← AdExchanger
AdExchanger
Frank Predictions For 2025
00:51:00
Tune into the first episode of the year with guest Andrew Frank, VP and distinguished analyst at research firm Gartner. We cover a lot of ground, from the impact of multiple major Big Tech antitrust trials to practical AI use cases (minus the BS).
Loading summary
Transcript115 lines
- [00:00]
Andrew Frank
Foreign.
- [00:13]
Alison Schiff
Welcome to Ad Exchanger Talks, the podcast devoted to examining the issues and trends in advertising and marketing technology that matter most to you. I'm Alison schiff. Welcome to 2025 and you are listening to Ad Exchanger Talks, our first episode of the new year. Hope everyone had a restful break. And all I can say is that Santa or the Hanukkah Genie or whoever got me a great gift this year, the gift of no major news during the holidays. My guest this week is Andrew Frank, a VP and distinguished analyst at research firm Gardner. It's pretty cool to have the word distinguished in your title. Gotta say. We do a retrospective on the year that was and a look ahead into 2025. All the greatest hits, practical AI use cases, the Google AdTech antitrust trial, the Omnicom IPG merger, how brands can gain trust in a world of lies. Very dramatic. Plus a few predictions, of course. But first, please save the date for CTV Connect, taking place March 12th and 13th in New York City. And hey, don't just save the date, snag yourself a ticket. Ad Exchanger is joining forces again with synopsys Ad Monsters and chief marketer to host a can't miss summit on all the key issues and opportunities in Connected TV. Learn more and register@ctvconnect.com hey, Andrew, welcome to the first episode of Ad Exchanger Talks of 2025.
- [01:51]
Andrew Frank
Great to be here, Alison. Thanks for having me.
- [01:55]
Alison Schiff
So I'm not very creative. I'm going to use my same favorite question that I asked guests that I asked them all of last year and the year before because it's a really fun question and I get great answers to it. So what is something about you that not a lot of other people already know?
- [02:11]
Andrew Frank
Well, one thing that comes to mind is that I spent the first part of my career as a video game designer back in the early days of console games, and that's where I got my start in the media and technology space.
- [02:27]
Alison Schiff
Are you a gamer today?
- [02:29]
Andrew Frank
I am not really a gamer, no. I occasionally dabble, but I, I can't really call myself a gamer. Not enough time in the day.
- [02:37]
Alison Schiff
There really is not enough time in the day. And you're going to see from the questions that I have for you, we're going to jump all over creation. So much happened in 2024 and so many interesting things I think are on deck for creating 2025. But before we do, I also saw on your LinkedIn that you have a BA in music from Wesleyan, which is very cool. I mean, do you Play an instrument or, or instruments. Do you sing?
- [03:01]
Andrew Frank
I do, I don't sing. At least not in public. I, I do play though. I, I play guitar and bass and before co. I, I played around a lot, but I haven't been doing so much lately.
- [03:15]
Alison Schiff
So your journey is from music major to video game designer to the agency world to industry analyst?
- [03:23]
Andrew Frank
Well, yeah, and I wrote a lot of music for video games, so that was the connection there.
- [03:28]
Alison Schiff
Excellent. So this is kind of a little music reference. There's been no rest for the Wicked because Wicked was a big movie in 2024. So no rest for the Wicked this past year. I mean, I don't know if I'm suffering from, from some form of recency bias, but it feels like 2024 was jammed like wall to wall news with these internecine industry struggles. And there were acquisitions, some small, some huge, the usual navel gazing, but really kind of intense navel gazing and then developments that broke out of the insular walls of this industry and into the mainstream news cycle. And we'll talk about one of those in detail. The Google antit. Both of them. There's the search trial and the ad tech version. But was 2024 a particularly crazy year or am I just getting old?
- [04:25]
Andrew Frank
I. I'm not sure either of those is the answer. I think it could just be a natural cognitive bias that, you know, when you're living in times of chaos, the most recent chaos seems like the most intense. But it's been pretty chaotic for a pretty long time, hasn't it?
- [04:40]
Alison Schiff
It's been extremely chaotic. Although I, I feel like in December's of yore I had time to work on my end of year pieces. I would just kind of chill. I didn't get a lot of email. Massive news wasn't breaking like Omnicom merging with ipg, which we'll also talk about. But yeah, I mean so much happened this past year, maybe tons of stuff happens every year. But anyway, there is a lot to noodle on about what the year ahead holds. So yeah, I'm just going to jump around. We're free range in this conversation and I want to start with the Google antitrust trials first, the search trial. Will the fact that Google lost do anything, do you think, to hamper their forays and their investments into generative AI search, which is a big thing for them?
