← Marketecture: Get Smart. Fast.
Marketecture: Get Smart. Fast.
Episode 148: Paul Bannister Shares Research on How Consumers Feel About AI-Generated Content
00:42:05
Loading summary
Transcript140 lines
- [00:00]
A
This podcast is brought to you by Maloco. 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, 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. This podcast is brought to you by Adelaide Media. Verification and measurement are undergoing major disruption. Legacy players are pivoting to performance. Advertising, AI is reshaping, brand safety and attention is replacing viewability. Adelaide is leading the shift with au, a new way to assess media quality that scores placements based on their potential to drive attention and outcomes. Before your ads run, think of it like a credit score for media. Finally, a clear view of quality. Before you buy, take the guesswork out of your investment strategy and try Adelaide AU on your next campaign. Welcome to the Market podcast. This is Ari Paparo. I'm here with Eric Franchi and our guest today is Paul Banister, who is the Chief Strategy Officer for Raptive. Eric, you know Paul pretty well, right?
- [01:38]
B
Yeah, we've had him on before. He is a go to around all things publisher and publisher monetization and just with, you know, everything moving so quickly these days, I think he's going to be a great person to dig into. And he did a job. Great presentation. I think that you saw that inspired this.
- [01:54]
A
Yeah, yeah. Like last Thursday, I was double dipping. I was at the Roku Summit and then I was also at the Bidswitch summit. They're like a block away from each other. And at the Bid Switch Summit, Paul gave a presentation of some original research on consumers and AI. I love original research. New insights I hadn't seen previously about how consumers react to AI content. So I asked him to come on the show and to present some of that. So it should be pretty interesting.
- [02:23]
B
Awesome. And then we've got a bunch of news. As usual, earnings wrap up and maybe, you know, kind of like a quarterly overview. We should talk about what we've learned. We've got AI, we've got a couple of random stories as well. So this would be a fun one.
- [02:35]
A
Yep, let's get into it. So we'll be back in a second with Paul. All right. Welcome Paul Banister, the Chief strategy officer from Raveptive and I think a two time guest on architecture. Is that right, Paul?
- [02:48]
C
Sounds right to me. Yeah.
- [02:50]
A
All right. Yes.
- [02:51]
C
Happy to be here. Thanks for having me.
- [02:52]
A
I think after three times you get a robe. That's what we're thinking.
- [02:56]
C
Oh, nice, exciting.
- [02:56]
A
Or a blazer robe. The rope. I think there was a joke on Saturday Night Live where Steve Martin had a robe. I think that's where it's in my head. It got, it got accepted. So, Paul, we wanted to have you on because you're obviously one of the kind of deep thinkers on the publisher side in advertising. But also I saw you last week at the Bid Switch event, or maybe it's two weeks ago. I know it's all blurring together. And you had done or had collected a bunch of research on consumers and AI. And I wanted you to kind of walk us through what you found because I thought it would be interesting to our audience.
- [03:30]
C
Yeah, definitely. Happy to talk about it. So we probably have like two big data sets that we pull together. One is a big survey that we did of 3,000 US adults, demographically representative, blah, blah, blah, all those good things. And it was really about their views of AI content and what they think about it. That's like bucket one of data we have. And the other data set we have is we have a lot of traffic data across, you know, 6,000 sites. And we actually looked at a group of, forget the exact number, but around 4,000 sites that we've had traffic data on for at least the last three years. So we have a lot of information about like traffic changes by vertical and the different kinds of sites and how AI has impacted different areas and things like that. So it's kind of like two different things that we, we pull together into kind of one big story. I'll talk about the survey first. So in the survey what we did is we had across multiple vertical verticals of content types. We had human generated content. There was like real stuff that you find on a website today. And then we had AI generated content. And it was not like a slob. It was good AI stuff. It was like wrote it like it was like pretty good quality things.
- [04:39]
A
Okay.
- [04:41]
C
And we then in the case of the survey, we sometimes labeled the content as either human generated or AI generated. So people knew what it was and other times we didn't label it. And then so basically four cases, AI generated, human generated, labeled or not. And those are your kind of scenarios. Then we showed those to all the people and we asked them questions about like, what do you think about the Trustworthiness of this content. What do you think about, like, do you feel connected to it? Do you feel like you understand the author? Do you feel like you, you know, would tell others about. There's like a whole bunch of different questions like that. We also integrated advertising into the content and we asked people their thoughts of the ads in the content and what that was, and we had a couple of different learnings come out of it. I think one thing that was interesting was while people, when the content was labeled, people did trust AI generated content slightly less than human generated content, but not terribly less. Like it wasn't a huge gap there. And again, I think part of that was because we were, we gave them like, content that when you write it, you're like, this is pretty good. Like, it was not junk.
