← Marketecture: Get Smart. Fast.
Marketecture: Get Smart. Fast.
Show Me the Money: Terence Kawaja’s 10 Bold Bets on AI, Ads, and the Future of AdTech
00:22:56
Loading summary
Transcript3 lines
- [00:01]
A
This is a paid ad from our friends at Vibe. If you're running performance marketing at scale and still treating CTV as an awareness only channel, you're leaving money on the table. Vibe Co is the most advanced streaming TV ad platform built specifically for performance teams, giving you the same self service, control, targeting, precision and real time optimization you get from Meta, but on the big screen. Right now Vive is offering a 50% spend match up to $200,000. Head to Vibe Co and mention this podcast to claim the offer.
- [00:40]
B
This podcast is brought to you by audiohook, the leading independent audio dsp. Audio Hook has direct publisher integrations into all major podcast and streaming radio platforms, providing 40% more inventory than what could be accessed in omnichannel DSPs. What's more, audiobook has full transcripts on more than 90% of all podcast inventory Enab enabling advanced contextual targeting and brand suitability. Audio Hook is so confident that in addition to CPM buys, they offer the industry's only pay for performance option where brands can scale audio and podcasting with peace of mind knowing they are only paying for outcomes. Visit audiohook.com to learn more. That's audiohook.com. Hey this is Ari. We have another great recording from our Architecture Live event. This one is with the legendary Terry Khawaja of Luma Partners. Terry does his predictions of the top 10 things to know about Ad Tech and he does it for those of you not watching on YouTube. He does it dressed as bad bunny from the super bowl and that's why you might hear a little Spanish in this presentation. It was a full house brought down the house. Terry is hysterical as always and quite prescient and several of the predictions seem to already be playing out. So I hope you enjoy this conversation with Terry Khuaja from the marketector Live event in early March.
- [02:09]
C
Buenos Dios, marquitector. Ari said go big. It's a big room. All right, so here we go. We're going to walk through 10 predictions. I don't know if they're predictions. First up is. Yeah, so I posted this last week because, you know, people have different opinions and viewpoints as to what is likely to transpire at this interesting intersection of advertising and AI. And I agree to disagree with some of the folks in the room, by the way. Actually, I kind of don't disagree with Andrew. But the point is it's more fun this way and I'm super happy that we have smart, dedicated, passionate people who write about prognosticate about what's going on in our industry, it helps. So first, let's start off with a disclaimer. 10 predictions. Really? I think this space has got so many unknowns. There's three words I always use when I talk about AI and advertising, and those three words are I don't know. I don't know. You don't know. I don't think Sam Altman knows. Okay. Nobody knows how this is going to transpire. Or perhaps if I want to get more in fitting with the meme, no say. Okay. All right, so here we go. Good. I'm glad he's doing the Spanish. Mine's not so good. Prediction number one is we will see a separation of the wheat from the chaff in 2026 as real technologies manifest in financial performance. Whereas, as I like to derogatorily refer to them, posers. Keep talking about AI at conferences. Show me the money. Right. And so, you know, I tweeted, I posted this when we were formulating our AI Lumascape. An impossible task, by the way, to try and parse companies. Yes, of course, AI first. That's fairly straightforward. The hard part are the fast adapters. Okay, fine. If they've done it by acquisition, that's easy. If they've built a product that is a centric sufficient to warrant inclusion on the AI Lumascape, that's hard to parse. We're not technologists. So I basically said, you know, there's going to be a lot of people claiming they have AI and then there's going to be the AI posers. And when I posted about it, it got a lot of attention. So obviously I could probably augment our revenue with some merch if I sold this shirt. But nonetheless, we tried the impossible. This is the latest version of the AI Lumascape. Again, an impossible task to figure out who should be on and who shouldn't. And so we kind of punt it. We basically, you know, this got a tremendous amount of attention. 