Possible 2026 Day 1: AI, AdTech, CTV & the Future of Media | Signal & Noise Compilation

This is Signal and Noise, bringing clarity to data, text, and AI. Hi, Rio Longa here, Signal and Noise. We're coming close to the tail end of day number one at a possible event in Miami. We're at the Mad Connect, Mad Mansion here, which is a nice spot. Do you agree? Yeah. And I'm here with the one and only Sarah Caputo, or as you say, Long Island Caputo. right? I prefer kudos. It's good to see you. Good to see you guys too. So we've got a lot of great stuff to talk about today. I think the conversation we had before, bring your own models, some new standards that are coming. You sounded really excited about us. I'm really keen to dig in. Yeah, I'm not sure how much you've been talking about that today, but the advertising protocols are coming and they're forming the committees now. It's been something that's been open for about six weeks. So I've been consulting with different CEOs and founders on getting part of that process proactively. And it's actually going to break down everything from the model, which is anthropic, there's audience protocols, there's algorithm protocols, as well as other measurement governance ideas that are coming to the table to really actually level set programmatic landscape for the first time in, I think, 12 to 15 years since the original standards were set by the IAB Tech Lab and others. So it's very exciting. So these are separate from, like, the recent protocols that were, like, for, like, ADCP, for example, or UCP. There are two UCPs. I know it gets a little confusing for people, but, like, how do these standards relate to those? It's really crazy how many acronyms they've come up with. We love them in ADCP. And I actually studied for, like, four hours one night to try to actually digest it and understand it. ADCP is actually a part of the AAMP. The AAMP is the broader overall acronym that they have created. It says the Agentic Advertising Management Protocol, and that was set at February 22nd. And ADCP is the subset of that. If I'm not mistaken, exactly. And ADCP's contract protocol, which is ADCP, is actually the layer where we think and anticipate that the buyers are going to actually gain more control. Right now, the front door to buy is typically the demand-side platform, the DSP, and that's where the brief goes in, audience, like how do you actually want this to show up, what type of inventory. We just load it there, right. Exactly. It's all the basics of how we run a campaign. So consider it the front door. That actually is likely to push now to an agnostic layer above the DSP, even SSP. So the DSP kind of goes headless then? And, I mean, it's more visionary at this point in time, but in theory, yes. It's actually going to end up being the front door where everything lives in the buyer's hands or the agency's hands within the ADCP, which is the ADCPs protocol. So all of that can actually happen now agnostically above the technology. And I'm thinking through different elements like bringing your own algo or models that are available, like a chalice solution. So Chalice Talk, right, yeah. They're going to be able to actually come in and think about the power this will actually give the brand or the agency to actually create their own controls and use that consistently across all partners, especially if you have a less than modest brand, more aggressive budget. They're going to be able to actually take that decision a little bit better but with more transparency and control, again, theoretical. So it sounds like a lot of functionality is currently in a DSP. And this is a trend that's been going on for a while with companies like Chalice that are taking all the bidding algorithms, right? And other people have launched things to do that. So I see why this would be great for an advertiser, probably great for an agency who wants more flexibility. What does it do for the DSPs, though? For the DSPs, I have an article coming out on this in the next few days where my theory is that the walled gardens stay walled. And frankly, from a compliance perspective, the data needs to live within their ecosystem. It's really difficult to share that out regardless. But you think about Google and Amazon remaining the way they are, and even Meta falls into that category. They're not going to share anything different. In fact, they might have owned and operated inventory anyway, right? Right. Plus all the extensions for office. But you're right. I mean, that's theirs. They control it. Yeah. Hence the term world garden. Okay, that makes sense. For open web then? Open web, I think if you're not owning any first-party data, and that doesn't even just mean the tech or the rails, I mean actual data, I actually think it becomes a little bit more, I don't want to say commoditized as the right term, but you have to really lean in, and this is where I've been trying to advise publicly on some of these independents that really lean into service and strategy. That will still ultimately play a huge role in the agencies and the brands that need that arm, that extension of buying, how to measure, how to actually optimize those campaigns and truly map that back to some sort of incremental return, is going to be very important still. So you think about like the StackAdapt, the TradeSk models, where they're really known for their gold-level service standard. I would go all in on that and really over-index on that because it's still very much needed. It's interesting. I noticed that trend with DSPs. Like, for example, as a big global pharma, I helped them with an in-housing initiative over several years. They did not add buying, but like data, tech, a lot of process related to it. The agency was still doing the buys. They even in-house the contracts and went paper direct with the platforms, including the trade desk. What I thought was interesting was the trade desk had almost like a forward-deployed, I don't want to say engineer, but a forward-deployed employee who really had a desk in this time. It was pretty cool. I thought it was a fantastic model. If you need to differentiate, that's the way to do it. Yeah, at Allstate, I had eight hours. I had DSP hours, and I did that actually when I was CDO of a midsize agency to make sure that there was an extension. They actually had space in our offices to do that exact thing. That's smart. So, I mean, think about it. If their functionality migrates, like, okay, bidding, algorithms, some decisioning, if the whole front end of the DSP gets subtracted away, right, in theory, like, their take rates will probably need to fall. But to your point, if they add wraparound versus buy these other things, it could also be a different model, though. I mean, I think it's going to be a pretty big shakeup in the space because you're selling software that has a take rate versus selling consulting hours, regardless whether it's time material or a fixed fee with your margin is going to be, let's say, you know, roughly 50 percent versus with software where your margin might be, you know, it's like 90 percent. Right. It's my guess is this will be impacted, though. Yeah, that's actually where I think that buyers are, and I always say, are willing to pay more for service. And I think about that and how it equalizes on the back end for the brand and the end campaign. I don't care where the hours show up. If my agency is spending less because they're getting more from their desk, it all equalizes. Like effort and value out is how I consider this to be. So in terms of that economic pattern, you see that the independents do actually usually have a higher take rate because of that service. So ultimately, you're right. They have to make up for it in some way, shape, or form. But you can actually justify it to the brand or for the brand if they're, you know, basically contributing to what I call FTEs, the full-time employee model. Because that's money that's being saved somewhere in the chain. Yeah, I mean, there's a lot of debate now. I'm sure you're in the debates about EFC Multnomah impacted, right? If a lot of these things are automated, I don't see a world where there were no people, but there certainly will be fewer people, which is probably a good thing anyway. Yeah, so even with Agendic and the front end of some of these DSPs that are launching, I had a conversation with someone from Goodway earlier, and I was basically saying that there's so much that our teams can't even do to get into the supply, where it's showing up, the ad load, the frequencies, and other things that are actually in the data, if you dig enough, that we can't actually, again, optimize to the extent that we should. So I'm hopeful that it's less time clicking buttons to actually set things up in more tedious tasks. And we actually move it to signals, signal and noise. So the signals are there. they're actually becoming more granular and at a scale we've never seen because of the LLN's ability to actually parse and pull that out. Well, I think that's Chalice's whole, not whole, but like a part of their proposition, as I understand it, we had Adam on the pod a few weeks ago, is that all of that incredibly rich signal on the supply side, which DSPs only, because of the speed of the processing and the volume of the data, they can only take a small subset of it, really, in order to make their decisions. Like his explanation was, well, we have all this rich signal. We're building an LLM that we can deploy to places, and this LLM is going to have all of the signals to make much better decisions on. So it makes sense what you're saying. Yeah, I hope so, and I hope we can inverse that also to contextual. That's something else that I find interesting. If you actually start to model around the behavioral signals that exist, the signal doesn't exist because of the consent problem and issue that we're all facing. So I think that there's ways to use that data. Then you get away from the ID trap, too. like everything needs to be, and even the way the DSPs, like the way they buy, a lot of people like are not super deep in that take, might not know this, but like traffic with IDs is prioritized by DSPs. So I think having the able to pull in better contextual signals, it should be very powerful. Yeah, I believe so. And the data on what's actually missed in the targetable population is huge. There's a lot of good inventory that's clean that actually fits the market audience that doesn't actually see the time of day. So it's inflating CPMs and constricting the supply where we really shouldn't. So hopefully opening this up a little bit will help us to solve that problem. What's the timeline for some of these protocols? I mean, I'm not going to ask you for a timeline for, like, DSPs. Maybe you have a guess, but, like, just the timeline for these protocols being launched, and, like, what are you seeing in terms of experimentation so far? They're actually starting now. I mean, Index Exchange just launched part of buying within their own system. because they also, the SSPs have quite a bit of signal. And then you're seeing parts of the ecosystem put together the agendifying, either from research and planning to the actual setup. Then you've got the optimization and that measurement and attribution on repeat on the back end. So those are the four cycles that I really, you know, brought my head around in terms of the milestones. Each of those are actually starting to become more agendic. So you've got talk to your data. You've got the setup on the front end, chatting with the interface in order to optimize. I am also consulting with a few of the measurement and optimization kind of partners that show you agnostically above the DSP what you should actually optimize to and why. And then you can also flip between the models. There's one partner called Slingwave I really love that's got some interesting tooling. But it's happening now. It's really a long story short. It is happening. So it's a matter of when we get to actual agent to agent and going together to pieces. It's just going to be trial, error, and time to kind of start to push us forward with more momentum. People don't realize even, like, you know, programmatic when that came out, there was a lot of, you know, started small. There was