- [05:38]
A
Right.
- [05:40]
C
And so like, you know, when things were labeled, clearly people felt good about things. When things were not labeled, people were very suspicious. They really didn't like it. And they, and to some level, while they couldn't label it correctly, in a lot of cases, like we asked them, like, do you think this is AI generated or not? They were very bad at guessing whether it was or not. Interestingly, even when they were wrong about what it was, their, their emotional feeling was worse with the AI generated content than the human generated content. So it was like a very interesting subconscious impact on people in terms of how they, how the data showed up for that. And you know, so that, that point we called kind of like the, the AI stink. If you were suspicious that it was AI generated and you weren't clear about it, your, your trust levels went down like 30, 40, 50%, like really big drops in like how you thought about the content. And, and the same held true for ads integrated into the content. Like, so that was kind of its own big problem there.
- [06:32]
A
And Paul, that's because you asked them, you asked the survey respondents, like, do you think this is AI generated? Or something like that.
- [06:39]
C
One of the questions, although to that point we asked them often, the, the order of the questions was randomized so many times they got that question much later in the survey after they already answered a whole bunch of just like feeling questions. And so it was kind of, it was pretty interesting in terms of how that shook out. So, right.
- [06:57]
A
So AI labeled content. If it was AI labeled, people felt a little worse than human labeled. Right, Right. But if it was not labeled and they suspected it was AI, that gave it the stick.
- [07:10]
C
Exactly. So that was kind of a very interesting thing that that came out of it. The other big interesting Thing that came out of it was is this point that we assumed that as people become more experienced with AI, like these sort of issues might go away because it might be like, okay, I know what I'm getting here and I understand, I feel good and whatever. And so one of the questions we asked people was kind of like where they are on the spectrum of like, AI usage, whether you're total newbie, you know, been using it for a little while, or like, you know, heavy user sort of question. And the interesting thing was that this like, lack of trust issue kind of looked like a curve in that data where novices felt like, didn't trust, didn't trust. It felt bad. Like those numbers went down. People who were like, somewhere along the pathway where they were like moderate AI users, they actually were like, this is good. Like they felt good about the content and that stink issue was like far, far less for them. But then the most experienced actually swung back up to almost the same level as the novices, where the people who were really AI savvy were like, yeah, I don't trust this stuff at all. Like, I'm totally suspicious. So that was kind of the other really interesting learning was kind of this curve that came out of the data in terms of like, trustworthiness based on how experienced you are with AI.
- [08:25]
A
Yeah. I want to just point out that, you know, AI has been generating content for like 10 or more years. Like, you know, sports scores and like stock articles have been written by AI. Like well before the LLM era. I remember seeing there was a narrative content, I think was the name of.
- [08:44]
B
A startup in Chicago.
- [08:45]
A
Narrative science.
- [08:46]
B
Yes.
- [08:47]
A
I don't know if consumers realize that, but like, it's been pretty common among the kind of like, commodity writing for a long time.
- [08:54]
C
Yeah, yeah, 100% agreed. Although, like the scale at which it's happening right now and the, and the content area is just far, far higher than it were.
- [09:03]
A
Yeah. Did you find any differences by vertical? Was it like with sports news, et cetera?
- [09:08]
C
So we asked a question in terms of where are you most or least comfortable with AI generated content? And, and, and to some level, like some of it was pretty obvious, like the, the type of content that people were most comfortable with, AI generated content was simple, factual lookups. Sure. Yeah. Like, if it's just a fact, like, I don't need a person telling me that, like, AI will summarize that. Great. And be really good on the reverse side. Like, people wanted personal stories, like they wanted a human to write it. Like that makes sense. But Then there was a whole bunch of inconsistencies, like the second most, the second highest vertical in terms of people not wanting AI content. So wanting human generated content was parenting and family. Which kind of rings true. Like you learn something about your baby, like you kind of want like a person to tell you about it. But the interesting thing is it doesn't hold up in our other data set which where we have all this traffic data. And one of the verticals that's been hit hardest from what we see at least from traffic loss due to AI is parenting and family. And so people are saying that they want human generated content, but doesn't really seem like that's actually holding true. They seem to be fine to ask ChatGPT, like how do I put my, you know, colicky baby to sleep or whatever, whatever the question is, which is pretty interesting.
- [10:21]
A
I'd be scared to know what ChatGPT would tell me about my baby. Like, well, so let's transition. So we've seen reports that similar web put out some really aggressive numbers, like 30% to 40% declines in traffic on major news sites like CNN. Other folks have sent, you know, 10, 15%. So what are you seeing on traffic?