1.6 million views, likes, downloads, comments and messages. That's great. A lot of interest in this topic. But we believe that this will manifest if you just look at Applovin, for example, from the time they launched Axon 2.0, their AI technology, how their sales inflected and how their margin improved and how their revenue per FTE inflected and spiked. So that's demonstrable financial performance in faster growth, higher margins or greater revenue to FTE ratio. We should. If your AI is that real, it should manifest in your financials. So what are we doing? We're defaulting to including you on the AI Limascape and then we'll check in in a few quarters. And if you haven't sort of had the similar performance as meta as applovin, I know those are extreme examples, but again, it proves the point. Then, you know, maybe, maybe, maybe you shouldn't be on the AI lumascape. So we're calling 2026. Show me the money. Show me the money. Show me the money. All right, so that's one. Let's go number two. Prediction. Numero dos. Numero dos. Despite naysaying about advertising as a modernization strategy for LLMs, it will, we believe, be the most important revenue driver. Necessity is the mother of invention. And I've heard the counter arguments, but they don't work. Already the these four LLMs have raised almost $300 billion, including OpenAI's $110 billion round announced last week at a $840 billion post money valuation. They're going to need a return on that money, folks. And so for that, it ain't going to come from subscriptions, which are up running around 6 or 7% and declining as more and more users use AI. Let's not also forget big tech is investing. These four companies alone investing 700 billion with a B every single year. So that money's got to find a home. Like I said, it's not coming from subscription. So ads, I think are gonna be a reality. As we said, just like the new entrants into media, all roads lead to advertising. And yes, I'm well aware that, you know, Anthropic came out and said no, and perplexity said, no, we're not gonna do ads. Trust me on this, ads are coming. All right. Okay. Number three, despite OpenAI's initial launch of ads based on a CPM, the LLMs will evolve to a performance or affiliate model as it drives consumer commerce transactions. It surprised a lot of people. Oh, right. It's the comparative advantage of intent data. Right, I forgot my point. I talked about Fiji Simo, who is the Sheryl Sandberg of OpenAI, as this sort of having an exciting new future with an AI driven advertising interface. Because again, here's a company with an $840 billion market cap or valuation, most valuable private company in the world. They've got a new consumer interface, a new product, new pipes, new value proposition, new business model and new ecosystem. Oops. Except for the CPM part, that was a little bit of a flaw in the ointment. And I have a idea as to what is driving that and essentially it's to garner a sufficient amount of what I call advertiser density. Because if you think about data, all data is not created equal. Right? Here's our sort of hierarchy of data. And yes, the base element, the impression is of course divided by 1000 is the M in CPM. It's what we based most of our advertising on. And yet what is an impression? It's a proxy for a viewable impression. What's a viewable impression? It's a proxy for the user's attention and so on and so forth. Context is a proxy for audience, audience is a proxy for intent. And intent, ladies and gentlemen, is the golden elixir that drives outcomes. And what is an outcome? It's simply advertising perfected, right? That was the reason why you spent the money. And so yes, OpenAI starts with CPM because that's the fat part of the pyramid. That's where the most advertisers are. And if they get them on the platform, we believe it's inevitable that they move up to intent and leverage that comparative advantage. So think about it. AI answers are like search, intent on steroids. So anybody entering the keyword phrase Cleveland divorce lawyer, Google knows three things about that person. They live in Cleveland, they're getting a divorce and they need a lawyer. And I'm going to submit that there are zero, absolutely zero, casual searchers of the keyword phrase Cleveland divorce lawyer, unless of course you fall into that category. And so every divorce lawyer in Cleveland has bought that keyword sequence to the efficient frontier, to the max, because you're guaranteed to get customers. So now think of that. And by the way, so it's a very, very high intense signal. Now take that to the AI chat world. The user is going to start with an on average 25 word initial prompt and then follow that up with multiple prompts, giving a tremendous amount of intense signal to the LLMs. So search was the most amazing thing in the world, created a quarter trillion dollar industry. We believe that AI advertising is going to do the same, only bigger and faster. All right, so LLM advertising will not be vertically integrated, leaving opportunity for a vibrant last mile ecosystem to emerge. I know, good news to this room, right? Except Google, right, that's a little flaw in the ointment. So here's the thing. It makes no sense for LLMs to fully vertically integrate into ads. They're horizontal platforms where with many downstream verticals. So yeah, they'd have to build out advertising. They also, they're not gonna build out all of these other verticals. As well, their revenue model favors platform economics. Right. An app store gets 30% of everything, so why not let an ecosystem thrive and you just collect a toll along the way? Advertising is extremely fragmented and requires deep vertical expertise. Just look at the Lumascape. It requires distribution, not just intelligence. So great, well done