- [05:34]
Andrew Frank
Yeah, I think it might, but I think that the market forces are actually a much more significant factor right now. I think that the fact that people are changing their search habits, they're embracing Conversational interfaces for search and trying new things and new products and offerings. I think that that might have even more of an effect on Google's dominance of the category than the antitrust stuff.
- [06:04]
Alison Schiff
Have you personally been messing around with chatgpt or Perplexity? Because I have used Perplexity a bit. I was reticent. But I did use it to prepare for a panel during Advertising Week. And it's funny, I didn't end up using any of the stuff, but it did help me get some thought starters going. And I was so resistant to it because my brain is better, I thought. But yeah, what about you? Have you been messing around?
- [06:32]
Andrew Frank
Yeah, I try to play with everything and I certainly have played with Perplexity and chatgpt and a number of the others, PI and so forth. I mean, I think that it's very interesting to see how they all have different takes on not only the resources out there, but the meanings of your questions. It's. It sort of brings home how limited the keyword approach to search really is because there's so much that's unsaid in a keyword type prompt.
- [07:06]
Alison Schiff
Although, as exciting as it is, there is some danger there, I guess, in giving people one answer to a question, even if there are links below it. I was reading recently about a teacher. They had been posting on Blue sky about an experience they had in class. A student made a statement that was categorically wrong. Something about the Greek language, like the Greek language is a combination of seven languages. And the teacher said, no, that's not right. And the student was recalcitrant and said, no. I searched for it and I found that answer. And it turns out I think he had been using Perplexity, although I might be misspeaking. It might have been a different search agent, but. But the teacher could not get through to this kid because for years teachers have been telling their students to look things up for themselves. And he did. And to him, that is search. He searched for it and there was the one answer. And it's not like he has anything to compare that to. So that became his accepted truth. And then the authority figure standing in front of him in the form of a teacher could not get through to him.
- [08:14]
Andrew Frank
That's a very sobering example. I do think that we're going to need to think about how we educate people differently about the nature of truth and to be more skeptical of the truth that seems to appear on their screens.
- [08:29]
Alison Schiff
So I was searching your name using Google, actually, because I wanted to ask you some good questions during this chat and I came across this presentation that you gave at Gartner's symposium, I think it was in Denver last year. And the title slide was how to Gain Trust in a World of Lies. So I, I didn't get to attend, unfortunately. But that is one spicy title slide. So, yeah, I mean, what specific lies are we talking about? And I mean, how can brands just to bring it back to, to marketing? Like, how can brands gain trust when we live in a world that's chock full of obfuscation and straight up like mendacity, but also in a world where truth is turning into truthiness and it's hard to find your facts?
- [09:25]
Andrew Frank
Absolutely. And that's been a very important topic for us over the past year or so. We're definitely trying to develop a comprehensive answer to that question because obviously you can't just wave a magic wand and create trust. But we do think that there are operational things that organizations ought to be doing. We call it trust ops that can help basically institutionalize the things that you need to do in order to make sure that the content that you both use in your organization and produce is, is truthful and verifiable and, you know, has provenance and so forth, and that you're able to detect and engage with false content as it becomes apparent that it's again, corrupting your market or your industry or your, of course, brand itself. So it's, it's a very deep topic and it's one that I think we're going to see a lot more of over the next year or so.
- [10:30]
Alison Schiff
I mean, there can be no trust without truth. I can't be the first person to have said that. It's like a nice little catchphrase.
- [10:37]
Andrew Frank
I can't say it often enough.
- [10:40]
Alison Schiff
So let's go back to the antitrust trials. I want to nerd out on the. The ad tech one. Putting aside that it was surreal for me because I was there for the first week to in Virginia to hear witnesses under oath in federal court talking about header bidding and digging deeply into ad auction dynamics. It was just such a strange experience. What were your impressions reading about it and hearing about some of the revelations and the dynamics and. Yeah. What were your main takeaways from. From what came out of that? From the testimony?
- [11:15]
Andrew Frank
I mean, I probably had some of the same feelings you did about the sort of collision of worlds. Right. The, the tech world suddenly being aired in court in the legal world. I mean, one of the impressions I had was, you know, what took you so long? It's not like this problem arised over arose overnight. You know, Google has been playing the, the market maker and the, the seller of media for decades. And you know, you, I'm surprised that these issues took so long, perhaps because they were so buried in technical jargon and, and complicated systems, but it still seems like it, it has been a long time coming.
