Episode 174 · Season 7 · ·

The AI Surface

Ari Paparo, Founder & CEO of Marketecture Media, on where advertising money ends up when people get more of their answers from AI and visit far fewer sites.

Ari Paparo, Founder & CEO, Marketecture Media. In conversation with Justin Cooke.

Read the interview
Share
Ari Paparo, Founder & CEO at Marketecture Media
Ari Paparo, Marketecture Media

The AI Surface

0:0034:00

Two decades inside the plumbing of digital advertising, from building and selling a demand-side platform to writing the definitive account of Google's ad dominance. Paparo argues that the story of advertising has always been about distribution, and that as consumers get more of their answers from AI, the money will follow whoever sits closest to them.

Paparo has spent his career inside the machinery of the ad market. He ran a demand-side platform, Beeswax, which he grew and sold, he sits on boards as an entrepreneur and investor, and he wrote Yield, a history of how Google built, bought and bullied its way to advertising dominance, bookended by the invention of programmatic advertising around 2005 and Google's federal antitrust trial in Virginia. Today he is Founder and CEO of Marketecture Media, which he started four or five years ago to help buyers skip the first vendor meeting, and he records a weekly podcast and writes a regular newsletter that calls products as he sees them.

In this conversation with the host, Paparo argues that the profits in advertising accrue to whoever touches the consumer, and that the AI surface, whether ChatGPT, Gemini or a future agent, is becoming that surface. He traces a 40 to 80 percent collapse in search referral traffic, explains why aggregators lose and utilities survive, and reframes SEO as AEO rather than a dead discipline. He is clear-eyed on scale: OpenAI's run rate looks closer to 150 million dollars than its billion-dollar goal, and Google's move to AI search matters more in the long run. He warns that chat is an open market today but agents could lock consumers in, that there are no breakout AI-first ad tech companies yet, and that generative creative has become a tidal wave. Whoever touches the consumer keeps the money.

If people get more of their answers from AI and click through to far fewer sites, where does the money physically end up?

“The profits should accrue to whomever touches the consumer”

The story of advertising is always about distribution, who controls distribution to the end consumer. It goes back to newspapers, who held a local monopoly for decades. So the profits should accrue to whoever touches the consumer, whether that is Gemini, ChatGPT, or ultimately an agent. The reason it holds is a very distributed world of millions of advertisers. Where newspapers had 100 advertisers, now there are a million, they are fragmented, and if you are next to the consumer, you are going to get the money.

Is there a version of advertising inside AI assistants that works commercially?

“Something like 80% of purchases from OpenAI ads were new customers”

ChatGPT is showing ads based on what people ask. As a first-order analogy it is search advertising in a different form, but it is richer, there is far more context in a prompt than a search term. I interviewed the CEO of Criteo, an early OpenAI partner, and he said something like 80 percent of purchases from OpenAI ads were new customers. That means it is not bottom funnel, it is mid funnel, and that is a really big difference.

If the market really gets opened up, who benefits, and does chat stay open or do agents lock people in?

“There could be some really significant consumer lock-in as we move to agents.”

Search turned out to be a natural monopoly, where the winner is determined by scale, so there can only be one. The data flowing to Google made its search better every day until no one could catch up. The real question is whether a consumer AI environment is unstoppable, or more like accounting software you can switch between. Chat looks like an open market, you could move between Gemini, ChatGPT, Claude and Perplexity without losing much. But agents know your bank, your email, your preferences, so there could be some really significant consumer lock-in as we move to agents.

From Business of Marketing

Put yourself forward or recommend a guest for us to consider.

Request an interview →

What stayed out of the book, and what was the strongest story you told?

“All three of those companies broke their promises effectively.”

The strongest chapter, a story that had not been told, was 2018, when Apple released Apple Reader, Facebook announced Facebook Reader and Google announced AMP within six months of each other. All three had the same vision of moving the web forward, ad-supported, and publishers thought this could be their saviour, a path out of fragmentation, lack of identity and bad ads. But all three of those companies broke their promises effectively. Apple was hostile to ads, Meta would not build a real partnership and pulled it, Google said the right things but did not do them. That was the lost opportunity. If one had worked, it could have changed the consumer's relationship with content.

If companies hire fewer juniors, how does the next generation learn the business?

“The AI is your intern.”

This is the crux of the AI question, whether it eliminates or creates jobs, sometimes called Jevons paradox, if things get cheaper you use more of them. For young people, they get the negative side first because you do not need interns, the AI is your intern. On the other hand, if you come in AI-pilled and know how to automate your own job, you have a big leverage effect, far more than the boomer in the corner office printing his emails. And it has never been easier to start an ad agency, one person can have more impact than a team of ten a couple of years ago.