- [10:42]
C
So everybody's a slightly different story. It is, you know, the, the, the two best indicators we've seen are content vertical matters a ton.
- [10:51]
A
Okay.
- [10:53]
C
You know, sports has been hit hard, travel's been hit hard, personal finance like things like that, that are like quasi factual in, in kind of nature, those have been hit pretty hard. Food, which is one of our biggest verticals, like not really hit so far like it seems some people are using chat GPT to get recipes, but not that many it seems like. And so like that like vertical is a big factor. News we haven't, we don't have as many news sites. My data there is not amazing, but from what we see, news hasn't been in a store. Partly because search already hasn't been a great traffic driver for news for a while now. We see actually those sites getting more traffic at this point. I feel like it's like 2x the traffic they get from Google Discover than they do from search. And so I was like, we haven't seen like a big hit to news sites because it kind of, it happened years ago to some level. And so that's a bit of a, that's inconsistent with what we see. But you know, everybody's got a slightly different data set.
- [11:47]
A
So there's also, I don't think they've put AI overviews in the news tab. Yet to the extent that people use Google News, that's a good point now. Yeah. So. So what was the. What's the headline though? What's what percent down across. Across categories.
- [12:03]
C
Ranging from 0 to 30 and more in the 10s and 20s? I would say so, like, I think so. I think our data is pretty consistent with a lot of other data we've seen out there. From that perspective, at what percentage do.
- [12:16]
B
Publishers hit the panic button?
- [12:19]
C
Well, I think they already do because it is also like, my numbers are averages, and that means that some people have been fine because the other big thing is bigger sites generally do better than smaller sites. And we've got a pretty big range from big to small. And I personally have had multiple conversations with content creators of smaller sites who are like, I am closing up shop because I have lost too much traffic. And so those people didn't take a 20% hit, they took an 80% hit. And the game is. The game is basically over at that point. And so it's kind of already happening where people are saying, like, I'm giving up here. But on the reverse side, like, a big thing we talk about with people is like, the number one defense against all this is having a really strong brand. If users know who you are and want to come to you, they will find you and they will find you in different channels in different ways. And like, like, that's what you got to do. And so, like, I think that's got to be the biggest focus for people.
- [13:13]
A
Right. And when you say the big sites are getting less impacted than small sites, is it brand or is it also, you know, email newsletters and multimedia content and stuff like that?
- [13:25]
C
It's both, for sure. You know, to some level. Also, the bigger you are typically means you have a better brand, which typically means you're more diversified. I mean, if you're a small site and you were 90% SEO, like, that's a whole lot of eggs in one basket, and that is likely not going to work out well for you.
- [13:40]
A
Yeah. Yeah. Okay, so what. What can we learn? So especially from the consumer survey, what. How are you taking that survey results in action? What are you telling your clients?
- [13:50]
C
I mean, a big thing for us is, you know, we've taken a stance for the sites we work with that we only want, well, we say human generated content, but really what that means when we, when we boil it down is we want content that has significant human involvement. And so we're trying to ourselves, like, draw the line in terms of, like, if somebody spins up a site and produces 5,000 articles in half a day. We don't want that. If somebody figures out a way to use ChatGPT or whatever tool at scale to help them go from producing five pieces of content a week to 30, but they're editing and checking all of them before they go live. That sounds great to me. And so how do you. So a big thing for us is how do we enable people to do those sort of things? And we're actually, we're building some tools right now to help people with that because we think that's a big thing. I'm like, to compete in the future you have to use AI in a deep way, but you've got to do it in a way that, that is good and ultimately people can trust.
- [14:39]
A
So like, let's say I just want to throw an example to think, see if I'm thinking about it the way you're thinking about it. Let's say I'm a recipe creator and I have a recipe, I have a recipe for, you know, whatever. I, you know, German chocolate cake. And then I use AI to create the gluten free version and the low fat version and the, you know, white chocolate version, etc. Is that a good use of AI? Is that kosher?
- [15:02]
C
I think again, if you are reading those and checking them and validating, I mean, the case of food, I would prefer if you cooked it to make sure that literally worked. But if you're, if you're a professional baker and you've got a pretty good sense of what works, like that's probably okay. Yeah. What the human involvement.
- [15:19]
A
Yeah. Let the AI raise your children. But chocolate cake? Absolutely not. We got to check it.
- [15:23]
C
Yeah. All right.
- [15:26]
A
I want to transition to another subject because while we have you here and you're, you're such a deep thinker on this stuff, what adcp? So adcp, we're now like three or four weeks after they did their big demo and announcement. And one of the big use cases that some of the proponents are pushing is this idea of buying media across many small sites, automating the IO process. And that's right up the alley of raptive as you represent many, many small sites. So what is your current thinking? How involved were you in ADCP or are involved and what is your thinking about that use case?