- [12:03]
Alison Schiff
Well, do you think Google will lose the ad tech trial? It lost two counts for the search trial. I think Google's going to lose.
- [12:11]
Andrew Frank
I'm not going to make a prediction on the outcome. I think that if you look at the history of antitrust, there's losing and there's losing. Right when Microsoft lost, they managed to appeal and get the sanctions softened and so forth. So I expect that the resolution isn't going to be as simple as losing, win or lose.
- [12:35]
Alison Schiff
No, for sure. Even though I would love a yes or a no answer, I respect that. Are your clients asking you about the cases and the ramifications of these cases? Like how does it come up for them?
- [12:49]
Andrew Frank
It does come up. Marketers all over are trying to game out the future and it's so complicated. And this is just another variable, I suppose, that you have to think about as you're trying to play out the scenarios for how you might need to shift your media strategy, your martech budgets, even your strategy for where you're going to source various ad tech components or even gen AI components since they're all part of the same complicated big tech market. Right.
- [13:26]
Alison Schiff
What though? Like what can brands or what should brands do to prepare themselves for the aftershocks of these trials? Like right, right now I know that the ad tech trial isn't decided as of when we're, we're speaking now. But beyond speculating and pontificating on LinkedIn, which everyone loves to do, what what can companies actually do now?
- [13:52]
Andrew Frank
Well, they can look at more experimental capabilities to test out different platforms, different strategies, so that they have a sense of how they might pivot if they find that they need to diversify their approach to media. And certainly there's been a lot of consolidation of media spending over the years. And so I think that the most obvious thing is to sort of back off on that and see if there aren't other channels and technologies that might, might be able to produce results as good or better than just putting all your budget into one or two platforms.
- [14:37]
Alison Schiff
In the spirit of trying new things. What's cool to you? What kind of channels or what kind of tools would you advise or suggest that a client messes around with or Is it more like this is the advice, you guys should go and try new things and then you kind of leave them to it?
- [14:53]
Andrew Frank
Well, we try to be a little more helpful than that. I think that one of the big trends that you're seeing as a challenge to the conventional approach to media allocation is the rise of data collaboration platforms, particularly with retail media and also with some big publishers who are experimenting with monetizing or utilizing their data much more in their media operations. So I think that's one area that's very fruitful for testing. You know, can you collaborate in a privacy friendly way with publishers and retail channels so that you can recover some of that lost signal that helps with targeting and measurement, which has been of course, the big story since the partial demise of cookies.
- [15:47]
Alison Schiff
Ah yes, cookies. I remember when these podcasts were like 90% cookies, 10% other stuff, and now it's totally flipped. But data collaboration and clean rooms. I know the FTC is about to change, I think quite dramatically with the new administration. But just recently the FTC put out a post saying that they don't think data clean rooms are the panacea or the perfect privacy safe solution that they have purported to be. What's your take on how data collaboration will play out in 2025, considering that the FTC has started to poke around?
- [16:30]
Andrew Frank
Well, like almost any technology, I think data clean rooms can be used in various ways along a spectrum of, you know, safety and compliance with the spirit of the law. I think there's plenty of examples of data clean rooms being used in questionable ways, but I think that the technology is sound and I think that as, as the FTC maybe puts a little bit of pressure at least on the transparency of this systems. I think organizations are going to have to figure out how to do data collaboration in a way that respects not just the letter of privacy.
- [17:11]
Alison Schiff
Yeah, the word respect, it doesn't come up as much as it should because of the fact that it's not really a word that's business related. It's so important, you know, asking yourself, what do you want it? What would you as a consumer want to see? Or what would your aunt or your mom or your sister or your brother or your friend think if you described XYZ practice to them, like, would they be shocked by it or a little bit put off? I don't think people do that enough.
- [17:43]
Andrew Frank
I think you're right and I think that's a really important point. And maybe it's another point that we can raise in terms of where we think things might be going in 2025. Trust and respect are coming up a lot more lately, especially with AI governance and trust. And also I think you, you were talking recently with Jared Belsky about the principle based buying, which is another area of trust, you know, and I think it's another area that's kind of confusing for marketers because they want to trust their agency and believe that their agency respects their goals and is really working for them. But the lack of transparency in a lot of these systems is I think, really where you start to undermine the trust factor and question whether there really is a respect that you are expecting from your, from your agents or other partners.