Never miss an episode. Follow Business of Marketing.

Read the full transcript

Lightly edited for readability

Speaker 1: Welcome, Ari Paparo from Marketecture. It's great to have you here.

Speaker 2: Yeah, thanks for having me.

Speaker 1: So you've spent your y- your career inside what I would call the plumbing of advertising. You built and sold a company, you sit on boards, and now you're running kind of Marketecture Media, and, uh, really, y-you also wrote Yield about how Google took control of the ad market. Um, and I think this is gonna be a fascinating conversation because it's where the advertising kind of money ends up if people get more of their answers from AI and visit fewer, fewer sites. So let's get going. Um, on that point, um, where the money ends up if people get their answers from AI and click through, um, through to fewer sites. You've spent, you've spent your career inside the machinery of the ad market. If people get more of their answers from AI and click through, through to fewer, far fewer sites, what are you seeing about where the money physically ends up?

Speaker 2: Sure. Um, the story of advertising is always about distribution, who controls the distribution to the end consumer, the person who's enjoying the media, and that goes back to newspapers who had a monopoly on local distribution for 50 years, 100 years. Um, so, uh, the simple answer is that, uh, the AI surface, what the consumers use, so if the consumer's using Google and Gemini or if they're using ChatGPT or if, uh, ultimately they end up using an agent of some kind, the, uh, benefits, the profits should accrue to whomever touches the consumer. Given the reason that's the case, it wouldn't always be the case, but the reason that's really the case is because you have a very distributed world of literally millions of advertisers, uh, which is a wonderful thing that so many advertisers exist. As, uh, back to the days of the newspaper again, you might have 100 advertisers, now you have a million. And so they are fragmented, they're gonna spend money however they can, and ultimately, if you're next to the consumer, you're gonna get it.

Speaker 1: And plenty of publishers have built real businesses on search, on search traffic. What are the ones handling this well doing differently now?

Speaker 2: Right. Well, search traffic has declined, uh, by various estimates, something like 40 to 60 to eight- maybe even 80% over the last two years, depending on your category. Um, and the worst affected are folks who were just aggregating data from various sources and not adding a lot of value. Um, and there's also a dynamic where publishers are playing a cat and mouse game with the AI crawlers and using things like Cloudflare to either block or to potentially try to extract money from those aggregators, right? So it's a real mixed bag. Um, the more unique the data, the more it's a utility for consumers, um, the less effective it is. To give one example, um, if you, in the United States, people play, uh, what's called fantasy football, where you pick NFL players and things like that. That's a media business, having a fantasy football, uh, portal, but it's totally not replaceable by AI. The tra- you're gonna go there, maybe you'll use Claude to, like, check your, check your rankings of things-

Speaker 1: O-optimize your team maybe

Speaker 2: ... Right. Yeah.

Speaker 1: Yeah.

Speaker 2: But you still have loyalty to the media source in that case. Um, whereas something like a, um, a review site where you're writing reviews of different products and you're making money on the affiliate fees-

Speaker 1: Mm

Speaker 2: ... you really have to think about how that's gonna change in terms of traffic, outbound clicks, uh, affiliate fees, a lot, there's a lot of moving parts there.

Speaker 1: And perhaps entering somewhat a new world of, of creativity in what content we create as publishers.

Speaker 2: I think that's true. I think one of the biggest differences, I'd say, between the traditional media world and the tech world is that the media world thinks in terms of pages and content and eyeballs, and the tech world thinks in terms of data and APIs. Um, and in, in a world of an agent, you know, right now agents are the new hot things. There's Muse from Meta, there's Dots from ChatGPT. What they want to do on behalf of the consumer is get data, right? And if you're a publisher, you need to start thinking about the world that way. Um, it... and an example once again would be, you know, reviews, product reviews, right? Is the only way to consume a product review to download an HTML page a- and read it, or what if there was an ability to extract those product reviews, including affiliate links, in a structured form to then be read and interpreted by a Claude or a ChatGPT, and maybe the revenue model survives, and you're still adding a lot of value in that circumstance.

Speaker 1: So is that what you mean when you talk about sort of a, you know, a version of advertising inside AI assistance that works-

Speaker 2: Yeah

Speaker 1: ... from a commercial perspective?