a lot of trial and error. It's not like it took off overnight and suddenly, you know, most traffic was programmatic or was biddable. And it took a while. So that does not surprise me that we see something similar in terms maybe a little faster. I think tech accelerates a little faster now. And certainly with AI, I've never seen things change as quickly. But you're right. I mean, difference between things being possible, you know, no pun intended, and things actually being, like, being released. Those are different things. Yeah. The acceleration's there, but you always have to test thoughtfully. Always have to test budget and test thoughtfully. That's just my biggest advice. Honey, that's a theme that's come up a lot today, Sarah, like the whole, like, back to basics. Like, okay, we, like, marketing really shouldn't, like, there's no magic formula. Like it really should be about testing things and trying things and seeing what works and then seeing, okay, can things work better, right? And then maybe something works today, but it doesn't work tomorrow. Or maybe something worked yesterday and it'll work next year. So I think constant testing and trying and being flexible and not – we're not trying to prove or disprove anything. We're just trying to see what works, right? Right. I always love tests. It's always what I start with in any sort of brief. Usually at campaigns, and you're in this world, there is always a research brief to get to the media plan and then also the creative. But where does the measurement brief show up? And so I'm a huge proponent of three briefs at onset, including the measurement brief. I like that. Testing and learning needs to be in there. Your source of truth needs to be in there. You've got micro and platform metrics, and then you've got your out-of-platform metrics, and then you've got your BI layer, which is the layer I love to live, is making sure that that's intersected as much as possible to truly understand incremental business. So the measurement brief is something that I really want people to take this conversation. Incrementality. I love it. You can't. And that's another one of these words that I think it came down. Maybe it started last year, but like getting, okay, like proving that your marketing actually is effective and that it works and being able to understand what works, what doesn't, and heaving up where things work and maybe dialing back where they don't, right, or they work less well. Yeah, in platform and out of platform. And the last point on that is that these tools are starting to allow to do that. There's match market incrementality to be very difficult to do. You have to actually assess the data, power, lift score, all the things that are included in measurement and testing. There's actually platforms now that can automate that as long as they have access to your business intelligence to break it down and then give you the markets and the test protocol. It's kind of thing. Any big trends in terms of measurement? I know there's been a lot of people saying how they don't, platform metrics, I mean, shouldn't be trusted. There's a big debate on Twitter recently. Some people saying they should be, like, they're better because the signal's straightforward. Other people say, no, you really can't trust them. They can be gained. Any thoughts there? I always start with a layered model, like a short-term, mid-term, long-term. Long-term is going to be more of your, like, longer trackers, you know, and then we'll fit into that. Sometimes it can be a little bit more frequent. But on the short term, I actually like to use the longer term and midterm tests that are done outside of the platform and basically incrementality tests, which is my gold standard. I do those frequently, usually every three to six months to understand what haircut to take to the platform metrics so that you can still actually use the platform metrics that are overreporting. And then the second piece to that is just making sure that tagging and tracking is constantly checked and that you're not, you know, with any blind spots, double counting or over counting in ways that you can actually control. Because at the end of the day, they will over report in most cases. But if you know what haircut to take to that directionally, you can actually make very quick decisions and then constantly do that on repeat so that you have this like checks and bouts. I like that. So it's not like, oh, they're good or they're not good. It's okay. Well, like there's something you can glean from these. Yeah. Don't, you know, take everything with a grain of salt. Correct. Like, compare things to eliminate duplication, overcounting. And, like, I think when you look at several different, let's say, reporting environments, looking at the same data, like, when you extrapolate, you can find the truth. Yeah, you can definitely find directional results. And then from a source of truth perspective, it is the longer-term tried interior testing. And just as long as everyone knows it's going eyes wide open, that sometimes those models take some time. So, like, new, for example, new platforms and new partners are going to take time to ramp up. Regression modeling requires a lot of data, you know, and a lot of history. So those things will take some time. That's why I love an incubator or a test budget. So it goes back to that other point is, like, don't actually put the test budget within the results that you should expect and don't actually make really quick decisions on anything you're testing that's new. Because at the end of the day, sometimes that takes a little time. Well, this has been a great discussion. Before we go, I did want to comment on the watch. That's vintage, right? It's 1985, yeah. I love it. Where did you get it? My director of technology, when I was CDO of the agency, gave it to me as a holiday gift. And his name is Chris Van Meter. Hello, Chris. And, yeah, it's just a good reminder of how outdated good tech gets so quickly. And for those who can't see it, like what brand is that? Is that a Casio? It has like the vintage, like old school keyboard on it. I think I used to have the same way I should have. It's probably worth a lot now. I should have saved it. I don't even know how people use it. The buttons are so small. It's insane. You need like a pencil or something. Yeah, small fingers. It's so hard. Yeah. Well, this is a very great discussion. As always, thanks for the insights. It was great to catch up. And we will, I'm sure we'll have you again on Signal and Noise for a longer form podcast. I assume this was a great little segment for Possible. Thank you. I appreciate it. I'll see you at the dinner tonight. For sure. Hi, this is Rio Lonnie here, Signal and Noise, and we're at Possible event in Miami, and this is the end of day one. It's been a great day. We've had lots of great discussions about everything from, of course, AI, supply path, optimization, attribution, targeting, measurement, you name it. To wrap up our day here, I've got a special guest, M.K. Marsden, And you are the CEO of Sales Sleuth, but you have a long and distinguished career in advertising and marketing that includes senior roles in lots of different companies, including Axiom, Epsilon. I can rattle a bunch of names, but it's a cool career. I think we can probably touch on that a little bit, but I'd love to first start out by maybe telling the audience a little bit about yourself and your current role as CEO of Sales Sleuth. Okay, great. First of all, thank you for having me tonight. I appreciate this. And by the shout out to Mark Sabatini for introducing us. Yeah, yeah. And that's all part of it, right? How you treat your network, trust that you build over decades and that you continue to build on. Totally. It's a big part of what we're doing right now at Sales Sleuth. But yeah, I started, I'm a computer scientist by accident. I wanted to be a vet. That didn't work, so I went into computers. And I started in IT software. So one of the companies that I was involved with was Nobel. We went from zero to $1.82 billion in seven years, and I thought this was magical. And I believed that computers were going to make the world a better place, and everything was going to be automated. I grew up on the Jetsons. It was going to be fantastic. So I got tired of IT software, and I was like, how do people use all these tools we're creating? So I went into MarTech, and there. You're trading the IT software. Well, yes and no. They're different, right? Like, it's a different problem that you're solving. So I created transactional email. I was like, oh, my God, we're in marketing, and everybody's order confirmed and shipping confirmed is in courier text and is ugly, and it's open. Everybody opens it all the time. So let's invent that. And now what I'm looking at is just B2B sales is a mess. And post-COVID, we've lost procurement. You've got 22 people involved in every buy decision. To thank for creating procurement. You're welcome. Yeah, whoever thought we'd want procurement, right? We can go back to our politics from last night to talk about things we didn't think we would want that we got. So now you've got 22 people involved in purchasing enterprise software or enterprise solutions. They don't even know how they buy as a company. and here are the poor sales reps with no proper insights. There's nothing in Salesforce. There's the lies we tell the finance department so that we don't get in trouble and we might get our commission checked. So there's no intelligence. Tagging pages and putting predictive analytics, that's not intelligence. That's tagging. So I was sitting around doing it the old-fashioned way, and I was like, no, there's got to be a better answer. So we decided to build the first buyer-centered sales intelligence solution for enterprise. And selling to salespeople is fantastic. It's interesting. I think that, like, you're looking at within MarTech and AdTech, and I think you could add sales tech, service tech, right? These are distinct. I don't think they get enough coverage, to be honest with you. And I think especially sales tech. Because you think, like, what is considered sales tech? Like CRM kind of is, right? Yep. Could also put like tools like Sixth Sense, like lead management, things like that, sales automation. But it's an area that is huge that's, I think, underappreciated. I would agree. And one of the things, I thought I would go raise a round of funding. VCs have no interest in sales productivity. So think about that. Why is that, do you think? Got me, right? Like you're investing in a company. Wouldn't you want them to be able to sell their solution? Like, wouldn't sales be the first thing? Like how you ship software, right? I know. So isn't it the most interesting thing that we've never provided salespeople with the actual intelligence they need to do their function? And they're the most critical function to the growth of your business. Totally. Just until you think about it, like, the best salespeople are so much more productive and better than the rest of the salespeople. and, like, I've been having worked with lots of salespeople in my career. Like, if a salesperson is amazing, you'll put up with a lot, right? Yeah. Because they're so good. They're so effective. So, like, what do you think – in my understanding, a lot of the sales tech tools, okay, let's find out what these best people are doing. Let's give them tools, processes, and solutions that can help the rest of the salespeople be closer to that. Do you look at it that way maybe? So I look at when we created digital marketing, we created a landscape where we could see what buyers were consuming across the entire Internet. And what we hadn't created yet was a reader to read all of that. And with the large language module, they're great at reading content. content. A lot of content very quickly, sure, yeah. Right? So we read everything about the seller's account, like everything, their SEC report, their Edgar report, what's that on Reddit, what's on Capterra, what's on G2, what's in Quora, everything. And then what we do that other people won't do and a software engineer would never figure out is we score it against what the seller sells. And that's where my 40 years of sales expertise combined with