- [18:42]
Alison Schiff
I'm glad you brought up principle based media buying, because I was going to and Jared Belsky, former CEO of 360i now he's the CEO of an indie agency called Acadia and he's been very vocally anti principle based media buying. He referred to it on the podcast that I did with him, this very podcast. He referred to it as stealing, which is like an extreme word to use, but it's certainly arbitrage. Do you think that clients primarily just want to know that it's happening and what the fees are or is just to have more transparency basically? Or is principle based media buying on its face a bad thing?
- [19:26]
Andrew Frank
I don't think it's on its face a bad thing, but I do think that the transparency issue is crucial. I think that if, if an agency or buyer wants to take inventory risk and put their money on the line, they're performing a valuable service for their client. And as long as that's transparent and the client understands what's going on and the deals aren't misaligning the goals of the agency with the goals of the client, I don't see any reason why agencies shouldn't make a profit on media if that they're taking a risk on. On the other hand, I think a lot of principle based buying doesn't work that way and it is simply a opaque ruse to raise more fees.
- [20:17]
Alison Schiff
I even think, and I've said this before on this podcast, but the term itself principle based media buying obfuscates. I mean, just call it arbitrage or call it something closer to what it is. Principle based media buying just feels a little bit euphemistic to me.
- [20:33]
Andrew Frank
Oh, totally. I think it's a complete smokescreen.
- [20:36]
Alison Schiff
So let's hop on over to the Omnicom IPG merger, just a little $13 billion mega merger that closed out 2024. And there have been a lot of hot takes on this, of course, go onto LinkedIn and you can't avoid them, the Omnicom folks, they're positioning it as good for market competition. And then you have publicist saying basically bring it is an opportunity for us. Of course they would say that. But I mean, fair enough. Arthur Sedun, I don't know if you saw it, he posted this very charming, very French video at the end of December making that point that Omnicom and IPG are just going to be crazy busy looking internally and focusing on integration so publicists can focus on new business and looking outward because they've already done their transformation. It's a, it's a good little video, about three minutes and it's, I think it's worth watching. But then again, there's the perspective that this is all about principle based media buying and then creating more margin opportunities. And bigger isn't necessarily better, it's just a conglomeration. Big isn't efficient, it's lumbering. Big isn't nimble. Not to mention that this is just, you know, inevitable. A lot of people are going to lose their jobs. What's your take on the merger? Good? Bad, Ugly?
- [22:00]
Andrew Frank
Well, I don't know if I have an opinion on its, on its altruistic value, but I do think that it will free a lot of talent. And I think that that that will affect both the tendency of a lot of clients to look at insourcing some of these operations by making more talent available as at least some of the talent goes native, as we used to say. And I think it'll also create more opportunities for independent agencies and maybe smaller boutiques to scale up their talent. And I think that that's probably a good thing.
- [22:44]
Alison Schiff
One more question before we take a quick break. Do you think that merger is all about principle based media buying? That's a hot take that you see plastered all over LinkedIn.
- [22:55]
Andrew Frank
I don't know. I think that there are other reasons why a merger like that might seem attractive to both of those companies. They've kind of been trailing a bit in terms of the, at least the perception of their adoption of new technologies and techniques. And I think right now the, you know, compared to WPP or Publicis or even, you know, Dentsu with Merkel, I think they've been getting a lot less attention for their innovation. And I think this gives them a chance to shake things up a bit. And I think that's not just about principle based buying.
- [23:34]
Alison Schiff
Sure. Anything as big as this can't be about any one thing. All Right. We're going to take a quick break and when we're back, we're going to talk about AI, of course. So stick with us. All right, we are back. And let's talk about AI Gartner's latest hype cycle for emerging technologies. You guys released that over the summer and you plotted generative AI as beginning its descent into the trough of disillusionment, which sounds about right to me. But lay some predictions on me like what should we expect this year when it comes to the practical application of generative AI to marketing, taking into account that people are getting a little skeptical and they're starting to take off their rose colored glasses a little bit.
- [24:36]
Andrew Frank
Well, that's definitely so. And I think that a lot of organizations are starting to confront some of the hard limitations and perhaps recognize some of the overhyped capabilities of some of these technologies. Technologies. So I think you are going to see a lot of backtracking. I don't know if we're looking at quite at the, like the Internet winter that we experienced at the early part of the century, but I think that you are going to see a lot of companies struggling to make these investments work.
- [25:09]
Alison Schiff
I had an experience when I was playing around with Perplexity that really threw me off. I. I asked Perplexity to write an article in the style of Alison Schiff from Ad Exchanger, which was just for fun. I was curious what it would come up with. First of all, I hope I don't write like that because it was kind of going for funny and goofy. I guess I am goofy. But it was all really cheesy. There were dad jokes in there. But what was really alarming was that it made up a quote. It completely fabricated a quote, something like privacy is the iceberg that hit the Titanic. And it attributed that quote not to a real person but to an anonymous DSP source. But I found that really strange and weird and a little bit scary and made me think, yes, this kind of technology, we need to put it in the trough for now while we work out kinks like that.