LLM. You've got a lot of signal, but now you've got to find inventory to place that against and then regulatory risk. Why would you want to take on that regulatory risk if you're an LLM? And after all, OpenAI only has 6,400 employees. They couldn't do it if they tried right? Now of course I said except Google. Because in every single one of these, Google is the exception. Gemini also is paired with Google Advertising. So they clearly have, they understand platform economics and they'll take their toll there. But they've got the advertising capability, they've got distribution, they've got reach, they have 191,000 employees. They could stand up a last mile solution in AI advertising. And besides, OpenAI has already partnered with numerous companies as I believe an initial sort of signal that. Yes. Now are they going to build some of their own tech? Absolutely. Or maybe they'll buy God Bless America. But the reality is they're starting off with these partnerships and I believe they need these partnerships going forward. Ok, ok, we're halfway through or almost halfway through. Number five, AI will create sufficient transparency that will jeopardize the principal trading activities and of the agency hold coast. Terry, what are you saying? Media kickbacks and arbitrage drive profitability for the big hold coast. Look, this isn't terribly controversial. You know this, right? And every single one of the major hold coast does this. It's, it's by the way, fine. The client business just isn't that profitable. So you know, they, they do factoring of your receivables. That's why you as an ad tech company don't get your check for 130 days. Seriously, 130 days. That's your money, by the way. And they're making margin on it. And of course they're driving principled activities of their own. I'm not saying it's terrible because largely they're doing it with the clients fully aware that they buy the media and sell them the media at a higher price. However, I believe that AI will bring about price discovery and transparency. That might make that a little bit awkward because I think we're going to miss those dark corners which is where they make their money. If it's not fully Disclosed. And there was an article in AdWeek last week, talked about or two weeks ago talked about how Dentsu and WP are moving away from trade desk's open path over transparency. And I saw that and flipped it around. I said, I actually think that open ads transparency is the problem for the ad holding companies. So I think the two components are right. The directionality is inverted. I believe this has more to do with that than it does with this notion of lack of transparency. Okay, prediction. Numero says. Numero says, oops. AI standards will standardize. I mean guys, we're already an acronym soup here with different parties doing different standards. Is there a lumiscate for this? Yeah, there is. You know, we got, we got all kinds of. And by the way, net, net, I think it's good that we're being intentional on the way in and trying to decide, yes, let's not let this thing run amok, let's figure out some standards. So like I am in support of adcp. I am in. I, I like the idea of being intentional on driving standards before we develop a multibillion dollar industry. So that I commend and I support. But I've also noticed the fragmentation and you know, maybe it's just an early thing and we'll get it all figured out and cleaned up, but I mean it should sort of consolidate, right? Logical. All right. 7. After years of being ignored, AI enabled creative technology will be recognized as the most critical tool to drive effectiveness. Creative is going from manual to automated folks. You know, look, Peter Drucker, famous management consultant, said many years ago, if you can't measure it, you can't manage it. Well, there's a corollary to that, which is you only manage what you can measure. Right? So we tabulate things and then act on our tabulation. And creative has just been. We know if you asked how important is the creative, everyone in their gut would say that's really important. Because an ad with shitty creative doesn't get a lot of traction. We know it in our gut, but we can't measure it. And if we don't measure it, we don't manage it. Which is why, you know, we have seen hundreds of billions of dollars invested in deals and conference panels all focused on data and media and almost nothing. It's crickets on the creative side. And I believe because it has been historically thought of as manual and it will be thought of in an automated way. So I think AI could well be the unlock that really turns things around. And I blame procurement. We need Ladies and gentlemen, a department of effectiveness is what we need at all these companies. And think about it applying in an AI context. We use an iceberg analogy where what you see above the surface is the no brainer sort of reducing costs on production. And it's clear that AI will have a massive, massive impact, all in the name of efficiency. But the bigger part, what's going to really move the needle, is the data that determines whether the creative has led to the effectiveness of the