Speaker 2: Yeah. There's a couple things going on right now. One- the most obvious one is that ChatGPT and OpenAI are showing ads based on what people are, uh, using those, uh, those bots for, what they're asking, which is, uh, peop- the f- as a first order, um, analogy, it's search advertising in a different form, but it really is pretty different from search advertising because the formats are different, the inputs are different, they're richer. There's a lot more context in a prompt than there is in a search terms. There's been some stats about how much longer prompts are than search terms. Um, so there's some real differences. And another difference, I, I recently had the pleasure of interviewing the CEO of, uh, Criteo, a- which has been a early partner of OpenAI, and what he said, I, I may get the number exa- not exactly right, but something like 80% of purchases from OpenAI ads were new customers, uh, meaning that that is a, not a bottom funnel, it's a mid funnel. Uh, and that's a really big difference.

Speaker 1: That's huge. Um, so in that context then, who's feeling it first, and, and how long have they got before it really hits the numbers?

Speaker 2: Um, well- Obviously, publishers who are, um, who are dependent on referral traffic are the ones feeling the biggest impact of AI right now, and they're in somewhat of a panic because that revenue's going away, they need to replace it. In, in the short term, uh, we've heard that prices are going up so that the remaining traffic may be more valuable. Um, and that's a dynamic that linear TV saw as consumers cut the cord. The revenue was able to be maintained for some time, but that, that'll end its course at some point. So that's the biggest effect. Um, I think advertisers, especially those that are very direct response, are diving in here, but the scale's not there. Um, you, you're not being able to one-to-one replace search with OpenAI or other advertising yet, um, and it's also a little uncertain how it works. And the other thing we've heard over and over again is that the organic traffic, um, is a big question mark for CMOs. Uh, the question is, "How is my brand ranking? How are we being displayed in these AIs?" There's a bunch of vendors, well-funded tech startups who are trying to tackle this problem, but there's not a unified answer. And we've heard, uh, I think we had Nate Elliott from eMarketer on our podcast, and he said it's the number one issue for CMOs is h- understanding my ranking in AI.

Speaker 1: And you talked about scale, sort of where are we on the, on the volume, the volume kind of charts now and-

Speaker 2: Right

Speaker 1: ... you know, and, and how fast do we think we're gonna kind of reach that inflection point?

Speaker 2: If you're talking about paid advertising in AI, uh, the, uh, we don't really know is the short answer. Uh , the longer answer is there's really one participant, which is OpenAI with ChatGPT, and they have said that they have a goal of a billion dollars in revenue for the year, and it, uh, by the early indications, there was recently some news that came out about the run rate being closer to, like, 150 million. So we're s- we're recording this in October, seems pretty unlikely they're gonna hit a billion. So-

Speaker 1: That's a lot less than I thought.

Speaker 2: Right.

Speaker 1: You know, when you think about, like, you know, the kind of the race to a billion and, you know, Uber kind of p- passed that threshold, what, 18, 24 months ago?

Speaker 2: Yeah. Yeah, they're speeding down the highway while changing the tires, right? There, there's a lot going on there. So I'm not gonna really s- it's not a indictment of their business model. It's just there's a lot going on there, right? So you have-- So that's one data point. The other, the other big data point is obviously Google. So Google has Gemini answers. They, every time they give, um, earnings reports, they talk about how great everything is going. But they also are not necessarily putting ads inside those AI answers. Um, so there's a very murky situation about how Google Search will or will not transition to ove- open, to a AI-based search, and that- that's probably more important in the long run than OpenAI's actions.

Speaker 1: And I have, have to kind of think about sort of, like, that early adopter and how kind of open to ads they are versus-

Speaker 2: Right

Speaker 1: ... uh, there's, there's a kind of sense of, like, the demographic that's kind of just absolutely kind of consuming the hell out of kind of search, right?

Speaker 2: Yeah. Well, yeah, but also keep in mind that those early adopters are the ones who are paying, and if you pay, you don't get the ads, right? So the whole point of OpenAI's ad model is to get to 100% of the eligible consumers participating when only call it 10, 20% are ever gonna pay, right? So it's the same as Netflix's strategy. Netflix had the advantage of saturating the paid market and then entering the free market. OpenAI is trying to do both at the same time, which is a little trickier.

Speaker 1: It's a fascinating time to be alive.

Speaker 2: Yeah, you could probably say that for the last 20 years, right?

Speaker 1: Right. Tell me about it. Um, so look, if you were advising a brand on where to put budget over the next three years in the context of what we just described, what would you tell them to stop doing now?