- [16:02]
C
Yeah, so, you know, we're one of the, I don't think we're technically a founding member, but we were like in the launch and we've been kind of involved, you know, from before it was was announced. We've got some people kind of working a little bit on like the spec and trying to help figure those bits and pieces out. And we are definitely talking to companies now about like running tests to actually get dollars to transact and we've plugged them into our GAM and all that stuff. So we're pretty involved. It's also early that like, it's hard to really know how this is going to shake out. Like, I'm not concerned about the outcome here. I think that, you know, small sites are always going to be challenged from an advertising perspective in terms of making things work and understanding the right ways to price things and the right ways to, to, to ultimately drive the most value. And so I think that, you know, for what we do, I think that adcp, you know, if ADC can be a part of like removing a bunch of nonsense that's in the programmatic ecosystem and creating a cleaner, better ecosystem, like that sounds great. And I think we will still have a really important role to play in that.
- [17:04]
A
Since you're in this business pre AI, how important is the workflow or the use case where an advertiser wants to cherry pick some of your thousands of publishers? Is that a use case that people want?
- [17:17]
C
I can say from the direct side and the PMP side, most people, a lot of people want some level of site targeting. Most of them are pretty happy for us to do it for them. And we say like okay, you want this contextual targeting, this you know, demographic targeting based on domain or whatever, whatever it is. And I'm like, people ask for that but they're happy with us doing it. We don't see that many people saying like we want these 13 sites and not these 24 or whatever like that that happens, but it's a lot less common.
- [17:46]
A
Right. So yeah, but the use case of like, hey, you have 9,000 sites. I want to traffic, I want to buy an IO against certain like data segment across those sites or performance across those sites. That's a use case that you currently entertain totally.
- [18:00]
C
And that happens a lot for sure.
- [18:02]
A
Right, right. So, so I think at cp, one of the ideas is that you should be able to do that without an intermediary, which I'm, I'm not entirely bought into and I assume you're not since that would be your business.
- [18:14]
C
I mean, I'm confident that if that can make things so much more efficient and so much better that we will find a great business model that lives within that it might be slightly different than what we do today. So I'm not, I'm not concerned and I want to figure out how to make it better because if it spends more money on the web, like that's ultimately going to be good for us in some way.
- [18:30]
A
We're not replacing Paul Bannister with AI. All right. Okay, great. This is a good conversation. We're going to transition to the news of the week because we have a lot going on. We will be back in a minute. This podcast is brought to you by Playwire. Tired of ad monetization? Platforms that feel like black boxes even when they promise you complete control? Playwire is launching a truly self service ad platform that gives you actual control. Manage the components of your ad tech stack you want or let their AI handle the heavy lifting where you don't. Your choice, real time analytics, transparent reporting and zero hidden fees. Mention Markitecture and get 50% off your first year. This offer is only available through the end of Q1 2026, so act fast. Visit playwire.com ramp-self-service to take control of your revenue. That's playwire.com ramp-self service.
- [19:25]
B
All right everybody, we are back with the refresh. We have a lot to talk about this week. Maybe we'll start with earnings, the final earnings recap of the past three weeks.
- [19:35]
C
We've been doing a bunch of this.
- [19:36]
B
Stuff so maybe we can take a step back afterwards and talk about, you know, what we've learned in the market. So I'll take through it quickly. We could talk about any of these, any specifics afterwards. So this is end of last week as well as this week. So TTD had solid earnings. Revenue up 18% quarter year over year for quarter Vyant. Also solid earnings, revenue up 12%. CTV was up 46% for Vayant. They said audio similar to what we heard last week with Magnite is growing very quickly and they won Molson cores which they've been making a lot of noise about. So some great stuff out of out of Vayant. They were also talking up there autonomous platform that you know, basically they're positioning as a open Internet alternative to PMAX and Advantage plus AppLovin. Another monster quarter up 68% to a $1.4 billion quarter.
- [20:33]
C
Teas was interesting.
- [20:34]
B
Tees was unfortunately revenue was up a bunch. Revenue dropped 42% largely due to the Teas integration from Legacy Outbrain. But the company swung to a big loss which you know, had a real effect on the, on the stock. And this is where one of those stats that you pointed to in the Previous segment already came from teeds. In their call said they've seen a 10 to 15% decline among what they call premium publishers. We'll talk about that in a second. We talk about some open web stuff. Then finally, Pubmatic solid quarter revenue up 5% again. CTV up 50%.
- [21:08]
A
Yeah.