my computer science work comes in. So why do we not have sales tools? Software engineers know nothing about sales. No disagreement. They don't, right? Right. And so it took a salesperson to stop and crush my brain for 18 months and get the help of people like you and Sabatini. And it took large language modules to get out of the sausage factory of first-party data and third-party data to be able to, like CEOs are publishing their mobile phone numbers in DMs that are in public domain. I've seen this more and more. I think it's a really interesting development. I'm seeing CEOs put their numbers, put their emails, and I'm imagining people reach out to them, right? I mean, they must be. No, people like visit. It's like his phone number is in his signature line. What do you mean you haven't called them? What is wrong? It's a couple of days, but like the salesperson, I have no idea, no way how to get nobody older than this file. Right? Because it's like they can't imagine it. It's like it's right there. He put it on his LinkedIn profile. Which apparently he didn't go to, right? Exactly. And partially it's because you're managing a book of business that's 20, 30, 40 accounts. You're spending 20 hours a week updating Salesforce, attending training, doing QBRs, doing RFPs that you're never going to win, that are cut and paste from the last RFP that didn't win. So you're spending all this time doing these things that aren't even material to your selling company. I remember at a large consulting firm I worked at, RFPs were like when you're on a bench, you would respond to RFPs. I remember the success rate was under 10%. It was just like a death march. You would do these 150-page responses. I remember, like, I'm not exaggerating staying up all night to do these damn things. This is when a junior consultant and then knowing that, like, there was zero chance you could have waited. It was totally demoralizing. And you're giving away all your IP. You're giving away your pricing. You're giving away all your IP for nothing. You're getting nothing in return. In every RFP I've ever written, in any company I've ever led, I've offered $100 in the RFP response. If you read this bar, call me. I'll give you $100. I have never paid it. Never once. And I've done that. So, like, you put it at the end of the RFP? No, it's not. Okay. Like, in the privacy bar, they're governance. There's something you know that they're just, like, never going to read. And they don't. And you're like, it's all insane theater. It's theater. It's what it is. It's sales theater, I guess you would say it is, right? It's like, we're going to pretend that we want this, and you're going to pretend that you care about making it. And you're going to pretend it matters, and I'm going to pretend that. We're going to pretend we read it, right, and then we're going to award it to McKinsey anyway. Yeah, because no one gets fired for that. Because the CEO is former McKinsey and he's best buddies with the partner. I mean, that's just. He just played golf with them. It's on his Facebook page. We can tell you that. Well, like, there's a lot of software sales, too, right? It's like, oh, like, because the salesperson went golfing, you know, or this kid goes to school with the CEO's daughter. Totally. So it is about relationships. But if you don't know what their problem is, like imagine starting a relationship where you ask no questions of the other person. It's kind of like dating in 2026, right? Like it's bad. It doesn't lead to anyone. It seems to be as bad, right? It's bad. Rest assured. You know, how can you build anything? So if you're not doing discovery, if you don't have a good sense of what their business problem is, how they're buying, what they're worried about, what their existing technology is, you can't. And if you've got 22 people that you have to know on each account, and if you've got 40 accounts that you're accountable for, humans can't do this. So we took the human sales brain, the best of the human sales brain, and put it in a platform so salespeople could do what they're really good at. And is this the salesperson in any industry, or is there a specific vertical you're focusing on? So I haven't, and I got a lot of advice today at Possible to focus on verticals. It's a self-learning system. In theory, it can work really anyway. In theory, it can. The more complex the service offering and the more complex the purchasing, the better for us. We can make a bigger difference faster in complexity. So if you have one product and 10 customers, you don't need me. If you've got a product offering of 70 products. 10,000 customers. And there's 22 people on the purchase. If you're a data center guy, there's 40 people involved in the purchase of a data center. where those people came from, who they have relationships with, what their bias is. Some are just influencers. Some can derail it, like the legal guy or the privacy guy or the infosec person. They're not signing a check, but if you piss them off or you run afoul of them in any way, you're not going to win that deal. Totally, totally. And they're going to take great joy in derailing it at the end. So, again, like people put, you know, 16 months of effort into it on the buy and the sell side. and you don't even know who the guy is. He's working at home in Utah and you're working at home in Chicago and you've never met and you never will meet because they don't come to the sales meetings. So the buyers don't even know and then they get bamboozled so they're not getting what they need and everybody is just putting so much wasted energy in the process. What did you see today? I don't know how many sessions you got a chance to go to. I know you spent some time on site. Like anything you see jump out that you think is really interesting, that changed your perspective, or that's maybe a cool trend that is coming this year. I mean, I was here all day, so I didn't go to a single session. I'm actually very interested in hearing. So I got really, really affirmed. I think I'm really on to something. I didn't bump into anybody who is doing what I'm doing. I didn't bump into anybody who said, oh, I don't need that. I didn't bump into anybody who's crushing quota, hitting their growth numbers, and everything's working. We have unbounded enthusiasm at these events, right? But it's like, okay, I have a widget and you have a thingy and I have a problem. There's no assurance of enthusiasm. No, no, no. It's unbounded. It's so loud in that lobby I can barely hear. I had a headache last year. And I think for a couple of days my hearing was slightly damaged. It was so loud. It was worse than the last concert I went to. Tell me about the volume of people stuffed in that bar area, but then the ceiling echoes. It doesn't. It's so freaking loud. I remember it was terrible last year. So I went to the AI sessions, obviously, because that's what I'm interested in. We get helicopters here. We don't get loud conference rooms. Yeah. It's the guys coming home from their hedge fund to their $20 million houses in Miami. Again, like the way that we're using the tools, I always try to solve the problem from a brand new lens and a brand new design. What we seem to be doing is taking crappy data or a crappy original structure and doing more of it in AI versus sitting back and saying, what is it really from a brand standpoint that we're trying to cause here? And to whom and how do we do that? Right. We just seem to run to the next innovation. And I'm all about the pause. Like, stop. There's some – we have too much stuff. Well, I think in the industry, we – I agree with that. We love our trendy, cool, newest, shiniest objects. Yeah. We chase after them incessantly. And every six months, a year, there is one, right? I mean, there's a new thing. And, you know, maybe AI – being within AI, there's different – like, it'll be the Gentic AI. And that just really started at the end of last year with some of the LLM releases, right? Yeah, the way that we're calling agentic AI now started, you know, a couple of years ago. If you're my age, you've known about AI since before you started in your first year computer science class, and you worked in predictive analytics, and you worked in data and machine learning. And so, you know, again, like here goes the marketing wagon of AI and agentic frameworks and, you know, workflows and things like that. And we've been doing a lot of this stuff for a really, really long time. Sometimes maybe some of the old things weren't wrong, right? And we shouldn't forget that. It was interesting because I actually was a, I don't want to say computer science dropout. My first year of college, I studied computer science. It's funny. I remember this was back when it was like COBOL, FORTRAN, PASCAL, right? That's what I did. And then I didn't love it, I'll be honest, right? Which is, you know, I liked coding just for fun, but I didn't like doing it for assignments, right? And then I was talking to my father. He's a nuclear physicist. nuclear physicist. I remember he was saying how COBOL is such a great language, you know, because it's a pretty object-based language. He was saying how he has programs that he wrote in the 60s, mid to late 60s, that he runs in the detectors or particle accelerators that work better than the stuff they have. I thought that was wild. Yeah, no, and it's true, right? So I started in the same environment that you did. My first company that I'm Canadian, The first company was an object-oriented database, and we sold it to the Americans, and I end up in the States because my colleagues don't want to come. And then we regressed to SQL. I was like, why are we going backwards? And now we're just finally really into object-oriented databases. This was 1986, right? We have these texts, and then we go running around in the new VC-funded, shiny thing that has lots of marketing, and we don't ask ourselves, is this worth it? Is this the best thing? Is this the tool I need? And I'm a little bit worried. I'm writing, you know, some stuff on AI governance at the enterprise level and things. It feels a little bit like we put nuclear bombs in children's lunch pails. Right. We don't let people build bridges without a certain degree of education certification. We don't let. Learn the hard way. Right. Or we do. Yeah. I think it was in Washington state or something like that. Right. Or the boat hits it and everything collapses. We don't let, you know, people fly airplanes. Whether they're gas or electric, because without the proper training, that's the proper training. Why are we letting people with no training write in large language modules and pretend they understand data and pretend that they understand the needs of the enterprise? It's dangerous. Look, getting back to sales, I mean, I think that, and I wrote this article about this, published the end of last week about like this, how, I call it the IRL, the in real life rebellion, how like the whole, what made me think about this was I remember I was doing a stint as a head of commercial for a software company last summer and we sent out thousands of emails to a really good list. We had thousands of opens, not a single click. And first I thought, okay, the email's broken. Like the links are broken. Something's wrong. I mean, having worked in email marketing for years, I would joke, I'd probably sign insertion orders for billions of emails going out right back in the day because a lot of early media was email. Yeah. I didn't realize that. So I knew a lot about email marketing. Maybe not as much as you, but I knew a lot about it. I remember thinking, oh, it's just something's broken, right? Then we did all the testing. It wasn't. And I realized the way people interact with brands, work with companies, make decisions on what software they're going to buy and which companies they want to work with has totally changed. And I think that this means that where the rubber meets the road is, yeah, marketing has to change. I think it's going to be more, maybe more pulled and pushed, but also sales people become more important. I agree. If people are tuning out, you better have a good sales team, right? Yeah. Yeah. For sure. And when the salesperson actually gets that meeting or that nanosecond, I literally gave an elevator