AD all again, driving for effectiveness. By all means, do both, but you can think of the former as tactical, whereas below the water is strategic. Okay, here we go. Number, number. Ocho. Despite significant changes brought about by AI, the Lumascape will adapt and live on. Hey, more good news, right? So as Rashad Debacciola famously said, if agencies are like cockroaches, ad tech is like a virus, it keeps mutating. So, you know, I came up with this slide last fall where I said, you know, again, there's the wheat for the chaff. Point. An admission the Lumacabe is a fragmented ecosystem of largely undifferentiated companies built in the ZIRP era, which eke out their existence on rev shares and kickbacks. Harsh but true. Right. The result is an inefficient supply chain with economic capacity, perverse incentives, shenanigans and fraud. What's needed are fewer players doing more volume and lower take rates with better quality. In other words, rationalization, like happens in every other industry. Our industry refuses to grow up. I won't grow up. It's the Peter Pan of industries. And some believe that this won't come in the way of an evolution. It'll come in the form of a revolution that blows up the existing programmatic infrastructure and starts over. And you know who I'm talking about. Oh yeah, comrade O'. Kelly. You know, making ubiquitous posts on LinkedIn last fall about how it's all over. And of course Scope three is the answer. No, just kidding. But I think directionally he's absolutely correct. I just think, you know, it won't quote, unquote quite be that magnitude. And by the way, you know, he was so prolific in these posts, they were posting like three times a day. One wonders what kind of automation technology he uses. Ah, that makes perfect sense. Perfect sense. Oh, hold that, there's more of that coming. Okay, all right, so let's go. Number nine, the penultimate. With a steep technology curve and much uncertainty, the single most critical skill set in adtech is agility. Like I said, going back to the Beginning. Nobody knows how this is going to go. So what does that mean? It means product strategy, product development, go to market, partnerships, everything. You should be going into the next five years saying, I tell you what, let's just stay on the balls of your feet, folks. You gotta be adaptive, be like this dog, because things are gonna happen and they're gonna happen fast. But certain things that I think you can pretty much count on, right? Scale always wins, right? That's a no brainer. Easy wins. The easy button. Why do people spend on meta and Google is because they make it, it works and it's easy. Man. If we could make the open web work and be easy, we'd do better. Performance wins, clearly, because we're in an age of outcomes and quality. By the way, facilitation wins. This means don't interrupt the user, let them. And I think LLM type advertising will be seen as facilitative, not interruptive. Google wins because they always win. Sorry. And we believe consolidation is inevitable. Of course we would say that, because prediction number 20, number 10 is that the ad industry will be substantially changed by companies that are new to the sector we call these new entrants. And they'll largely be buying their way in. It's fascinating to see these companies, they tend to have one thing in common. They have large pools of consumer data. And some of them, most of them, operate with a core business that's very low margin. So media margins of 40 or 50% look lavish and so that's why they're all entering it. But at the end of the day, the reason why they're doing it is not just to build a new business, but it's to contribute to their core business. And yes, several of them have done it via M and A. And don't worry, there's plenty more M and A ribbons to be handed out, so, you know, we'll be generous with those. Besides, really, an investment banker telling you that there's going to be more MA is his final prediction. I'm shocked, shocked to see Luma pitching M and A as a strategic solution. You're winning, sir. Oh, thank you very much. Okay, well, by the way, every one of my presentations ends with more M and A. So there you go, get used to it. And by the way, we got a bonus. My bonus prediction, ladies and gentlemen, is that Ari will launch some new venture before the next market event. I don't know what it will be. There's just not enough Xanax to keep him still. I know this because he's my neighbor, and he just keeps coming over all the time to borrow the wi Fi, the sugar, whatever. See the next parody video. So, yeah, Bad Banker, by the way, maybe in hindsight, I should have called it AD Banker. I don't know. But anyways, thank you for coming to the halftime show. And remember, halftime shows are what happens after a halftime show is the best part of the content. Right? And coming up, marketecture's got an amazing lineup. They've got a guy who performed amazing feats and basically separated himself from the pack as clearly the number one. And then he hit headwinds. But enough about Jeff Green. Okay, so, so, so, okay. Maybe I do deserve the big beds. Bad Baker. Thank you.