- [21:08]
B
Anything specific you guys want to touch on before we just kind of bring it back to what we learned this quarter?
- [21:13]
A
Look, teeds. I'm worried about teeds. So they have enormous amount of debt service because they took on debt to do the acquisition. And when you have a lot of debt, you do not want a 10 to 15% decline in your top line to suddenly come from an external factor. And I talked to someone earlier this week, I was like, what khat do? And the first thing they said is they have to renegotiate the debt. And when that's your first priority, that's not great. That's not a great position to be in. So like keep an eye on that space.
- [21:43]
B
Yeah, yeah, no, unfortunately. Paul, anything call out to you?
- [21:48]
C
No, I mean it all seems like, I don't want to say standard stuff, but like these are mostly directions we've seen from these companies in the past. The teens thing is a little bit of a surprising point for sure, but it does jive with everything else we're hearing from other angles. So. Yeah, makes sense.
- [22:01]
A
Tease is very exposed to open web right there. There's their TEED is video, but it's a lot of web video. Yeah, yeah. The other thing I'm interested in is this. More Molson Coors win for Vyant. You know, the, the conventional wisdom's always been that there's like three major DSPs, maybe four if you include Yahoo, and then the sort of peloton of sub 1 billion dollar dsps all sort of compete with each other. And the question is, is Vayant breaking out? Is Vian gonna join the top four? They don't have the revenue. Like they're pretty small company compared to the other ones. But if they can win top 100 advertisers onto their platform and take it from trade desk or whoever, I don't know where Molson was migrating from. That's a big deal. That's good for that, good for that company in stock.
- [22:52]
C
Yeah, yeah. Competition is good.
- [22:54]
B
I agree. And if we believe that AI is going to rewrite the script, so to speak, and create an opportunity for new winners, I think you're onto something. Right. Because they might not have the revenue today, but what do they have? They have A bit of a head start in AI at least from a positioning perspective. They've got good assets in terms of both the household ID as well as the legacy Iris TV content id. And they're really trying to champion this idea of making the open web, the open Internet, accountable. Measurable ability to be outcomes based on.
- [23:31]
A
Don't forget smart, aggressive founders, not private equity.
- [23:34]
B
The most aggressive founders haven't competed with them for a long time. Never count out the Vanderhooks. All right, what do we learn this quarter? Almost every company was up. You know, the demand side companies are up more than the supply side companies in general. The platforms, namely the walled gardens, they were up more than anyone. So I think it's the continuation of the trend is number one. Number two, we're seeing the impact of AI. I go back to the stat that you pointed out Ari, when we were talking about meta, where fully 30% of their revenue was being driven by the end to end AI ad products and others have been pointing to that. And then for the open Internet, CTV is driving like all of the growth. Everybody's talking about 50% quarter over quarter, 50% year over year with, with CTV and you know, a little bit of mixed in terms of the, you know, state of display. Open, open web. If you're, you know, kind of using AI, I think it gives you a good, a good nice lane there.
- [24:40]
A
Yeah, 2026 is CTV plus a political year could be, could be gangbusters for a lot of these companies.
- [24:47]
B
It should be. Yeah, that's a great, that's a great point. Let's, let's go back, talk about open web. Let's talk about some of the stuff. I think that Paul, having you here is going to be a good conversation. So there was couple weeks ago, I don't remember if we talked about it when it came out Ari, but Digiday had a good write up on it. I think it was last week or basically Forrester in their 2026 predictions. By the way, to differentiate you need to do your 2026 predictions like two early, before 20.
- [25:19]
A
Yeah, it's like pumpkin spice latte. We're gonna do them in September.
- [25:23]
B
Exactly, exactly. Yeah, they had a number of, of predictions but the big one that you know, I think got everybody talking was they predict that display budgets are going to be cut 30% due to AI. And the, and the reason why is twofold. There's a couple of quotes. We maybe we pull them out, maybe we don't. So number one, it's traffic, right? If traffic is down 20, 30%. You can imagine. Ipso facto spend goes down 20%. But then there's the unaddressability. Unaccountability. Right. Where it's like very difficult. If you've got a choice of spending on a walled garden, which supposedly $in, $3 out with the unclear $in, what do you get out of the open Internet? And you need to make a choice. You wonder where that choice is going to, is going to go in a rational buyer's mind. Paul, how do you think about this stuff? And most importantly, like, you know, how do you counsel your clients, your publishers to position against this?