pitch coming down the building today. In an elevator. In an elevator. I love it. That I waited for forever. It was like my first real elevator pitch. I mean, the guy goes, what do you do? And I said, oh, I have. Funny you asked, right? Yeah. And he's like, I need that. He's like, I got an ICP of 1,000 and things like that. I'm like, fantastic. How do I get in touch with you? Oh, take a picture of my LinkedIn profile. Boom. Connect with a guy. Done. Between floor 30 and ground. I love it. People want to talk to people about their business problems. They want to look you in the eye and know that they can fix it. And you could have sent this guy 50 multi-touch emails. You could have, like, lit him up on Facebook with the same audience. You could have, like, and he wouldn't have cared. Nope. Right? Would have tuned it out probably. Completely. Wild. Completely. So, yes, I think salespeople are critical. I want to be the champion for the quota-caring salesperson, account manager. Say it, you know, we fire them. We ruin their careers. We, you know, they have. We put up with them, but without them, we wouldn't have any revenue. Totally, totally. That's funny. Well, this has been a really cool conversation. I'm glad we got to have you appear in the pod, and glad I got to hang out with you and get to know you this time. And we would love to have you back for a longer-form discussion on Signal & Noise at a time soon. I would love to do that with you. Awesome. Thank you. Thank you. Bill Longacre here for Signal & Noise at a possible event down in Miami. And it's a beautiful day. This is day number one, and I'm here with Obriana Alvarez Lopez. It's a mouthful, I know. It is. It's a lot. I love it, though. My name is Rio, so I get a lot of – it's a little easier. But great to have you here, and I appreciate you coming down to the mansion here. Yeah, thanks for having me. Excited to be here for today. So love to talk about, number one, tell the viewers a little bit about you, your background, and the company, Ignisio, that you're the CMO at. I am, yes. So I'm Aubriana, and I've been in the data and ad tech space for a little over 15 years now. And so I started off with working with lots of data, lots of IP intelligence, and, you know, that was working with advertisers. And then I worked at Samsung Ads for a while, so cut my teeth a little bit on the CTV side of things. Was this right when you were kind of starting out Samsung app? Yeah, it was pretty early days. I was there kind of in the beginning of that transition, and it was really interesting to see how that developed over time, right? And then it was three seven figures for that side of the business, and now it's massive. So it was really fun. It's a huge revenue stream for them. So it must have been cool during that beginning in New York. It was so fun because it was kind of like this family inside of this massive organization, organization and to see how that tracked and, you know, grew, it was really, really fun to be a part of. But something that, you know, over time resonated is that there's a lot of data, right? Data was the new oil 20 years ago, and then that expanded. And then now we have all of these tools and technologies that have come into this industry to make things more rich and, you know, better and more exciting. And I just saw, you know, there's still a gap in terms of like how things come together, right? Even within one organization, how do those tools and technologies come together? And so that's really, you know, the premise was at Agnesio, people should not be shipping spreadsheets and sending slacks at midnight. We do a lot of that. We still do that, right? A little bit. But, you know, there's things that we're like, we should not be paper pushers as human beings. And we have a lot of worth and getting back to human connection. And so the premise was, there's a way to fix this. And I think we can use AI to do so. And so Agnesio was born. So looking at the problem statement so it like let me see if I getting this right So like all of these manual processes within AdTech I mean is that fair to say you think that what broken or that a place where there tons of room for improvement right Yes. I think fundamentally we have moved from less tools to so many more tools. However, those tools are very siloed. And so what happens in one area of the business for an organization doesn't translate to the next. And so there's humans that are standing there and stitching things together. The analogy that I like to give is that I think ad tech is like an airport. Let's call it Atlanta. You know, I was in Atlanta for 11 years. Nobody's thinking of it, right? You know, you have a train at the airport, and there's no air traffic control. So imagine, you know, everything is having human handoffs. There's humans standing between different data sets, and they're directing that together, just like at an airport. And then on top of that, each of those tools or technologies are speaking different languages. So, yes, fundamentally, humans are stitching things together. Something that should be maybe one workflow is various and separated and takes weeks and months to do. You think about it from the planning, there's groups of people doing planning, there's other groups of people doing activation, there's different people who are probably involved in creative and audience development and third-party data, and then different teams for measurement. You're right, there's tons of handoffs. research and what we found is there's 53 days for the average campaign to go from conception to actually like liftoff and that's about 17 different people. Long timelines a lot of people I guess is not surprising for anyone who works in media but I think for people who don't have a media background it's probably a shocking number. And then add in how many tools they're using maybe 20 25. I'm not a math person but if I were to like think about all the complexities there I'm pretty sure it's a high number. It's a big number. Yeah. Yeah. And then, and the, and the number of people, number of tools, the time minus self, like I, again, I don't think that's super shocking. Nothing to go wrong. Yeah. Never. Right. Yeah. I mean, I think about every handoff, there's an opportunity for something to go wrong. There's an opportunity for someone to do a fat finger mistake or something to put the wrong thing in here. Even I was working with one client just on looking at are the right, is the right creator for the right ads serving up based on the metadata and like which geo they're in, which audience segment, which offer even. It would be shocking because compliance can sometimes be really low. And not only from a compliance standpoint, but from an actual like operations standpoint, how often do you get that information back in real time or in time enough to actually do something about that? The campaign's already run. Sure, yeah. Ads already started, budget's already spent, right? Yeah. Yeah. So, yeah, I agree it's a big problem. So AI, I guess the premise here is that AI opens up the possibility to finally do something about this, right? Well, yes and no. I'm glad you asked that question. I think, yes, I think there's the ability for AI to orchestrate. And when it's done in the right way, there is, I think, a lot of endless benefits to this industry. where I think we're getting it wrong, right? The second, though, is that the conception maybe is that AI just helps everybody move faster, but AI in the setup and the way that we have things in this industry today is just chaos at machine speed, right? Like we are just providing yet another tool, yet another technology, yet there's no operational layer that's stitching those AI tools and technology together. So unless there's an operational layer or orchestration layer to bring it all together, I think there's still a big gap. Who do you think is going to, like, okay, so you have the different tools, so it sounds like the tools are going to stay there, but there'll be a layer built on top of them that, like, you know, they can connect to that by MCP server, and then there'll be some kind of agentic plane almost on top. Like, is this something that, in your opinion, a brand should own, an agency should own? Would this be a platform thing? like what would be the – where do you think this sits and resides, or does it not matter? I think it really does matter. One of the most important things, I think, is where it sits. So who owns it? Do I think a brand, an agency, a data platform should own an operational layer? Absolutely. And I think the reason why it needs to be owned is anybody who has any type of proprietary data or is bringing in client data, brand assets, they need to have some level of control. Now, who owns it and how it's owned is more important. And so is it owned specifically within their own client cloud, or is it something that they're shipping data out to? I don't think that anybody is ready to do that, right? You don't want to give up that control and governance. So AI – That's your big brand, yeah, for sure. Well, nobody does. I mean, you know, retention of what is your bread and butter or your, you know, je ne sais quoi, you don't want to be giving that over to anybody. And so that was the premise that we built Agnesia with, is saying, you know, we certainly want our clients to have speed, but speed without governance and efficiency is not really a win. Is it fair to say that a lot of the stuff being done in AI, like governance, the governance is where it should be, in your opinion? I don't think governance is where it should be in most cases. But there's a big spectrum of how companies are incorporating AI. Some are just like, no, we don't want to do it. some are like let's test everything but we don't really have a plan for how that's going to roll out i think that governance is not the antithesis to speed honestly it's really what the unlock is so if you have an idea what we do with ignitio is if you have an idea of how you want your governance to be that's the way that your setup should be from the get-go once you have that governance in place it creates trust and you can run fast right yeah so when the cmo trusts that you have an agentic layer and operating system that's making the choices that are best for your brand, then you're going to be able to move with agility. But whenever you're kind of stuck in governance post-tool integration, that's a tough spot to be in. Who are the users of Ignis are going to be? Are these media buyers on the agency side and the brand side? Maybe both the brands of in-house certain things. Who do you see being the people using this? There's a few different ICPs, if you will. There's a few different hands-on keyboard users and varieties. I'll give a couple of different examples. We work a lot with data platforms. So you think about, you know, companies that are providing, let's say, audience data. The way that that audience data is built and the way that they bring that in comes from a lot of different data streams. Perhaps they have an SDK. Perhaps they have some type of panel data that they're bringing in. And all that data has to be enriched. It needs to be normalized. It needs to fit with a structured taxonomy. And so the way that that process happens, you know, you have data analysts and engineers typically pulling that together. Well, that's a use case where a data analyst or somebody who sits on that side would use Agnicio to pull all of that together. And now instead of having, you know, a team of 10, perhaps they have one that can oversee a team of 10 agents that can pull those things together. Another use case would certainly be on the media planning side, as you brought up, Rio. You know, you've got planners they want to have. One, how do I very quickly and effectively respond to these RFIs? Because sometimes it's the first one that responds that gets the difference, right? And so how do I be really quick and effective? And how do I utilize my historical data, my historical campaign data, brands that I've worked on, products that I have available, and fill that out without having to rely on humans to respond in one, two, three days? Sometimes that's just the gap that gets the business, right? And so that's another key area. And