and say well you know, that is a, the way media planning should work is that, you know, there's some stuff that, you know, like cookie based and price very performance based that always would be better on a spot market. And there's other stuff like you know, upfronts and TV and stuff that's always been better on a futures market. And ADCP is just rationalizing that. Whereas for the past, you know, 15 years of programmatic, there's been so much effort to try to actually go the opposite direction and say, well just make all your inventory available on RTB basis and we'll figure out a way to make things like programmatic guaranteed or you know, reservate artificial reservations work. Right. And that's been a bit of an uphill, uphill battle, especially on things like live events where the actual pipes are not really good at doing it. So, so that's the, the way I interpret his conversation and I effectively agree with him. I think that it's insightful what he's saying, but I think that you also have to sort of de average it and say, you know, how much of this is true for open web display where you know, probably there isn't as much demand for this portfolio approach versus like TV where it's mostly a portfolio approach.
- [41:20]
C
I like that interpretation and I'll add to it because what I don't like the analogy. It's like RTB versus scp. It's like this binary. They're one or the other. It's a good analogy still. But ultimately, as I think about these things, when you look at large swaths especially of supply, the quality difference, you know, there's a huge range. Look at the open web or even look within the walled gardens. And that's the critical piece of making sure within a portfolio. That's the problem with MMMs, they're not granular enough. And so as you're looking at these things, whether it's with ad, CP or mmms agentic, it doesn't matter what it's starting to think about like not valuing all of this supply within a bundle the same and thinking about its effectiveness.
- [42:01]
B
Yeah. And I think that it is interesting to see what happens on the buy side. So we're talking about ADP as advertisers wanting things and doing things. But the buy side technology companies like the Trade Desk and Yahoo, DSP and the other DSPs are pretty quiet on this subject and because I think that this model goes directly against where, how they make money. And I'm very interested in watching what the reaction is going to be from someone like the Trade Desk on how this works. I mean Yahoo actually has been making some noise about it, but I'm not, not entirely clear how it fits their business model.
- [42:39]
A
Yeah, I think that's right. Kind of final thought on this. Zuck said a year or so ago, AI done right will improve gdp. But it's another way of saying it will radically grow the TAM of advertising because everyone will be able to sell more, be more productive, so on and so forth. Bach makes this take that this is actually not performance, but this is the way to grow the TAM of advertising because it helps to properly value media based on whatever metric right outcomes. And we see, you know, perhaps even because of things like ucp, a shift towards brand advertising, advertising powered by digital and ends up being, you know, this, this TAM growth. I think that's an interesting theory.
- [43:29]
C
I, I struggle, I struggle with that one with, with the TAM and with the percentage of gdp. Like unless people start buying more things Then those things are probably going to kind of stay the same. Grow Cokes tan, you know, like, like their revenue, but not Coke and Pepsi's together. And so I think that's the challenge.
- [43:48]
B
Yeah, the total amount of purchasing from consumers doesn't change. So if, if, if advertising became more efficient, it would have to come at the expense. It would either be make companies more profitable because they could spend less on advertising and get the same amount of Coke being purchased or it would steal share from less, less performant activities. The consumer they that Coke does and that could be as far ranging as stocking supermarkets, as trucks, as employees. But in no case do people drink more Coke than they did previously.
- [44:26]
A
We always end in this like weird theoretical macroeconomic moment.
- [44:33]
B
Brian Weiser is the only person who understands us. We should have him back.
- [44:36]
A
Maybe, maybe there's, maybe there's room for Diet Coke. I don't know. All right, let's move on. Let me ask you guys, prior to this week, had you heard of an ad tech company called Apple Cart?
- [44:47]
B
I had because I had done some consulting that related to them, but. Yeah, but that was the first time I'd heard of them.
- [44:53]
C
Okay.
- [44:54]
A
Not a well known name in ad tech circles. This is a company that has seems like both a self serve and a managed serve approach to reaching super influential people. Whether that's politicians, influencers, celebrities, power brokers, you name it. It's been around for over 10 years. Business was started in 2013 and I encourage everybody to check it out. We'll put the link. Seems to have a pretty sophisticated way of reaching individual people. Whose, I don't know, AOVA could be infinite somewhat like when Musk was saying you should advertise on X because the most powerful people in the world are on X. Basically there's an ad tech company that does it. They raised $100 million from Blackstone at $500 million valuation. I am so amazed.
- [45:48]
B
Yeah, it's a eye popping number from a company most people haven't heard of. I don't know if the valuation is justified based on the financials. Yeah, basically their elevator pitch is you want to influence this one person who is a senator. So you can't really advertise banner ads to the senator, but you could, you know, advertise to the 200 people most likely to talk to the senator at a cocktail party, including exactly that person's college roommates and you know, and stuff like that.