- [26:05]
Andrew Frank
I think that's true. I think there are a lot of these problems that have been raised and people have been sort of wringing their hands about them. But nobody has really come up with a good way to combat that stuff, particularly when you're talking about emulating people or making up false quotes or even deliberately training language models with misinformation for whatever reason. Right.
- [26:33]
Alison Schiff
It was for sure. It was though, a very interesting and bizarre experience to ask it to write something in my style because I'm relatively obscure. I mean, I've written thousands of articles, so there's a big corpus for it to pull on. But it's not like asking perplexity to write a paragraph in the style of Cormac McCarthy. I mean, I'm a random person and yet it was able to do it in, I don't know, seven seconds.
- [27:03]
Andrew Frank
Right. I had a similar experience asking for my bio from, I don't remember which AI system it was, but it came up with a completely fabricated bio for me that, you know, lavished me with all kinds of advanced degrees that I'd never earned. But I found it quite flattering.
- [27:22]
Alison Schiff
Yeah, at least it flattered you. At least it didn't demote you or say anything nasty. What's the BSE est thing you've heard someone say about AI and how it's used for marketing in 2024?
- [27:37]
Andrew Frank
Well, I did have one client report to me that a vendor had told them that they could use their large language model for everything. They would only need one large language model and they could basically write any application they could think of on top of it just by describing what they wanted the application to do. And I thought that was a particularly BSE thing to say.
- [28:01]
Alison Schiff
Well, because that just bleeds right into the AI is magic narrative, which is not the case. It's a technology.
- [28:08]
Andrew Frank
Exactly. And it's a technology that can go off the rails in very unexpected ways.
- [28:15]
Alison Schiff
Well, what's the. Not to be so negative, let's be positive. What's the coolest, most interesting or maybe useful application of AI to marketing that you've come across recently? Maybe for media buying or like improving planning or something that was good for optimization?
- [28:33]
Andrew Frank
Well, I'm working on some research now about how organizations can fine tune models with their brand, their literal brand assets, to encourage them to produce actual branded content. You know, Genai doesn't tend to render logos or product images very accurately, but overcoming that so that you can really produce content that's commercial ready. And by the way, copyright safe is I think, a big breakthrough and will really empower a lot of organizations to finally approach that ever present goal of personalized content at scale.
- [29:21]
Alison Schiff
Right, I read about that perspective of yours. It was this Martech piece from last summer. The headline is it's time to teach AI about your brand. And yeah, basically marketers should develop custom AI models for their brand and then use it for their own branding or to brand themselves. So not, not to maybe build their own model, but to tweak an Existing model with brand specific training data so that it gets you. But unpack that for me a little bit more. I mean, that, that does seem, it seems challenging, but maybe it's not so hard.
- [29:59]
Andrew Frank
Well, like everything, it's, it. There's a range of difficulties. There are simple things that, you know, hobbyists can do to fine tune models, and then there are complicated ones. I think that what it does, though, from a marketing perspective is it encourages you to think about your brand vocabulary and what are the significant assets and themes. And, you know, the AI is pretty good at imitating the style of a brand, but it doesn't really get the substance. So if you can use AI to do all the rendering of your ideas and variants and so forth, it focuses you much more on the substance. How does our advertising and marketing relate to our values, our goals, our strategy, that sort of thing?
- [30:50]
Alison Schiff
Well, style over substance reminds me of the Coca Cola campaign that was created with AI over the holidays. It was, it struck me as kind of weird and yucky, but I'm sure that's just because I'm not ready for it and I'm a Luddite and whatever. But it did strike me as strange to be talking about the real thing. Coca Cola is the real thing. And using AI to create this campaign that had become so iconic and it really just felt like aliens had maybe infiltrated our television screens and done research on us and then beamed back what they thought we would want to see. Right. Like, it just felt off. It felt like some kind of photocopy of a photocopy.
- [31:38]
Andrew Frank
I understand, and I can't let the reference Alison referring to herself as a Luddite seems a little out of the ordinary. But I think that the Coca Cola ad is a good example of, you know, the risks that you take when you try to use these tools to convey some sort of authenticity. And it really has a boomerang effect, doesn't it?