Speaker 2: What to stop doing? That's a really good question. I think, I think brands are pretty smart about realizing what to stop doing, and they, they, um, tend to prioritize their dollars based on ROAS and, like, number one is always, like, Meta, number two is Google or something along those lines. Uh, it's not necessarily gonna change, right? Um, so you go down the list, and the less and less effective things start getting cut. Um, but what's most important, I think, is that the basics about having a really good website that's really informative and being on top of SEO, and those things have not changed, right? And some people are saying, "Well, SEO's dead." Well, it's not. It's basically just become AEO in a different form. Um, there's quite a bit of evidence, uh, also again, eMarketer, um, uh, sh- put some s- put some slides up during the, our recent conference in Chicago, uh, that showed that the brand websites and the retailer websites were two of the most important sources of AI data when searching for commercial products, right? So they can't be ignored. It's re- your views on a retailer matter enormously, so does your website, those things, right? So keep focused on the basics is my answer.

Speaker 1: I love it. Um, you wrote Yield, and that was really all about Google buying, building, and bullying its way to dominance. Which of those three did most of the work?

Speaker 2: Sure. So the book Yield, uh, it was bookended by the invention of programmatic advertising in around 2005 through Google's antitrust trial in, um, in Virginia, the federal, US federal antitrust trial. Um, and, um, and really I think that the build was the most significant there. So the subtitle of the book, for those who haven't read it, is, uh, How to-- How Google Built, Bought, and Bullied Its Way to Advertising Dominance, and obviously the bullied part, it gets the most attention. Uh, but the reality is, is that programmatic advertising in 2005 and '06 was a total unknown. It was considered a remnant, last-ditch way for publishers to make money. Um, and the idea that it was going to become the pipes effectively for all digital advertising, which it is today, uh, with some exaggerations. I mean, basically anything from streaming TV to mobile to desktop is entirely based on programmatic pipes. That was very unrealistic at the time, and Google did quite a g- quite a job both of building the ad exchange, AdX, to becoming the majority market share leader, um, which did involve some bullying. But also, uh, we don't spend enough time talking about DV360, the buying platform that is effectively number one in market share in the open web market, or at least was for a while. It's unclear exactly. So lot of innovation from my friends at DoubleClick and Google and, um, and they did a, a, quite a job of supporting the open web through those dollars, even if they had sharp elbows and some competitors ended up with black eyes.

Speaker 1: And if the market really does get opened up, who benefits?

Speaker 2: Yeah.

Speaker 1: Buyers, publishers, or a different set of big companies?

Speaker 2: Well, by opened up, are you, you thinking about the OpenAI's challenge to Google primarily? Um, right. Well, I think the problem in the search market, not the display market, the search market, is that it turned out to be a natural monopoly, and a natural monopoly, it also was a, uh, illegal monopoly according to the courts, but it was also a natural monopoly. A natural monopoly is when a market is, um, be- the winner is determined by scale, so you, there can't be more than one winner. And in search, what happened was that the data flowing to Google made their search better every single day. At some point, it became impossible f-for anyone else to build a search engine that was as good, had as much customer loyalty, and all of those things. Uh, and you could do, you could argue Bing is pretty good, but it's just not good enough for consumers. That's really the bottom line. Um, and so th-this is the real question. Will, uh, will a consumer-based either agent or AI chat environment, depending on what ends up winning the consumer space, will it be unstoppable because it's so much bigger than everybody else, or will it be, uh, more like a piece of software where you can choose which piece of accounting software you choose to work with for various reasons, either economic or brand or whatever? Um, and it appears the early betting, if we're sitting here right at the end of 2026, is that chat appears to be a, um, a open market where there's not that much difference between Gemini and ChatGPT and Claude and Perplexity and all of those, and consumers could probably switch between them, um, without losing much. Um, whereas agents, meaning like Muse and Docs and, uh, Grok, GrokBot, I think it's called, um, they know everything about you. You connect it to your bank account, to your email, to... They know your preference. They know your wife or husband's name. They know I just trained my Muse on which sports teams to bet on, not to bet on, to, to follow. Um, makes it a lot harder to switch. If I was gonna switch, I'd have to teach a new agent all of that stuff, right? Um, so there could be some really significant consumer lock-in as we move to agents.

Speaker 1: And maybe some legal challenges around sort of who owns that data and the exportability of that perhaps.

Speaker 2: Well, in the US it'll be fine. In Europe, I'm sure they'll have to register with the German authorities before they take any actions.