- [46:22]
A
It's crazy sophisticated stuff.
- [46:25]
C
Yeah, yeah.
- [46:26]
A
And that's basically what they say on the website and shout out again to Vibe. Maybe this was a co. It probably was a coincidence because this was a physical outdoor billboard. They had this awesome billboard in San Fran that basically said, hey, advertise on Vibe. You can, you can reach Mark. And it's a. An image of Mark Andreessen's very famous, you know, kind of, kind of head which is bald and, you know, kind of quite large. I was dying. I thought it was awesome. Another viral moment for Vibe. But it's in this same kind of category, right?
- [46:56]
B
Yeah. Ovoid, I guess would be the word for Mark Andreessen's head. I pointed that out on Twitter and that's why he blocked me. So he's going to block this podcast next if, if we keep talking anyway. All right.
- [47:08]
A
Want to cover off Liftoff and then we'll. We'll call it.
- [47:10]
B
Yeah, sure.
- [47:11]
A
All right. So we have a potential ad tech IPO. Liftoff filed their S1. Liftoff is a combination of Liftoff, which was tech, and Vungle, which I think was more of a mobile ad network. Rumored to be raising 400 million. Looking to raise 400 million in the IPO. Some stats on the business. It's 140,000 apps use its services to effectively do the same thing I think that they do with Applovin, which is grow by advertising their apps on other apps. 2025 revenues of 519 million. Net loss of just over 48 million and it's carrying over 1.85 billion in debt. Interesting to me is, you know, I tried to really parse through and as a, you know, kind of quote unquote desktop guy, it's a little bit hard to figure this stuff out. The difference between Liftoff and Applovin beyond just like size and scale. It seems that Liftoff may be a little bit more demand centric and app because I don't know if they have the mediation where Applovin has the mediation, which I think creates this big network effect. But there's smarter people that mobile that can certainly correct me on this.
- [48:19]
B
Yeah, so I didn't read the S1 yet, but effectively historically liftoff was the DSP and mobile DSP and they were known for their algorithm, but also they did a lot of creative services. And I think that is a common through point of a lot of these mobile ad businesses which is, you know, how do you create these really compelling ads that interstitial or click to play kind of things to make make the result work. And Vungle was also in that business. So Vungle had an SDK. When they were independent, they had an SDK. They were probably top 10 mobile ad networks. And they also did a lot of creative. They would do like these cartoon videos of your app. So if you were an app that wanted to advertise on Vungville, they would create like a cartoon version of your app to explain it in 5, 10 seconds. And that would drive conversions. And they're both, I think, acquired by KKR and merged together. And yeah, I think this is very in the line of applovin, smaller, obviously. And maybe I don't think they have this unified auction like Max that applovin acquired. But otherwise they're in the same ballpark as Maloco and. And geez, what's the company? No, AppsFlyer is just measurement analytics.
- [49:43]
C
Right?
- [49:44]
B
AppsFlyer is just analytics. It's that LoopMe is the other company that's probably a little smaller than this. They're all doing similar things in app performance advertising with some mixture of services that's not entirely just algos.
- [49:59]
A
Yeah, makes sense one to watch though, because more successful IPOs is only a good thing for ad tech. And with that you want to call or you want to cover this less antitrust thing we put in there?
- [50:11]
B
Yeah, there's not much to say here except everyone and their cousin is now suing Google. So this week we had. I only have Atlantic the magazine and Penske, which owns Variety and a bunch of other services. They both sued Google for antitrust reasons, but there was a couple of others that came out yesterday. I don't even have their names in our notes. It's really becoming a bit of an avalanche of folks who are diving in because the Virginia court already declared them in a monopolist. So really the lawsuits are all about damages, and it could be a free pretty large check coming in to any of these publishers. I think they probably are hoping to be consolidated in some form so they don't all have to do their own cases. The biggest case is coming to trial, I think next month or March, which is the consolidated case with Gannett and the Daily Mail in it. And then the people has a very large case. So it's just more on this Google antitrust stuff keeping the lawyers very busy.
- [51:10]
C
Enough said.
- [51:11]
A
All right, well, this has been a packed episode. We should call it here. Eris, thank you so much.
- [51:15]
C
Thanks for having me.
- [51:16]
B
Absolutely. Thanks.
- [51:16]
A
Eris, thank you for subscribing to marketecture.
- [51:20]
B
New interviews are added every week at.
- [51:22]
A
Marketecture TV and your favorite podcasting app, Sam.