- [32:02]
Alison Schiff
Well, yeah. And I also, I don't want to be someone who rails against progress or yells at the sky. And brands do have to experiment and take risks or nothing interesting will happen. But it just, it felt very off to me. And I think the best way for a brand to put that kind of stuff out there is to say, we know this is kind of off, but we're playing around. What do you think? As opposed to here's our holiday campaign.
- [32:31]
Andrew Frank
Like everything, it depends on how you're using it. I don't think that we're talking about, you know, lights out, creative production of a campaign. I think you still need humans in the loop to. Yeah. To make sure that that authenticity, that voice is coming through and that it's on brand, not just stylistically, but again, substantively. As.
- [32:56]
Alison Schiff
As we say, when do you think the accustomed to a brand AI model thing will become mainstream? Like 2025 or maybe later than that?
- [33:08]
Andrew Frank
I think it'll happen at different speeds in different industries. I mean, I think you're already seeing, you know, fashion and apparel kind of moving out in front of a lot of these technologies and embracing AI in their production cycles. I think that a lot of other agencies, especially regulated ones, are going to take a lot longer. But I think you're talking about a transition that will happen over the next two years.
- [33:33]
Alison Schiff
Okay. I mean, that's pretty soon. That's really fast, actually, considering it wasn't even something people were talking about seriously two years ago.
- [33:43]
Andrew Frank
Well, I'm not sure. I mean, we've been using AI in content production for some time in some sense. Right. It's getting more powerful and, you know, the whole way of prompting it with natural language is changing. But I would say that we've had AI tools in creative suites for longer than that.
- [34:03]
Alison Schiff
Fair enough. What about like AI governance and trust? Because AI is developing really rapidly, even though, as you were just saying, it's not brand new on the scene. But I do think it's outpacing our ability to regulate it. So I'm a marketer and I want to make sure I'm being responsible with how I approach AI. But I also don't want to cut myself off from opportunities by being too cautious. So what do I do? Like, how do I approach AI governance?
- [34:35]
Andrew Frank
Right. So we've been working with a lot of folks on building their AI council and getting that. Right. I think that whether you call it a council or a committee or a subcommittee, I think you need to have a new charter for a group of people from different areas of the organization coming together to really hash out some of these policies that are so new and complicated. And. And I think you need, you know, marcom obviously needs to be in there with IT and legal and hr. And I think it's a new opportunity for company leaders to collaborate and to start to put policies down that can guide the practices and adoption of these tools and techniques in a responsible way.
- [35:25]
Alison Schiff
Just like privacy. I think everyone needs to be in the room for these kinds of conversations. Like, no one should just own it. Everybody in an organization should be able to have their say. I mean, it doesn't all necessarily make it into the policy. But you do need multiple perspectives to make sure that you're not leaving out something really important.
- [35:46]
Andrew Frank
Yeah. And I think the other thing is that the people who are trying to contribute need to also learn how to see things from other perspectives. Right. So I think marketing and legal have been doing this dance for quite a long time, but I think that marketing and legal are probably a little bit behind the IT department in understanding some of the implications of AI, particularly when it comes to some of these unanticipated unexpected bad effects.
- [36:20]
Alison Schiff
So I don't know how to segue away from AI, which is scary but exciting to third party cookies. Funko I do want to ask whether your brand, clients, publishers, agencies, is anyone still asking you about third party cookie deprecation? I mean, I know Google's not officially deprecating third party cookies in Chrome anymore, but it doesn't mean that brands and publishers and everybody don't need a strategy for signal loss. But I feel like it's so much on the back burner.
- [36:55]
Andrew Frank
Well, it's a long fade out, I guess. We still do get inquiries about that, but they're certainly a lot less frequent than they were two years ago.
- [37:04]
Alison Schiff
What are people asking you?
- [37:05]
Andrew Frank
Oh, how to deal with signal loss. Whether X and X workaround is is kosher, that sort of thing. Whether something that is being proposed by a third party vendor to match data. Whether as we said before, in a data clean room or in some kind of third party system. Typically it'll be a marketer who's being told by the legal department that they can't do something that they think that they ought to be able to do. And so they want to beef up their argument about why it's either okay or understand why it's not okay.
- [37:41]
Alison Schiff
Is that like going to the other parent and asking for permission if the if mom says no, ask dad or if dad says no, ask mom or something.
- [37:49]
Andrew Frank
It's like going to ChatGPT and asking and how to argue with your mother.