Speaker 1: No comment. Um, so when you were writing the book about a company almost everyone in industry de- uh, depends on-

Speaker 2: Yeah

Speaker 1: ... how did you decide sort of what, what went in and what stayed out?

Speaker 2: Yeah. It's a good question because I was trying to write both a history of digital advertising and a concrete story about one group who marched forward and totally took over, and their sort of ethical questions and issues they confronted on the way there. Um, so there were definitely topics that I felt were interesting but didn't tell that story. Um, to give one example, mobile, mobile advertising wasn't, I, uh, wasn't that much of a part of the story, um, because it had totally different dynamics from, uh, web and display inventory. There w- there wasn't, Google wasn't the bad guy to, to make it simple. There was like a host of 20 different bad guys all stabbing each other in the back. It kind of could have made its own book. Um, so right in the middle of kind of the second half of the book, I had one chapter that's like, "Oh, by the way, mobile and video happened too," and I just felt like I had to put that in there because of the history there, but, um, it wasn't that relevant. What was really relevant, and I think probably the strongest chapter of the whole book as a story that hasn't been told, is about the efforts that happened in, I wanna call it 2018, where Apple released Apple Reader, Facebook announced Facebook Reader, and Google announced AMP all within six months of each other. And all three were like, "Hey, we're moving into this world that's really complicated, that has multiple screen sizes, that have different latency. We need to move the web forward into something new." And so all three of them had the same vision. They, uh, they all to some extent, maybe Apple not as much, were wanting it to be ad-supported, and publishers said, "Oh, wow, this could be our savior." Like we could move from the, the open web, which has always had these real business problems like, uh, fragmentation, lack of identity, bad ads, to this more controlled environment where w- it, things could be better. And all three of those companies broke their promises effectively. Uh, Apple was hostile to ads from the start, so no one made any money. Uh, Meta also hostile to ads that weren't sold by them and didn't build like a real partnership with and eventually pulled it. And Google, sort of benign neglect. Like they said the right things but didn't do the right things, and all three of those were dead by, uh, in different forms. Uh, Apple continues to this day, but the, basically not a, not an impact. And I think that was almost like the lost opportunity, right? If, if one of those had worked or if someone independent had done it, it could have changed the nature of a consumer's relationship with content.

Speaker 1: And in the world we're in now and where we're going, what's gonna change day to day for a media buyer, and what's gonna kind of carry on exactly as it does now?

Speaker 2: Right. So I think the story for media buyers for the past, you know, five or six years has been fragmentation because there has, um, since, um, Meta and TikTok and all, just name the next five platforms, all became must-buys, entered the portfolio, um, the ability to optimize across them and to manage creative and things like that has, has Been a story of complexity, and it's not changing, right? If, if, if anything, it's getting worse. Um, and I think that, um, you just have to accept that and, and deal with things like media mix modeling and, uh, a- and incremental attribution and all of these subtly more complicated technologies than the old, um, attribution world where you just basically put a pixel on everything and try to read the reports the right way. And, and I think this new world we're in is better fundamentally because marketers have to ask themselves the question, and not just ask themselves the question, they have to prove what worked and what did not, and ca- how can I justify the next dollar I'm spending based on that data? Um, so, so that's really the tool set that marketers need to be thinking about.

Speaker 1: To your point around sort of Google's benign sort of, um, sort of, you know, action around that sort of what was potentially a kind of new golden age, when you're interviewing these companies, as you do pretty much for a living, how do you tell who's got something real from-

Speaker 1: ... who's just sort of press releasing it out?

Speaker 2: Yeah, right. Uh, so, uh, I, I, uh, record a weekly podcast, and I also write a newsletter about every other week. So I'm effectively a journalist of a kind in this space. I do get a lot of PR pitches, a lot, a lot of press releases, a lot of, "Oh, you won't believe we ju- what we just released," kind of conversations. Um, and I'm, I, by my nature, just accept that there's something there. There's not total BS when they send 'em my way. Um, but you know, uh, you really have to ask like what's the adoption? Is this just a case study or is this a real trend? Why does this make sense? Um, you know, there's nothing really magical to it. Um, the one thing I will say, and I put this in my newsletter j- that came out today as a sort of a joke, but it's true, which is the, uh, press releases that says, "We just added an agent to our boring old SaaS product" are so boring. Like, my God. Like, oh, now with agentic, you know, basically it's table stakes. If you want a c- if you want a marketer or a publisher to log into your UI and fill out a bunch of form, form work, you know, and then use naming conventions to make sure they understand the data when it comes out of the report, and then download the report and reformat it for PowerPoint, you lost. Like, just that's, that's the old way of doing things. You need to be AI, uh, excuse me, API and MCP enabled today, and y- yeah, sure, put out a press release, but nobody cares. Like, just actually use the product.