then I would say finally on the brand side, yes, we're starting to see more of insights, right? So how do I effectively measure this across all of the different measurement tools and technologies that I have, but in one key place that I can quickly and easily ask questions in natural language where I don't need to do SQL queries to find out something? Or how do I leverage insights that I don't even know exist there, right? Like what are hidden audiences that I should be going after that I'm not leveraging? So I think a lot on that capability is using a niche too as well. So as AI starting to automate a lot of these processes, where do you see the role of people and humans in this process? Do you think that the whole thing eventually is automated and there's almost no people? Do you think people will remain heavily involved or too early to even know for sure yet? I firmly believe that humans will always be part of this industry. I think one of the beautiful things about being in this industry is connecting with the great humans that we get to. Humans of Atik are wonderful people. They are, you know, because let's face it. We're not like saving children or even puppies here. We are selling advertising, right? But I think that it's so important that we maintain that human connection. What I think AI does when it's used the right way, especially with Ignatia, the way we think about this, is that it frees them up to actually be strategic. I think we live in this world where we're constantly in this response frame of mind. We don't have enough time to ponder, to think, and to maybe allow our minds to linger a little bit about the what-ifs for clients, right, for use cases, and this is really what I think is an unlock for us. It's interesting you bring that up, because the one criticism I've been reading a lot recently about AI is, well, if you use AI, your brain is going to atrophy because you're not going to be doing certain things. And, okay, maybe there's an argument for that. But I think I tend to agree with your point, which is actually saying the opposite, saying, like, I didn't think about the way we run this podcast. There's no way we could do this with so few people and all the moving parts without a lot of AI. Now, AI sometimes isn't perfect. Like the discussion guide you sent you, you got your title wrong, right? So it does make mistakes sometimes. But when I think of all of the documents it prepares, briefs, summaries, incredible recommendations for questions that will come. We can do that in our own, but it will just take longer, right? So using AI to assist all that, I do think it frees us up. So I think I was just looking at this anecdotally, this one example, like, you know, from running the podcast, I think it does free us up to do more strategic things. Now, I do think people tend to prefer or gravitate towards doing the things they like to do. So I think for certain people who like operational things, it may seem more of a threat. But I think people who like to think strategically, use their gray matter, solve the big problems, I think it's a big unlock. I agree with you. Well, I mean, let me just step away from, like, founder, CMO role for a second. I think that it's a refinement, right, of my thinking process and strategy. So we've lived with this inner dialogue our entire lives, but now having something to kind of process that inner dialogue with just a little bit when it comes to your ideation, sometimes is the thing that takes your idea from, you know, two to ten, right? And I believe that that is where the humans are always going to be in the loop. I use it all the time to, you know, refine this idea. This is a terrible idea. And you don't want to have that conversation with a real human all the time. Sometimes you just need, what other, you know, challenge this. Play the devil's advocate to me, right? And I think that there's a place for that because sometimes it just gets your mind, you know, sparked into another dimension of what could be, what should be. and I think that very much applies in an industry that is fundamentally creative, right? We've taken the creativity a little bit away from what we do in some aspects and I think that we have the ability to get that back. So I don't know. I would push back on that a little bit. You know, I think that we did stray from creativity. I think this obsession with vanity metrics, performance, I mean, which, you know, pushing people through a funnel. We've talked to other people about this today, too, that maybe it went a little too far. I get it. That's how the Internet got wired, and that's what we trained ourselves to do. But focusing more on creativity, on messaging, and if AI helps, that's great. But I think you're right. I mean, I'm seeing people focus on it more, and the people that use AI the best, I think, almost paradoxically, are the ones who tend to be the most creative and have the most experience. I find that. Like what they're getting as an output from using an AI, using, let's say, an LLM, is qualitatively much better than something that, you know, someone who doesn't know how to write a prompt, doesn't have the background to use as an input, is going to get out of it. Yeah, and I feel like it's a little less about the prompt because are there people who are going to have their roles replaced by AI? I would be amiss to say there's not. There is. There is going to be. Every automation, right? Every automation. Like telephone operators don't exist anymore, right? I mean, and, you know, I don't think anyone's bemoaning that. So, yeah, continue. No, so I do think that that's going to happen. But there's an opportunity right now to figure out what does that mean for your area of expertise. And I think when you own control or narrative or input, you fundamentally can step into a leadership, you know, role into that area. So whether that's in operations or whether that's in creative, whether that's in storytelling, audiences, planning, or measurement and insight, there is a lot of, I would say, like, white space. Yeah, I agree. I also think certain jobs, I mean, most jobs, if not all jobs, will be impacted and will change, right? Many tasks will be automated using that. But what I think is interesting, too, like there's the example people give about radiologists, right, how there was a prediction that machine learning would completely wipe out radiologists, and there are actually more radiologists employed now. And part of that is there was a huge pent-up demand for radiologists that was just unmet. And then, okay, well, like this unlocked a lot more. It made them much more productive. So guess what? There's actually the unmet demand is now, you know, more being met than it was for. I think with certain things like engineering, like as we're finding, okay, a lot of – most coding is now done – I think more than half of coding is now done with AI, right? But there's – I think there's probably more – just as much if not more demand for engineers. I mean, I know there's debate about that, but I think there's a lot of pent-up demand for that as an output, like writing code, right? So my guess is there'll probably be more. I mean, we'll see. It may impact hiring maybe junior roles, but I think the jury's still out. We don't even know. Yeah, I'm not sure that we know. I think there's going to be shifts for sure, but how that pans out, I'm interested to see. Well, you're definitely in an exciting space. I am rooting for you, and I'll be keen to follow your updates and see how the CIOGNU CO progresses and all the news you have about new features, new additions, client wins. Brianna, thanks so much for coming and spending some time with us today. It was a fun discussion, and I look forward to our next one. Thank you so much, Rio. We're real long here for Signal & Noise, and this is a possible event in Miami, and we're in the beginning of day number one, and I've got a special guest, Pete Blackjaw, who is the founder and CEO of Brand Ranked AI. So great to have you here. Beautiful day in Miami, huh? What a great setup. It is a nice setup. I'm not going to go back. Can't blame you. So shout out to the Mad Connect crew for giving us this incredible look here on the rooftop. up. So Pete, excited to get you here to talk about your company, what you're doing. It's a hot space right now. I'd love to maybe, I think when we were talking before this, we were centering on trust and that being kind of a new currency that is incredibly, maybe increasingly important now. Maybe we can start the conversation now. I'd love your thoughts on it. I think the way to think about that is take a step back and just say, why are these cancer engines, some call them L-BOM. Why are they so popular to the tune of like 2.5 billion props a day? And I think they're seeing that cocky-sick growth because consumers are incredibly satisfied with what they're getting. And their responses are trusted. It's not that there aren't hallucinations or misinformation here and there, but generally that a lot of happy consumers that are getting fantastic results. So good that they're not always clicking. In fact, that's kind of creating a lot of turmoil in the industry. And so, you know, trust is everything in this environment. You've got these LLMs that are trying to continue to win the loyalty and the continued usage of these consumers. And, you know, that presents a lot of challenges for brands in terms of how do you show up in that type of environment and how do you show up credibly? I think what's different about answers versus search is that I think search, you could game it a lot more. You could manipulate it. Everyone tries to step ahead or two ahead of Google, right? They're always adjusting it. It kind of became a bit of a tragedy to comments. Everyone was trying to game the system. So you never really knew what was organic versus vague. And this is an environment where all of the answer engines, in fact, there's almost, I mean, bloody wars between Anthropic and Chat Chief Tea to almost see who can be the purest. who could be the least corrupt. We all saw the ads on Super Bowl. They built on a real insight. It's like consumers are like, what the hell is this? And so, yeah, so I think kind of preserving that trust is really critical. And then brands need to figure out how do you build credibility in that model? And it's not an overnight thing. They've got to do it over time. What? Like, how is Branding AI doing that? Can you maybe explain to the viewers? because it's a space that's nice and is changing quickly, and you're right. So fundamentally, this is my second trust startup. My first startup was called PlanetFeedback.com. We sold it to Nielsen, and it kind of did a similar thing where we metered social media, consumer feedback, and it's all about the consumer in power. Same model here where we are measuring what the answer engines are saying about brands. We're fundamentally measuring answer share, and we're looking at, I'd say our core metric is, our metrics roll up into a brand health and trust scorecard, which we believe and we know is kind of predictive to business outcomes. And we have three building blocks. Visibility, do you show up? Vulnerability, how do you show up? Because oftentimes you can be visible, but story is really negative. And so it just pushes the concern away. And then readiness, which is like, do you know how to market to algorithms and what we do, is we have about 80 clients, and we will test 200 trumps per day per answer option against those three different variables. And it's infinitely revealing the brand value. You know, you're standing, where you have weaknesses, where you have long-term issues to fix. And it works across the different LLMs. We measure six, seven, eight of them, and they all think differently. They're like different TikTok influencers. They all have their personalities, so to speak, right? College radical. They're very righteous. They stick it to the man. They attempt to default to ethics. If they think you're BSing on sustainability, they'll call you out. They'll find it to the Pentagon what to do, if I remember right. They ran into some trouble there, yeah. Sure. You know, they may be buying Super Bowl ads, but they're probably the last ones that are going to