- [26:15]
C
Yeah, my highest level thought is like, I'm skeptical of the number. It struck me as like somebody saying a big number to get attention, but clearly that's working. Yeah, exactly. Yeah. Mission accom. Um, and so a little bit like, I'm pretty skeptical of the number. That said, I do think that, you know, we all know it. Like, I think I really, you know, and I'm totally biased, but I think that the web has gotten a lot better over the last five years in terms of like removing problems, increasing addressability, increasing measurability, like making things better, but there's still a bunch of junk out there. And so for me, if 30% hit is really the number, I mostly think that's happening in places that are the places that should be cut down anyway. It's like the number of unviewable ads and ads on random sites that it shouldn't be on. And all these problems, like, if we, if we, if that 30% could be targeted to the right place, that can actually be great for the web because it would mean more money is in better places. And that's. And that will lead to more growth in the future. So macro. I don't believe it. And micro, I'm like, maybe part of that's good anyway.
- [27:18]
A
Shouldn't prices be going up? CPM shouldn't, like volumes down, demand's the same. Shouldn't prices be going up on open web?
- [27:25]
C
I guess the question is how much of it is. Like, is it the open Internet and is that money flowing to other places away from the web? So, so yes, probably. You know, our data is, our data is pretty positive this year on cpms. Wouldn't say it was like massively up, but like it's up. So what we're also, I think we're trying to do, and I've heard this from a lot of other big publishers, is we're trying to reduce supply in general. We actually lowered this year Our content ad density like average at the beginning of the year was like, I think like 25% and now it's like 21%. So we're trying to cut ads out in places because we're finding for the first time ever, it used to be that if you lowered ad density by 10%, you would increase advertiser outcomes, but you'd make 8% less money in the process. So like that was a loss. Now we're finding that there's actually a bunch of places where we're able to cut density or cut ad load in general and make the same or more. And so it's like finally some of the incentives from what we can see are beginning to align in a very positive way. And that's a really good outcome too. So we want less supply in general and parts traffic driven and parts kind of ad load driven. Yeah.
- [28:28]
A
So just quick follow up question. How do you calculate ad density? How do you get to 20%?
- [28:32]
C
It's similar pixels. It's. I don't know exactly. It's similar to like the Sincera A2CR methodology. It's a similar.
- [28:39]
B
Exactly how you mentioned it.
- [28:40]
C
Yeah, got it.
- [28:41]
B
Is there a killer app in terms of an ad format that, you know, if you, if you had your choice, this would be the one or maybe one or two of them that would be widely adopted across the open web. They're performant, they're measurable, they, they drive, you know, all types of outcomes.
- [28:59]
C
For Brent, I don't know if there's a single one. That's. The other thing we've done is we're cutting more and more plain display and replacing it more with like high impact formats. And that's like a big thing. So to some level, a little bit, it's a little bit case specific in terms of like the right unit. But like, you know, for us, like we've got a library of 10 to 15 different formats that we, that we can sell. And you know, our team is picking the right unit for the right place. But it's like, it's, it's a, they're just better units that perform better for the advertiser and that again allows us to remove the plain old display, which is definitely not as good.
- [29:32]
B
Yeah, that makes perfect sense.
- [29:33]
C
All right, cool.
- [29:34]
B
Let's move on.
- [29:35]
A
Let's talk AI.
- [29:36]
B
It was a big week with AI.
- [29:39]
A
Again, of course, every week, every week.
- [29:42]
B
And everybody basically announced the same thing. So Amazon had a bunch of announcements with Unbox, but they announced Ads Agent, which is an AI that sets up and optimizes campaigns. Google rolled out Ads Advisor, which is an agentic product that basically does the same thing. And Innovid launched a set of AI tools that are, you know, agentic and you know, help with campaign setup, optimization, creative generation, all this stuff. So it seems like everybody's, you know, pushing very, very hard in this agentic direction.
- [30:14]
A
Pour one out for Gigi. The Gigi's getting started. Yeah. So my quick framework, I've been telling everyone keeps asking me about AI and so my quick framework, which I think I've talked on this pod before, but just to be be obvious about it, is step one, automate your workflow inside one tool. That's what's going on here. Everyone's using AI so the traffickers and the ops people don't have to fill out 50 forms with naming conventions and all that nonsense that has to happen. Step two is integration of complimentary products so your creative goes to your ad server without a human being cutting and pasting things and without errors and with data correction and things like that. Stage three is integration between non affiliated companies. So the, this is where EDCP has a lot of investment. So it's like how do I buy a segment from Live Ramp? I put it in my, my DSP and there may be pricing and there may be negotiation involved or the ultimate thing is how do I buy media buyer to seller where we have adversarial relationship in a sense. So this is all stage one stuff. Very valuable. Glad to see it doesn't change the world. Every single ad tech company should be rolling out announcements like this in the next couple months.
- [31:28]
B
Absolutely. Paul, what's going on? Raptive with relation to this stuff?