Speaker 1: And to the opposite of that, sort of where are you seeing AI genuinely changing the economics of ad tech?

Speaker 2: Right. Well, I think it's a little too early to tell. Um, but you sh- would expect that the job of ad ops on buy and sell side would be better, uh, better, it would be improved, it'd be less busy work. It would, the quality would be there. You'd be able to change things more often to produce better yield and things like that. There's so much of ad ops that's drudgery and that, uh, effectively doesn't happen because it takes too long, right? Um, so that's one thing. Um, I think the, the other two things are, first is y- our industry has used machine learning for a decade, um, so many of our online systems are quite tuned already to optimization, and there isn't really an opportunity for AI to just come in and make it better because they're already quite well-tuned, right? So you see less of that, although many times the press releases will be there. Um, creative came along a little faster than I expected, um, so I was a little skeptical on the use of creative, uh, AI, generative creative, um, and it is a, it is a tidal wave, I would say. You know, basically being used internal at agencies for, uh, concepting, for prototyping, and then, um, also for final outcomes. But also in the social world, you know, the, it's widespread. There's no, there's no taboo against using an AI creative in, on the social platforms. I think that's just gonna continue full, full steam ahead.

Speaker 1: Many of the things that you listed off when I asked you the question about being more than a press release felt like to me more like kind of features-

Speaker 2: Yeah

Speaker 1: ... than a product, right?

Speaker 2: Yeah.

Speaker 1: And again, it felt like that's just a race to kind of table stakes and commodity, as it were. How do you defend, uh, the world of ad tech and, and, and, you know, what-

Speaker 2: Yeah

Speaker 1: ... what's i-- if you, I suppose if you knew that, you wouldn't be sitting here with me right now. But basically-

Speaker 2: Huh

Speaker 1: ... what, what, what's your take on that?

Speaker 2: Yeah, there haven't really been any breakout AI companies in ad tech. If you think about an AI-first company that started in the last five years, there are none that are substantial in size, um, right? Uh, the ad tech companies that are winning are the ones that already were part of the transaction, either a DSP, SSP, or someone else in that frame who's now using AI to make themselves more efficient and make their customers more efficient. Well, interestingly, I mentioned creative a moment ago. The breakout winners in our adjacent space are the ones who are doing creative things like ElevenLabs, Twelve Labs, and Higgsfield, all three brand names I had never heard of, uh, 12 months ago and are now, you know, billion-dollar valuation companies because they manipulate either audio or video or imagery in a way that supports our ecosystem. Uh, you could think of them as like fundamental platform pieces. Also unclear if they're gonna survive in the long term, but that's been what's happened so far.

Speaker 1: The other thing that sort of we're seeing is, you know, the fact that sort of, you know, s- basically brands are starting to kind of build things for themselves again-

Speaker 2: Yeah

Speaker 1: ... now they can, right? Sort of on the boards that you sit on and the buyers you talk to, how many brands are really building their own tech that plugs into the ad platforms, and how many of them are just kind of talking about it?

Speaker 2: Yeah, I haven't seen that much on the brand side. Uh, in my previous iteration when I ran a DSP, we often were trying to sell to the brands, and the brands' tolerance for build was almost entirely relegated to companies that were extremely direct response-oriented. Uh, y- as soon as you became brand response or brand and had an agency involved, the tolerance for a build-out really went down pretty precipitously. You might have a team, you might have a couple people in t- in-house, but you weren't building anything. What's ha- uh, what's been interesting, I think, a little bit more is the agency build-out, where virtually every agency, uh, the indies, the big ones, they have CTOs. Those CTOs have budgets. They're actually building things. The holdcos are buying things, uh, that are very relevant here. Um, you know, there's the LiveRamp acquisition, the, uh, Omnicom bought Flywheel, which is an e-commerce company. Like, there's a lot going on in the agency world because the barrier to running your own tech has gone down. Uh, y- the agencies stayed away from the DSP and SSP spaces 'cause they were so high-tech and so complicated and low margin, but that's very different from things like, um, workflow for organizing creators and, and monitoring their activities. That is a doable tech project and worth owning.

Speaker 1: And of course, they themselves under massive pressure on their kind of core and traditional business-

Speaker 2: Yes

Speaker 1: ... right?

Speaker 2: Exactly, right.