- [37:54]
Alison Schiff
So I want to do a little lightning round, kind of like Andrew's Hot Takes. That's what we're going to call this segment. But no one ever takes my lightning rounds seriously, which is why I don't do them that often because I say okay, I want five to ten second answers for real and then I get full length answers. But you've been very concise and you've said a lot in a short amount of time, so I'm going to trust you with a lightning round. I'm just going to ask You a series of quick questions. So five to ten second answers. Is curation just ad networks reborn. Yes, you got it. Thank you. All right, is Eric Suefert right and everything is an ad network or waiting to become one?
- [38:42]
Andrew Frank
No, everything is not an ad network.
- [38:44]
Alison Schiff
The ana, Tag and Fiducia, they just released their second programmatic transparency benchmark at the end of last year year. And there were some encouraging findings like less spend going to mfa apparently and more of every ad dollar going toward working viewable media. Good news. But will the programmatic supply chain ever actually be fully transparent?
- [39:09]
Andrew Frank
No.
- [39:10]
Alison Schiff
Ever, Ever, ever.
- [39:12]
Andrew Frank
Well, ever is a long time. You know, I think apps advertising may become obsolete before it becomes transparent.
- [39:20]
Alison Schiff
You're really good at sound bites. Okay. Will 2025 be the year when ad buyers finally get mad as hell and start pushing back on take rates and tech fees?
- [39:32]
Andrew Frank
I doubt it. I mean, they've already been mad as hell and pushing back, but substantively they don't have a lot of other places to go.
- [39:44]
Alison Schiff
Right. It's hard to complain when you just keep doing the same thing over and over again. Nobody believes you. Will regulators let the Omnicom IPG merger go through?
- [39:56]
Andrew Frank
I don't know.
- [39:58]
Alison Schiff
Is the open web operating on borrowed time?
- [40:02]
Andrew Frank
I don't think so. I think the open web has staying power. It survived a lot of challenges and I recall the quote that the open web sees censorship, is damaged, and finds ways to route around it. I think it's a powerful construct and I think it has a future.
- [40:22]
Alison Schiff
Hear, hear. Will Google actually be forced to spin off Chrome?
- [40:28]
Andrew Frank
I don't know, but I don't think it matters that much. I think that whoever buys Chrome will find ways to work with Google or its competitors to achieve the same kinds of economies of scale that Chrome has already achieved.
- [40:46]
Alison Schiff
Will we get a federal privacy law at some point in the next four years? No, I agree with you on that. And then, last question of the lightning round. Back in 2019, you, you gave a presentation at Pro Guyo New York, and little did we know it would be the last in person Pro Guyo for a couple of years. A little tear down the cheek emoji. But you said during your presentation that marketers felt betrayed because big data had basically let us down. Like the whole promise of yeah, yeah, right time, right place, right person. Like that was just fallacy. And all that this approach to data really did was increase the regulatory burden on marketers and reinforce a data oligopoly and a corrupt media markets with bad data and traffic, which is quite A condemnation.
- [41:37]
Andrew Frank
Oh, wow, I said all that, huh?
- [41:39]
Alison Schiff
Yes. Do you still feel that way?
- [41:42]
Andrew Frank
I don't see any reason to change my view on that.
- [41:46]
Alison Schiff
Yep. Okay, so I'll give you more time with this one because we're not in the lightning round anymore. But which ad tech or. Yeah, you were great. Which ad tech or marketing buzzwords would you like to just kill with fire in 2025? Like you just want to never hear them again. They shouldn't darken the doors of your ears.
- [42:13]
Andrew Frank
Well, one pet peeve I have is this term autonomous agents. And I understand it's a useful term. Autonomy is a useful concept in differentiating. You know why? These are not just fancy macros. But I think autonomy carries a lot of excess baggage when it comes to creating the illusion that these things are, you know, reasoning for themselves and approaching problems with the same level of self determination that we associate with human autonomy.
- [42:52]
Alison Schiff
There's also this trend of naming AIs after people and I think it lends to the same, you know, issue that you're, you're referring to because. Yeah, I mean it's, it's a tool, it's not a person. And if we call it Einstein, then we're going to treat it like it has a personality.
- [43:11]
Andrew Frank
Yeah. I do think the personification of AI is a real double edged sword. On the one hand, it does sort of help people feel comfortable with this new user illusion. But on the other hand, I think it also does suggest capabilities that, that are not really there. I, I was looking at this quote from Eliezer Yudkowski that said the greatest danger of artificial intelligence is that people conclude too early that they understand it. And I think that this is part of the problem with naming it is that you think, oh, well, this is just like a person. It's just sort of an artificial person, but it's not really an artificial person. It's really something quite different and quite subtle in terms of the kinds of errors and mistakes it's going to make.