integrate advertising into their core model. And they can afford to do that because they have a great subscription. That's a big model, too, right. And BRAC is a little bit all over the map. They're hard to predict, but they're edgy for sure. ChatGBT is kind of straight down the middle. DeepSeek has a much broader distribution of Asian sources, which is really interesting because if you say what's the best skin care product, there's a very good chance you're going to get a product from Korea. Yeah, so that makes that really interesting. And then perplexity tends to, you know, they're very in the citations, and they're still trying to figure out their models. So to some extent, their responses have changed quite a bit. So, yeah, we keep track on them, and brands cannot afford not to know. Because here's what's critical. It's not just consumers that are looking at this. like every single retailer is setting up teams to analyze brands before they pitch their products. Every financial analyst, every journalist, every NGO, every determined detractor is using LLMs as a cheat sheet to learn about brands. So there's a lot of different dimensions. You've got your buyers that are now using it before they buy, and you've got these influencers that are kind of addicted to it, so they can sound smarter about the brands, but they're often digging up dirt. or digging up things that maybe aren't totally accurate. So, when you think about it, like people are using a web world of those things today. This is just a new way of accessing information in a curated way where it's an answer engine and it's giving it back to you so you can see what you'd want it to be. You'd want to know what these things are saying. Yeah, but it's different than search. So, search, you were competing for one of ten blue links, and you could then kind of make your choice about what you thought was most relevant. Here, you've got everything rolled up like a brand smoothie. They're kind of taking all these different elements and they're kind of blending it together into what is largely perceived as a single source of truth. And you said hallucinations. There are negatives out there, but generally they're doing a very good job of bringing credible, synthesized answers to consumers that are almost like good enough where they're like, I don't even need to click. And the hallucinations have minimized over time, too. I mean, the hallucinations have minimized. I mean, I think it's under 5%. And I think if you use them correctly, you'll verify things anyway. So I think that is a little overblown. And I think it's a controllable, too. Listen, brands, you know, one of the most important issues for brands to understand is readiness, which is like, do you know how to market the algorithms? And I would say, you know, if I looked at all the brands that are possible, I'd say, despite all the talk about AI, only about 20% of them are truly AI ready. Do they really know how to market the algorithms? And one of the very first things I do when I talk to a brand or a CMO, especially the ones that are waxy and poetic about AI, I go to their website and I type in a really obvious question like, you know, tell me about your product. How does your product work? What's in your product? Blank, blank, blank, blank. It's funny. Commerce slap in the face. And so brands today are 99% optimized for conversion, not for curiosity. And that is where the digital industry has evolved. And I felt a lot of pressure when I was at Nestle, but it's led to a certain, like, cultural blindness to marketers where we've become answer illiterate. So the world's biggest advertisers that spend the most money and talk the most about emotional bonding do not know how to answer questions on their own media, which is ironic because what we've learned across thousands and thousands of audits is that your brand websites are the number one influencer of your answer share. But things almost need to be re-architected. So we spent 25, 30 years architecting everything for conversions and for the funnel. But now, like, with AI and AI search, it's completely changing that, right? So when I first started, remember the early days? We all talked blissfully and innocently called it interactive marketing. And so our first instinct was like, oh, we're going to be able to have an exchange, signal in, signal out. Consumers are going to ask us questions. In fact, a lot of the early innovations when I created the first interactive marketing team at TNG, we had things like Pampers Parenting Institute, Ty, you know, Stain Detective. They're all very like ask a question, get a response. And then what happened is that we moved very aggressively to digital targeting at scale. And a lot of that worked. A lot of that was based on algorithms, but we kind of lost the high touch. And then as E-commerce came around, everything was just about optimized for conversion. And so that left a lot of collateral damage. Now you've got these LLMs. If you really think about it, step up. They have become the most consumer-centric vehicles in the history of marketing. They are elegant, friction-free, ad-free, multilingual, multimodal, empathetic. And this is just the beginning, right? And they're only going to get better. Yeah. And what I often tell SMO, it's like this is a consumer expectations recognition, not an AI recognition. And brands, if they really want to be content ready, they almost like think about like what are the consumer needs? Like what will a consumer want to know about my brand? And then build the content. And that's how you influence the algorithms. Most brands, when I say they're only 20% AI optimized, they haven't even started to build the book of truth. Or they're heavily relying on outside parties. But the good news about LLMs, they will give brands the benefit of the doubt, especially if anything is even directionally negative. crisis, recall, you know. It tries to give a balanced answer. If it finds something negative, it'll usually explain it, and I usually can't remember something. That's my experience. Where else would you go? Like, health is becoming one of the biggest uses. Like, yes, third parties like MyoClinic, those are critical, but, like, the original manufacturer of the drugs is probably going to have the best instructions on how to use them. What's doing them, right? And so, yeah, so if I were an agency, I would be starting a whole new business and, like, literally re-architecting brand websites, making them, you know, content liquidity. I would be thinking about the assets that have the content but don't get read by AI. So, for example, we work with a lot of clients. A lot of clients are trying to make sure that they get full sustainability scores. They put a lot of effort into it. They promise to street it all the time. And we'll analyze them and say, oh, my gosh, your scores are really low. And you know the reason why? They're just not submitting it to the engines, are they? Yes. This is like where the organization is like dysfunctional. You took the time to make the content. A lot of the cart-com groups are ready to check the box off compliance. We did it. They're not thinking about getting read by the LLMs. Then you've got the marketers that are almost trying to game the system. And so I don't even think you have to game the system. You just have to make sure that you're getting the credit you deserve. And that's kind of what we do. We'll do these audits and we'll say, you know, I think you're better than this. You should be getting better credit for product performance, being sustainable. And here's what you need to do. You need to overhaul your FAQs. You need to re-architect your website. You need to think more about concept liquidity, not PDFs. And then we'll sometimes see results in a matter of weeks. But it does require organizational agility. This is why this is so much bigger than the search team. a lot of the search teams have kind of taken the mantle of AEO. That's not enough. You need the CMO to orchestrate this because it's very – it's kind of a governance issue that we're getting into. Well, because you're thinking they may have the content already, but it needs to be maybe taken from a PDF and put onto a website, and they need to make sure that they have all the hooks into these LLMs and they have a strategy for connecting with them and updating them. You're right. I mean, it's more of a big C communication issue, right, then? And it's even more complex than that. So I'm giving a big speech to one of the largest health companies on Wednesday, and the title of the speech is R&D is the New Marketing. And you may say, Peter, are you on drugs? What are you talking about? The reality is that the LLMs, especially in the topic, but even in chat, they cut through the spin and the slogans to the substance. So the brands that market their science to the algorithms win. So they're so smart. They will literally be able to. That's interesting. That must be so challenging for, let's say, a global pharma because R&D runs completely separately, right, from the commercial order, which is sales and marketing. I mean, it's a common marketing style as insights and marketing. You're right. Let's think differently about this. That's a really good point. Take in your academic R&D articles. We have made them more liquid so they are more available, and then those companies have gotten higher product performance scores. Because what is happening? Like, the key is not just to be visible. It's to be recommended. What makes you recommended? You have to be the best. And you have to kind of fall into this definition of value. Like, one of the biggest risks to brands right now, especially in, like, the consumer goods area, if you guessed. Biggest risk in consumer goods. The biggest surprise is that. I'll just. I don't know. Go ahead. Private label is showing up like crazy. Because the algorithms are so good at figuring out, okay, this competitor to Tide is 90% of the performance at 40% of the price. So they're not going all the way to price. But the bots are so good at knowing that it's statistically almost – It'll suggest the alternative to you right away. Yeah. And, in fact, a lot of the manufacturers, like Kroger, for example, has private label product. They're actually better than a lot of manufacturers on green. So you've got a lot of consumers that are – The Costco ones are excellent, right? Yeah, they're like, I want good products, good value, but they've got to be sustainable. And so this is where life's going to get really, really tricky. And so what I've been telling brands, like, you have really got to get your act together in justifying your premium in the age of AI. That's why I said this isn't an SEO. Well, I think then I think advertising has to probably come in, too, at some point. Retail media, that was the issue, too, right? But obviously you can advertise in keywords. You can advertise your brand, right? Which was, I think, the way of telling us. advertising is not going to necessarily change your organic answer. That's the trick. Now, some will believe that it will, and we got all sorts of models that are emerging. Amazon's got one that's sponsored inside the prop. You know, ChatTDT has got one ostensibly around. Google's a little bit fuzzy. They might be the ones that kind of get away because they've gotten away with a pretty blended approach. But, again, in the world of consumer expectations, I'm not sure there will be a high tolerance for messing with all the lines. So they need to nail this. And if they don't, like, they're not going to be represented at all. But they might be represented in a way they don't want to be if they don't do it right. And they may need to reboot some things. Maybe, I don't know if you remember, like, when I was first at Nestle running digital, you know, I spent a lot of time on paid on Ernie Notley. And I'd say that's more important than ever before. Like, our own media is actually very strategic. You know, at the end of the day, when I was at Nestle, I just, like, left websites for dead. It was like, oh, yeah, you can't. You know, market to where the consumers are, and they're not going there. Now, you know, websites, customer service, customer experience is really, really strategic because that's what