Speaker 1: Really interesting. Um, you know, in your view, kind of what should a marketing team own itself and what should it never try to build?

Speaker 2: Right. Um, it's a good question. I, I think that it varies by the marketer's breadth. Uh, obviously a small marketer should build nothing, uh but, uh, as you get larger, maybe you do want, uh, to own more. Um, I, I think that the, um, the question of do, do the ads work, uh, is so fundamental to the marketer and it should not be something that necessarily changes if you change your agency makeup. Um, so I would encourage some degree of investment in attribution, analytics, and incrementality, as well as potentially media mix modeling. Media mix modeling is an area where, uh, in the old world, there were these complicated models built on mainframes and things like that, and it would- they'd run once a year. And now we're in a world where both Facebook and Google have released open source versions, where the data is flowing between platforms quite liberally, and where you can be running an MMM weekly, right? So I, I think those are the sort of investments that make, that are your first step as a, as a brand. Um, workflow is probably much less important.

Speaker 1: Let's talk a little bit about sort of Marketecture for a minute today. The new- the, the, the newsletter that came out today. You know, we, we can, you can tell from this conversation, you're pretty direct about-

Speaker 1: ... products in the market, right? You know, and which is, which, you know, to me is of, of, of massive value, and I'm sure to many, many other kind of, uh, marketers-

Speaker 2: Yeah

Speaker 1: ... in, in the buying kind of cycle. But, you know, in all seriousness, sort of like what makes a kind of product interview useful to a buyer rather than a kind of another vendor presentation?

Speaker 2: Yeah, sure. So, um, what, uh, the founding of Marketecture was about four or five years ago, and the, the initial idea was, let's help people skip the first meeting, right? You, you've all probably been at a first meeting, you get a free sandwich, and a vendor's t-talking about their offering, and you leave the meeting with pretty much no idea whether it's real or not or what it does or how it works because the follow-up item is like, "We'll come back and do a demo once you sign our NDA," and things like that. So I, I wanted to get through that first meeting, um, by, you know, calling some BS or if not even BS. It could be very friendly. Uh, it could just be like, "Hey, I, I heard your pitch. Who are you selling to? What data are you using, and how does it cost?" Right? 'Cause those answers are often elusive, especially on the average website of a vendor. Um, so we do our best to do that. Um, we also cover kind of the press releases of the week, uh, and say, "Well, okay, is this a big deal or not? This just sounds like ordinary course of business," uh, and try to help people interpret it a bit. Uh, I'd encourage people to listen, and it comes out every Friday. It's called Marketecture.

Speaker 1: So you just talked about sort of, you know, the fact that you, and, uh, another statement about a new company has done blah, blah, blah, maybe a little bit unflattering. How do you sort of factor in the fact that that company might indeed wanna sponsor you, and how do you handle that?

Speaker 2: Yeah. So I'm the chairman of Marketecture, and I write the newsletter, do the podcast, which are my own voices, and I don't, uh, alter my opinions based on sponsorship. But then we also do sponsored content where I, uh, I, for example, was just at Amazon Unboxed event in San Francisco and did a couple of interviews there where, um, there was a commercial consideration paid for that. Uh, we try to just disclose it and be as straight as we can about it. Um, I will turn down interviews even if they're paid if it's a company I don't believe in or have some issue with. Um, and we've lost a deal or two because I say what I think and it doesn't always work for the sales force. But, you know, we try to manage it just like any other media company.

Speaker 1: Love it. Let's talk a little bit about the next generation of l- of leadership and where that's gonna come from. You know, if you're seeing agencies and ad tech companies-

Speaker 2: Yeah

Speaker 1: ... hire fewer juniors, you know, uh, you know, how, you know, how is someone gonna kind of, like you, learn the business from the inside?

Speaker 2: It's a good question, and I think this is the crux of the AI question, the question about whether is AI gonna eliminate jobs or is it gonna create jobs, sometimes called Jevons paradox, right? Which is if things get cheaper, you can use more of them, uh, right? It's, there were, there are more accountants than there were before spreadsheets were invented, right? That kind of thing. Um, so but as it specifically relates to young people, they will be the first to, uh, get the negative side of lower hiring because you don't need interns. The AI is your intern. On the other hand, if you're what's called AI pilled, if you come in as a recent college graduate and know how to automate your own job, you're gonna have a big leverage effect that's much more significant than the, you know, the boomer in the corner office who has to print out his emails. So, uh, so I think there's two sides of that, and there's also, of course, the entrepreneurial side, which is, uh, it w- it's never been easier to create, say, an ad agency. Uh, a single person could create an ad agency that has far more impact than a team of five or 10 a couple of years ago if you use AI properly.