- [44:02]
Alison Schiff
Artificial person or actually more dangerously like an artificial pet. Right. Like you just control it.
- [44:08]
Andrew Frank
Why? Or a therapist even worse.
- [44:10]
Alison Schiff
Right, yes. Yeah. And there have been some really unfortunate stories where people get too close to an AI agent and take irrevocable action.
- [44:21]
Andrew Frank
Yes.
- [44:22]
Alison Schiff
All right, well, we're, we're nearing the end, so we're at the beginning of 2025. New Year resolutions. I know they're made to be broken, but what is one of your personal New Year resolutions, if you don't mind sharing? And what's a New Year resolution that the ad tech industry should have for 2025.
- [44:46]
Andrew Frank
So personal new Year's resolution. I'm going to try to stay focused more on big picture items. I think that 2024 was a year when everyone got dragged into, as you put it at the beginning, a lot of sort of chaotic news cycle items. And I think that that causes you to lose perspective. So I'm going to try to keep focused on larger ARC stories. And in terms of the ad tech industry, you know, I think two things. One is re double down on the commitment to transparency and to really exposing to both publishers and advertisers how, how their money is being, is being spent. And number two, I think the ad tech industry has to get back to the position where you're working for one side or the other. You're not playing both sides, you're not being a market maker, you're not playing buyers against sellers to maximize profits. I think it has to be more about serving the constituency that you are contracted with.
- [46:08]
Alison Schiff
Yeah, the, the waters have become fairly muddy. Do you think, do you think people in the industry are too negative? Like, do we only focus on the bad stuff? Because this is a thought I've had and it's something we've discussed internally during ad exchanger editorial calls, just as a, I don't know, a philosophical exercise. But it feels like sometimes industry watchers and analysts, and I'm not, not you, but, but take like pleasure when there are revelations about inefficiency or a lack of transparency. Because it's like, see, I told you, like almost gleefulness about being right rather than focusing on how to make things better. I just get that feeling sometimes.
- [46:48]
Andrew Frank
Hmm, sounds like a bad sickness.
- [46:54]
Alison Schiff
Okay, so penultimate question. I've taken to asking people what I call my time machine question, which is like, what specific point in time would you travel back to and what change would you make if you wanted to alter the future of online advertising for the better? But I'm going to ask you a variation on that question because I've asked that question to people so many times and it's time for a new question. So I'm going to ask you my Connecticut Yankee in King Arthur's Court question. If, if you could rip a hole through time and space and live in a slightly different version of reality, what would that reality look like? For example, maybe you want to live in a reality where Lou Montulli, the 23 year old Netscape engineer who invented the third party cookie, never invents the third party cookie.
- [47:48]
Andrew Frank
No, I think that the third party cookie. I mean, that was, as I recall, originally part of the Unix operating system and it had some important benefits before it became this instrument of corruption. So I wouldn't do that. One, one thing that you could change in the past.
- [48:08]
Alison Schiff
It's a stumper.
- [48:09]
Andrew Frank
Yeah, it is a stumper, I think. Well, I'll give you one. There was a time, I remember a couple of decades ago when, when Viacom decided that it was time to sue Google for copyright infringement on YouTube because all of its intellectual property was being played without any compensation or attribution. And that case was never really adjudicated. It was settled out of court. I don't think that the details were ever fully revealed, but I think that was an opportunity for the interests of copyright and intellectual property to establish a precedent for compensating creative people for their contributions. And now I think we're in this world where this has become a huge issue. You know, I'd like to see a world where developers of large language models compensate the contribution of the creative people and the journalists who have contributed to the training sets that they've used to build their, their large language models. And, and so I think that, that, that would be. One thing that I think would be helpful is if we had established a clear policy toward compensating intellectual property owners for their contributions to technology.
- [49:38]
Alison Schiff
I obviously love that answer because I'm fairly certain that a lot of my bylines are feeding a model somewhere. Last question. Make one wild prediction for me, something totally out there, and then if you end up being right, I will buy you dinner.
- [49:54]
Andrew Frank
Ooh. Well, I'm going to build on my previous answer and say that my wild, outlandish prediction, it's going to be an optimistic one, is that within five years, the largest expense for AI model providers will become the royalties that they need to pay to creative talent whose works they've used to train and generate content with their AI models. I'm going to say that we're going to get it right this time and that there will be a large flow of cash to the people who are creating new content because we know that when Gen AI starts eating its own output, it. It tends to get worse and worse very quickly.
- [50:38]
Alison Schiff
Well, I hope to be buying you dinner in five years.
- [50:42]
Andrew Frank
I should have made it sooner.