- [31:33]
C
A lot of it. The workflow stuff I think is the big win. We've already built a bunch of things that help our direct sales team and PMP sales teams automate just so much of the manual labor and spreadsheets and things like that. We're working on a thing now that can take an IO and insert all the line items directly into the ad server. And just things like that that just like so manual. They're so painful. And the fact that it's 2025 and it's still happening, it's shocking. But like, okay, now's the time we got to fix all these things.
- [31:59]
B
Yeah, that makes sense. So much. I think like awesome. Low hanging fruit here. All right, let's, let's depart. Bit of a grab bag.
- [32:10]
C
Did you guys see this?
- [32:11]
B
Former CEO of a WPP agency that produces Love island is, is Suing the company for $100 million.
- [32:19]
A
Yeah. The Business Insider, Lara O'Reilly reporting on this. My big question is, like, WPB produced Love Island. That's the part that I was like, what is going on? This is a set of words I didn't expect to see in a row.
- [32:30]
B
Yeah, it's a WPP agency. Right. Remember this? You know, there's a bit of a, you know, kind of Frankenstein. It's called Motion Content Group and they specialize in reality tv. So sounds like this gentleman, the former chief executive, name is Richard Foster, was allegedly ousted after he did. Sounds like a whistleblower complaint on kickbacks and rebates.
- [32:57]
A
Yeah, I think the article business, the article goes into some depth about this, that, you know, getting a rebate, a media rebate, is not illegal in the US So basically you buy media, get a couple percent back. What he is alleging is that those kickbacks, for lack of a better word, were not either disclosed or shared with the customers whose media dollars it was. And then he was fired for that. Seems very possible.
- [33:24]
B
I don't know.
- [33:25]
A
I don't know.
- [33:26]
B
But we learned that WP has an agency that produces reality, reality tv.
- [33:31]
A
Another facet of this that I learned that's crazy is that like, you know, all the big agencies have these barter agencies that they don't like to talk about. And so a lot of this rebate stuff happens through the barter agencies where they do a media buy and then they get millions of dollars of unsold media as sort of a bonus. And then they're allowed to allocate that unsold media to either other clients or for trades. They might give it to an airline to get first class seats on the airline. They're unsold and then their executives fly around for free. It's really ridiculously weird. It's like ready for like a Thomas Pinchon novel. How weird it is. But that might be part of this.
- [34:11]
B
And going on forever, by the way, the barter agencies.
- [34:14]
A
Yeah, yeah, the barter. That sounds like the most fun job, like running the barter agency, doesn't it?
- [34:21]
B
The random, the random stuff in your office. All right. Yeah, exactly.
- [34:26]
A
You just. I gotta make a phone call in.
- [34:30]
B
DVDs over here, which.
- [34:34]
A
Yeah, exactly. It sounds awesome.
- [34:37]
B
Yeah. And I'm sure they're super profitable anyway. All right. One more kind of interesting one that I don't think anybody would have predicted was Nativo. So one of the OG native SSPs, you might have called them in a previous iteration, was acquired for over $100 million. So good job Justin. By Life360. Who saw that one coming?
- [35:02]
A
Did not see that coming.
- [35:02]
C
Not on my finger card for this year.
- [35:04]
A
No, it was not.
- [35:06]
B
Guys familiar with life 360.
- [35:08]
A
Yeah, it's family spyware. You like, you make your whole family get spied on by this company. So my son, without talking too much about my family, my son is like a hardcore like hacker and he, I've never seen someone so insulted by the idea of using this app. Like he was sort of like, you know, he was going onto the dark web to complain about his family because of this, this app.
- [35:36]
B
Anyway, so it's a, it's a company that has, I guess we call it, you know, very rich location and behavioral insights and they cannot connect it to advertising with a platform like Nativo that seems purpose built to, to help run an rmn. So maybe this is the beginning of the next, the next big RMN.
- [35:57]
A
Yeah, I think live live360 is sort of positioning itself as like another like super awesome, like a way to reach young people in some sort of of, I hate to say a privacy protected way but some like way because like honestly I mean the amount, the detail of what they track on your family is pretty extreme. So I do have questions about how safe they are. But yeah, being able to take that data and run it across, you know, many apps, many websites seems like an obvious use case here.
- [36:29]
C
It's. I also feel like, you know, I don't know a ton, a ton about the company but like they've done quite well. Like their stock's gone really high up over time. They seem to made a lot of good decisions. So, so they, you know, clearly they've got some master plan here and we'll see, we'll see what happens.
- [36:42]
B
Yeah, yeah, for sure. Hey, are you want to close it out with the big cloud X news?