the LOMs were. This is fascinating. So you're getting less traffic because, you know, people are going to these answer engines, and they're not necessarily clicking through. but your point is that the content is more important than ever. Yeah, you may get less traffic, but you may get higher overall conversion because the consumers are coming in, are actually really, really convinced. And so, yeah, so it's a totally different topic. That's wild. It is wild. Yeah, it totally is. I just finished a book called The Answer Economy. It's all about this notion that the brands that really understand, the winners are going to be those that know how to answer questions. Forget AI. Forget all the buzzwords about AI. It's just like it's fundamentally old-fashioned, you know, marketing basics, like anticipate the needs, make sure you have content, and then really make sure that you're, you know, those boring basics like product security. Right. Those things matter because it's not like a P&G side-by-side. It's much more sophisticated now. So it's back to basics. This is the stuff like, you know, as a young marketer, this is where everything started, right? We spent decades doing this stuff, and then maybe we forgot about it, gotten some bad habits, right? It's funny. I launched, you know, I do a lot of AI-generated songs, probably to a fault, but I've now put together songbooks for my clients. Oh, with all AI-generated songs? It just lessens. Okay, got it, got it. Okay, sure. So I released my first album like three months ago, and the very first song is called Pre-Condaborate Basics. It was like a regular album. I think he was the last. I was checking out. My clients went nuts. But it is funny. brands are like consumers. They have the attention span of a flea, so you really got to spoon-feed them things that kind of emotionally resonate. Engineering music is wild. Joe Rogan kept playing that one. It was, I think it was a 50 Cent song, but it was someone singing it with SoulSinger singing it. He kept saying, this is my favorite song. It was really surprisingly good. I did, just a couple days ago, I did a rap battle between Possible and Khan. You got to send me that one. And in my book, I have at the end of every chapter, I invented what I call the takeaway track. So literally every chapter has like a recap in song. Hey, whatever it takes. Podcast, song. Yeah. Lots of new formats. Yeah. Well, thank you so much for coming here. This has been a great conversation. You're in town the whole week for Possible? Yeah, here through Wednesday, I got a panel today on AEO. Still looking forward to that. And we've got a bunch of clients that are here. So kind of, you know, schmooze. Yeah, soak up some sun. Yeah. I'm already feeling really dehydrated. Yeah. We have some water downstairs. And thanks again for making the trip out here. And enjoy this show. All right. So we have Patrick Duggan here. And a little impromptu session. Let's do it. Just because, you know, when we're sitting backstage, backstage, I mean, over there, and we're having a little chat about all things AI, a really fascinating topic randomly unearthed itself, which is hard in AI these days. Everybody's just AI and then what, right? Yeah, exactly. Well, this one actually had a point and I liked it. So I just wanted to dig into a little bit, which is you talk about AI monetization and how there's almost like a mad scramble for AI companies now to go, well, we've got people using it, Yeah, but that's not making us the money we need to pay back all of the debt or pay back what we need to make what we need to make. Yeah. Tell me what you told me. Unpack that in terms of the individual versus enterprise. Yeah. And just add a layer of that, right? Like brands are thinking about I want to give access to AI, but no way to kind of gatekeep or guardrail. Yeah. the usage and consumption, which is where the model is going from like a pricing perspective, right? So if I unleash CLAW to you and I and we really find a way to use it, we're driving a lot of consumption. Well, that's driving usage out of the rails, right? So a lot of figuring out the balance. What we were talking about a little before is I, as an individual, may be using CLAW, ChatGPT, Gemini, others for my own personal use, mixing in some professional use. How do I keep the guardrails and the usage around? I've got my, you know, my guarded brand content or my, you know, individual PI or whatever the case may be that I need to maintain. I got to think about enterprise level brands for this. Right. And the AI companies are thinking about the same way. Yeah, right. And I was talking about this to someone yesterday as well. And it seems to be the case that a lot of a lot of AI companies obviously build by the token. So it's all usage-based, which is a model that CFOs are not yet used to. And therefore, it's almost exacerbating. They'll get a build you at the end of the month. It's way bigger than what they expected. And that is exacerbating the need to, I guess, maybe reduce headcount even. Well, I think that's the great unknown about how do I make sure that the users of this understand the impact of what they're doing, right? Like, oh, I built out this really great deck and I want to make a PDF versus a PowerPoint, whatever. Yeah. And it's so far downstream. Like, I don't know the token consumption. I don't know the actual usage that's occurring. Meanwhile, bills are skyrocketing, right? Or like, hey, like, I'm just playing around with it. I'm going to make a Mavis Beacon game. Like, I played that when I was a kid. Like, oh, what I can make, like, HTML coding. I don't have to do it myself anymore. Meanwhile, like, the bill comes at the end of the month, right? And to your point, the CFO panics. Yeah. And I think everybody with the great question is, is this going to take jobs? Is this going to make people more efficient? What does it actually mean? And, you know, how companies like B2B companies or MarTech or AdTech companies are thinking about that's so different than how brands are thinking about that, how that drives usage or engagement for them or what their end goals are, right? Yeah. And so as I'm thinking about it, I strategy talking with companies, it's really around what are you actually trying to accomplish, which is the same challenge that's existed for 10 or 15 years as we've moved into this digital era. It's a new it's the same question, new problem in the sense of how do I do it with AI? Yeah. Well, what are you really trying to accomplish? What are the measurable outcomes that exist in this world of it? Like people say outside the box, there's not even a box anymore. I think it's so far past that. So the reimagining of what's real and what's not is even the start of like, how do I get some real outcomes? Yeah, yeah. And it's wild that, I mean, the en vogue LLM of the day, I guess, is, you know, a year ago you couldn't have convinced me that anyone other than ChatGPT would be, ChatGPT Enterprise would be obsessing, you know, the employees would be obsessing over it. Whereas now it feels like with flawed design, flawed artifacts, live artifacts, called Code, obviously. It feels like Anthropic is the darling of the day in terms of what you just talked about, which is people using it to do X, Y, or Z. Is that what you're seeing as well? You know, in my personal usage, that's certainly how I feel. And I think as you look at the business models, one of the things I really love about Anthropic is their goal of scaling is really around governance and kind of guided responsibility on the product. The way that they talk about themselves and the generation of their spinoff really, right, was around, hey, all of this AI usage is out of control. How do we do it in a way and how do we make ourselves be the beacon of governance? Because that's the other thing people aren't talking about. Like, yeah, there's the government lawsuits and all these sorts of things. But when governance comes on a federal level, it's like how we talk about third party data, right? Individual states creating their own privacy laws and all of that. Well, what's going to be the step for that from AI, right? And I think their vision of we want to stay in front of that and make the Googles and the open AIs of the world have to spend so much money to catch us on governance is a really smart strategy. Yeah. Yeah. And yeah, the other thing that I guess was really interesting to think about is, I guess, the incoming waves of salespeople and go-to-market people. And so where do you think they're going to, like, if you're going to hypothesize on, like, who all these new salespeople go after and where the battleground is, is it agencies? Is it direct brands? Yeah, I think it's a great question. And as a partnerships guy, partnerships in my blood, it's fascinating to me, right? Because you're talking about folks selling a thing right now that you don't really have to sell, right? People are buying left and right. They want to get engaged as quickly as possible. They want to get these tools in their hands. So a lot of folks are looking at this and just saying, I don't need partnerships. I don't need this engagement. I don't need these co-sell models in a lot of cases, right? Google has worked historically really well at resale. I think a lot of people are throwing a lot of stuff against the wall to see what works. But I think as you think about the large consulting firms, the agencies, like nobody's going to give this revenue up that's historically been theirs. How do they protect their mode in terms of engagement, helping brands along? That's not going to change with the AI world, right? And so I think there's going to be this flood of usage. There going to be this terror of I overspent on consumption How do I become efficient on consumption is kind of the next level And I think there a space very much so where agencies GSIs and even hyperscalers and technology partners are going to play a significant role. And it's going to be kind of an evolution shift, but it's going to feel familiar in a lot of ways. I think the speed and velocity in which it moves because of the world of AI is what's going to make it feel different because it's moving so much faster. Yeah, yeah. It'll give whiplash for sure, but I think the same patterns will play out. I believe so for now. And this is a whole new world, so we could all be wrong on this, but I feel like we've seen a little bit of this before. I've talked to folks who compare this to the dot-com days, right? And it's like, where do you make your bets, right? It was infrastructure then is where a lot of the bets happened during the dot-com days. Who were the winners of the day? Applications, right? And so I think that there's an interesting play to see how this plays out as we evolve. Like, yes, there's a lot of momentum right now with infrastructure, right? The investments that, even if we talk about Anthropic, but we just saw AWS announce more investment, right? Google's made a significant investment. Competitors are investing in the same product. Who wins in the day in terms of infrastructure versus application, where applications go as it relates to this is going to be crazy to watch over the next 12 months. Oh, absolutely, yeah. I sit there and look at all the analogies that get drawn as to, oh, AI is like electricity, and, you know, the electricity company meters you, but the people that run the businesses make the money. Right. And it's that kind of thinking. But then also you do see Claude building tools that, yes, like other businesses are going to get steamrolled by even Claude design, Claude artifacts. Totally. There are so many dashboard companies, Sony. Yeah. Yeah, there's so many businesses out there that, I mean, that is, so the electricity analogy, it made sense when I heard it, but it isn't actually happening like that. Yeah. So, yeah, it's an interesting look at. Well, and, like, we think about ISVs and the way that they're building their mouse, like a mousetrap that leverages all these tools. But it's really just a mousetrap, right? So if you build something spectacular, how long until that's either synchronized on a larger platform that grants it more accessibility, or how long until some of these companies leading the charge are