Speaker 1: Talk a little bit about, you know, what you see from a board seat.

Speaker 2: Yeah.

Speaker 1: Sold businesses, you sit on boards. What separates the companies that just get bought from the ones that run out of road?

Speaker 2: Right. Well, as a entrepreneur and investor, the number one advice-- Actually, my number one advice is don't go to jail. Number two advice is don't run out of money, and those two things can come into conflict sometimes. So if you're running out of money and your option is to get the money in a way they'll send you to jail, remember the first rule is stay out of jail. So honesty, integrity is most important thing. Um, second thing is, um, you really have to understand where you are in the fundraising cycle. I've been in this long enough to know that there are some really negative sentiments against advertising businesses in the venture capital community. Um, even w- at one point when my startup Beeswax was almost doubling, like 100% revenue growth, I couldn't do a B round. Uh, everyone would say, "Uh, it seems very, you know, ad techy. It seems very, um, you know, non-SaaS." Um, and it was really s-a struggle. So you can't assume the next round's there. You have to be both growing as fast as you can, which often requires you to have some red ink, some loss, but also know where your next funding's coming from and being able to turn on a dime and get to break even with the least amount of pain involved. And, and usually where I see companies have trouble is in, on the sales side, in that, you know, the investment in R&D and product and those things, it's somewhat fixed. You can tell how much it's gonna cost next month, this month, whatever. Whereas on the sales side, there's often a hiring well in advance of revenue, and there's not as tight a management over that, um, that investment in new people, marketing, and things to get to the revenue because the sales cycles are long or it's unclear or you're going after whales and they take forever. Um, so that's really where you can get into trouble.

Speaker 1: Let's just finish up on three kind of quick-fire questions-

Speaker 2: Sure

Speaker 1: ... if that's all right. What's your go-to resource, the thing that you read or listen to, to, to that really earns the time?

Speaker 2: Uh, I love the Hard Far- Hard Fork podcast from The New York Times. Uh, they recently changed their hosts, but it's still excellent, and it's funny, it has got a g- good balance of AI boosterism and skepticism as well as just things going on in the general technology sphere.

Speaker 1: And where'd you get your inspiration from?

Speaker 2: Um, I think for inspiration, I certainly look outside of advertising and tech. Um, you know, I like to read a enormous amount of, uh, the trades as well as, uh, listen to podcasts, and just generally be on the, um, o- outside there fi- finding out what's happening in the real world outside of advertising. Um, one thing I always... I really enjoy listening to another podcast is, uh, The Town, which is a podcast about Hollywood, and that has absolutely no bearing on my life, uh, 'cause I'm not in Hollywood. I just like it. I like understanding a different business that I have nothing to do with.

Speaker 1: And talking of businesses, what's a brand that you are really excited about or admire right now, and why?

Speaker 2: Okay. Uh, there's a clothing brand called Buck Mason that's getting very popular. Um, and I like the way they're branding. So it's sort of, I would call it like a, uh, Banana Republic up- up- a Banana Republic/sur-

Speaker 1: Jake Cruish

Speaker 2: ... surf wear.

Speaker 1: Oh, a bit surfy. Okay.

Speaker 2: It's a little like California Banana Republic. And, um, and they're opening stores like crazy and w- and they are decidedly doing things that are trying to make themselves seem, like, differentiated. Like, the store in New York is near my house, and they actually have a Porsche car in the store, and it's not a very big store. You kind of have to walk around the Porsche to get to the clothing. Uh but I think that's quite, quite interesting. And my understanding is that the whole thing is sort of a new brand that's been fabricated in the last 10 years and is doing quite well. Um, and to get my attention as someone who hates shopping for clothing, they must be doing something right.

Speaker 1: And on that basis, thank you, Ari, so much for your time. It's been a pleasure having you on the show.

Speaker 2: Of course. Thank you for having me.

The Rewrite

Stay close to the conversation.

The conversations and intelligence worth knowing about.

About the guest

Ari Paparo

Founder & CEO, Marketecture Media · Season 7, Episode 174

Ari Paparo is Founder and CEO of Marketecture Media, the podcast and newsletter he built to help buyers cut through vendor spin in advertising technology. A long-time operator in the plumbing of the ad market, he is also the author of Yield, an account of how Google came to dominate advertising, and he sits on boards as an entrepreneur and investor.

Read the interview Full profile LinkedIn (opens in a new tab)

Share