- [36:48]
A
Yeah. So it came out this morning that Jim Payne, who friend, I call him friend of the Pod, but he's never been on the Pod. He is, he raised $30 million for his new in app monetization solution called CloudX. I think CloudX might be a temporary name. I think they may change their name but 30 million is a pretty big round for a debut. It was from a bunch of folks edition. DST Global, Eniac, it's Javelin, a bunch of others. It's a little hard to get to what it is. Exactly. So background here. Jim Payne is the founder of Mopub, sold to Twitter, did really well on that acquisition and then he was the Founder of Max Max got sold to Applovin and turned into Applovin's kind of secret sauce for understanding the web, like Unified Auction. So Jim is running a third game plan on this mobile SDK thing. And this time the difference, he talks a lot about AI in the announcement, but I think the real difference is that it uses a trusted execution environment, which is a new acronym the privacy people understand it called tee. And maybe Paul has more insight on how it works. But effectively what it does is it allows the auctions to be auditable because they're done in a sort of in its own container. And I think this is a play to give buyers like Meta and other other ad networks more assurance that their data is not being used by the SDK owner, who in many cases is Applevin. So that's my interpretation of the announcement and Jim actually agreed to be on the show. So we're gonna actually have Jim Payne on in two, three weeks and we can ask him about the infamous helicopter to the Preakness story that we, we broke on this pod many years ago and he objected to that is like, that's like a deep cut. Sorry.
- [38:41]
B
We will be able to call him a friend of the pod after that or maybe not if he gets really mad about the story. Paul, what you got?
- [38:49]
C
You know, the, the mobile app space to me is kind of mysterious. And so I agree with the point that Applovin is, you know, the 800 pound gorilla right now. But also people are like, what is going on here? So something that is more transparent and more clear, that could bring a lot of value to people. So it sounds interesting, but super early.
- [39:06]
A
Obviously they also their launch partners were included Meta, Liftoff and Magnite. So those are demand sources. They'll be bidding into this SDK. Pretty good list.
- [39:16]
B
Yeah. You know what is cool about this is some of the most successful entrepreneurs and this may be endemic to ad tech, they keep doing the same thing, right? So this is the third, you know, kind of like mobile app startup that Jim has done. After monster successes. Think about like the moat guys. They built, you know, the best like analytics platform ever. They're back building analytics platform. The universal ads guys. They just keep building ads APIs starting with like starting with Shift or brand networks. And they've done it like two or three times super successfully. There's something to this idea of people just like, you know, bringing back the old Playbooks, which makes me wonder, is there a DSP in someone's future?
- [40:09]
A
There's so many people who want me to build a DSP. Oh, my God.
- [40:13]
B
I know.
- [40:15]
A
I could build a DSP and then exit for like $100 million, like three times in a row if I wanted to. Like, there's so many people want to buy a dsp, but then you have to build it and then buy it, build it, buy it.
- [40:25]
B
But should we stop recording and have.
- [40:27]
C
A different conversation now? I'm ready to go. Let's do it.
- [40:29]
A
Let's do it. Yeah.
- [40:32]
B
Amazing.
- [40:34]
A
As another side note, while people are looking for business ideas, I think doing something like what Jim Payne is doing in CTV would be really interesting. I think the current state of, like, the auction mechanics in CTV are really messy and a lot of duplication and stuff like that. And having like a neutral SDK that serves the ads would be a pretty interesting play, but worth we're thinking about.
- [40:58]
B
Inboxes are open. That's a neat idea.
- [41:00]
A
All right. This is a great conversation. I think on that. We'll, we'll, we'll call it. Paul, thanks so much for being here and sharing your insights about consumers and AI. Can people get the slides or is it. Or is it.
- [41:14]
C
I'm happy to share the ones that I shared at the bit switch thing, so we'll figure out how to get them out there.
- [41:19]
A
Okay. Reach out to Paul on LinkedIn or wherever and he can provide you the slides. It was pretty interesting stuff. So thanks for being here.
- [41:26]
C
Yeah, thanks very much. Appreciate you having me on the Podia.
- [41:29]
B
See you next week, everybody.
- [41:30]
C
Thanks, Paul. Thanks. Thank you for subscribing to marketecture.
- [41:34]
A
New interviews are added every week at marketecture TV and your favorite podcasting app. Thank you for listening to the marketecture podcast. New episodes come out every Friday and an insightful vendor interview is published each Monday. You can subscribe to our library of hundreds of executive interviews at marketecture tv. You can also sign up for free for our weekly newsletter with my original strategic insights on the week's news@news.com.
- [42:08]
C
And.
- [42:09]
A
If you're feeling social, we operate a vibrant Slack community that you can apply to join@adtechgod.com.