like, ah, like, we can build that into the solution or the tool itself. and, like, great ideas immediately been monopolized, right? So, like, how you moat those sorts of things or, like, make your presence there, I think, is a real challenge for brands building really, really cool things leveraging AI. It's a really fascinating topic, and I'm sure we're going to come around to it. But thank you for tutoring me on this impromptu session, and we'll catch you next time. Yeah, sounds great. Listen, I always love a good improv, so greatly appreciate your time as always, and hope you have a great rest of the week. Thank you, mate. Cheers. Here's the noise beaming in from possible day one, Miami, 2026. I've got a very special guest, Nathan Thomas, the CEO of Thomas Media Consulting. Welcome. Thank you, guys. Glad to have you. He is an expert in supply chain, supply path optimization, and everything associated with it, right? Yeah. I think that's a good way to put it. The supply guy here. The supply guy. So I think, you know, on that topic of there's been a lot of talk about supply chain optimization, Yeah. Supply path optimization. Excuse me. We're not talking manufacturing here. And a lot of the black box challenge that happens within the programmatic ecosystem. Sure. Of whether it's, you know, Dr. Fu talking about take rates. Right. And publishers not, you know, sort of getting the short end of the stick when it comes to getting paid for monetizing their properties. Where do you think a lot of the where are we today? And like, have we fixed some of these problems? Has SBO lived up to the hype and to the, you know? Yeah, I mean, I think it's an interesting topic because both sides kind of approach it from such a different angle, right, the demand side versus the supply side. And me really dealing my whole professional career dealing with helping publishers monetize themselves, right? You know, sometimes you can feel a little bit at a disadvantage where you have the demand side, you know, mandating a lot of things, wanting a lot of data, wanting a lot of transparency, but then really on the flip side often not providing that, right? Yeah. And, hey, I want to see – As a feedback loop. As a feedback loop, right? Like, hey, you're buying me for this amount. I'm not getting log-level bid data back. Or why am I on a black list? Or, you know, I'm just making some easy examples here. But it really has to go both ways, right, on the transparency spectrum. And, look, I think over the last years a lot has gotten better. It's still a mess of, I would say, ad tech vendors and intermediaries. intermediaries and God knows what else is in the middle. So I think even more so than transparently understanding, okay, I'm buying this supply through this channel, which is obviously extremely important, is also, okay, what are all the steps involved in it? What are all the middlemen involved here, right? And how is that affecting my media dollar on both sides, right? Yeah, yeah. We had a conversation with Dave Rosner earlier today and Joe Lux last week from Permitive, and they're solving kind of this problem. And one of their big theses is that advertisers should focus, and Dr. Fu talks about this as well, should focus on going direct, right? Because publishers, one, aren't taking advantage of the amount of first-party publisher audience data. 100%. It's not being passed through to the DSPs, right? There's only a small segment of that data. And so what do you think about that topic in terms of publishers and brands sort of joining, you know, sort of that Venn diagram between publisher and brands with consumers is kind of the third piece. And cutting out the intermediaries. 100%. I mean, it's so relevant and so important. I mean, if we think about it really from a high level, what's the most direct supply path you could do? You can hand an I.O. to a publisher and they run your media. Zero hops, right? But, of course, that's typically not what's wanted. We want to buy through different channels at scale, and that's when all the mess kind of happens. Yeah, and should it get a complexity mess? I mean, you know, because if you're hiring an agency, right, they can do principle-based buying to give you volume discounts, which is kind of a reach play. Sure. But is it too much to ask your agency to say, hey, manage these 10 platforms or these 10 publishers? These are the places where 80% of my traffic is going to come from or my audience is going to be at. Is it too much to ask agencies to do that, or is that complexity argument real in terms of there's just so many places where you can buy direct? Yeah. We need to simplify the actual buying process. It goes a little bit to this kind of obsession with identifiers and audience play, right, where, hey, I'm looking for these specific identifiers, and the only really way to scale across that means you have to buy it really broad across a big swath of the industry. Yes. If it's going to one publisher, it may not work. I mean, that's where I think, you know, obviously this is a little bit of a different topic, but there's so many competing new protocols coming out. Yeah. And one area where I think I'm really fascinated with AdCP and those things, that those direct-to-publisher executions that are really kind of traditionally would be handled one-off with an I.O., that is – AdCP is actually very good for that, right? Where can agents kind of collapsing that chain and you executing with individual publishers at scale, where beforehand it was a lot of back-and-forth, and I could see why that's really not feasible on the agency side. administrative work behind it. It's got a lot of work to be done. It's not baked yet, but I think that's a good step there. And then to your point on the first-party data, I mean, a key topic is I do a lot of work on that with many publishers, right? They are sitting on the audience, right? The audience comes to them. On a treasure trove of data, much of which they're not leveraging in the advertising supply. Exactly. I mean, often it's just I'm the sending bid requester, right? And as you mentioned, with DSPs kind of throttling inventory, QPS, identifiers not being there. And they're doing that to control bandwidth. Yeah. It's toll-taking, right? Because it becomes too costly to bring in all of a publisher's data. Exactly. It would be more revealing from a contextual, behavioral way. It would be more revealing about who their audience really is and what their kind of propensities are. Yeah. Right? You think that that throttling, because you mentioned like ID matching, right? And they're pushing for sort of a reach. Yeah. Is that really an onboarding plan? We're talking about like you're onboarding a certain number of IDs, you're matching them against what exists within that DSP, which might only be a 50%. Yeah, exactly. Right? But are the days numbered when it comes to onboarding? I mean, like, are we in a better place from a publisher data control perspective to not even need onboarding anymore? Yeah, and I think that really depends what's supposed to happen with the data, right? If it has to leave your publisher wall, so to say, that it can go into someone else's system, and that system needs to have that data to make buying decisions, then yes, you have to obviously onboard it into that other system. Yeah, you've got to match with that. You have to match, and that gets back to kind of the ID obsession, right? I mean, every DSP and probably SSPs and others are primarily looking for their IDs, right? Like that's most important to them. Yes, it can transact on other things. But by doing so, you also leave out a massive part of the Internet where there are no IDs. Or, you know, those IDs that are looking to match are not present. So, you know, working deeper with the supply side to figuring out, okay, what kind of outcomes is the advertiser looking for and what does the publisher have? Yeah. And try to match it up as best as possible. I mean, that seems to be like a really fruitful connection. And a more direct data connection, right? And with sort of the growth of data lake houses, data warehouses, or data warehouse native applications. Clean rooms, you could argue, as well. And it does allow for kind of a pseudonymized match between data sources. But I think the challenge that I've seen is that advertisers and agencies just haven't had the level of access they need to that publisher data. So publishers aren't taking enough advantage of the rich data resource they have, which can more effectively drive better segmentation, better targeting, hopefully better outcomes because it's more aligned to just a much broader profile of an audience. Yeah, and I think the other thing to mention there is that, look, there's only so many publishers that have all this authenticated traffic, right? that Login uses. Is there going to be a ton of publishers that just don't have that? Yeah. So if that's all you're looking at, then, again, you're ignoring it. It's an authenticated premium publisher. But do you think that that's – well, let's get on that path. Do you think that the authenticated audience member – and Rio and I were talking about this with Joe in our last podcast – do you think that reflects really what media consumption is going to look like going forward, where more people are going to be basically interacting with less media channels. We looked at our own experiences, and we're like, you know, how many news publications do you read? How many TV publications or media companies do you subscribe to the Netflixes otherwise? And, you know, I could probably count the total, and it's less than 20. Yeah. So it's not a huge amount, right? And do you think that's reflective of overarching consumer behavior in general, where people tend to focus and authenticate more frequently and have a tighter relationship with certain publishers versus just defaulting to kind of the open, the long-tail web. Yeah. I mean, look, I think it's kind of a mixed answer, to be honest, right? I think what you said is totally true. I mean, if I think about myself, what websites do I go to and what apps do I use? I can probably count it on, like, one hand. Yeah. It's like it's the same ones all over. So all of a sudden, which kind of goes back to the supply path quality, If you're seeing a ton of impressions and volume on random apps and websites, is that real? Yeah. Because, I mean, if I'm trying to read CTV, isn't everybody on, I don't know, Netflix, Hulu? And then it kind of really safers off. Yeah, there's a few authenticated creative publishers, and then you've got all of this other inventory. They may not even be appearing on the television. Right, right. What I will say, though, is this, is that we can't forget the fact that there are kind of niche audiences. that may actually have a lot of scale sometimes that aren't just not on those traditional platforms. Yeah. And I think a lot about, so gaming has always been near and dear to my heart. Oh, that's, yeah, let's cover that topic. It's just a phenomenal area where, you know, you're kind of insulated from AI, right? You're not going to let AI play a game for you. Yeah. You're not going to defeat the purpose. So you're insulated from that traffic drop. You've got a captive audience. It's a human captive audience. Vastly underserved. I mean, based on the, you know, advertiser date, global ad spend is like 5% on gaming or something. Yeah. Whereas the attention on gaming is more than CTV for like Gen Z. And we talk about advertising. Can we break down the types of sort of theme advertising? I think of app loving and sort of this circular, let's advertise and promote other games within gaming. And that has historically been the biggest revenue driver on the app side. Yeah. Advertising other apps. Yeah. Yeah. Which is interesting because that made them be really good at that performance aspect. Yeah. Which, like, oh, this is outcome. Like, the app has been kind of doing that for a minute, right? Yeah. I want people to download or something. And can you replicate those outcomes when you're now dealing with brand advertising? Yeah. Whether it's a direct response to that, like your local auto dealership, or if it's a product placement. Yeah. Can a company like AppLovin duplicate the experience that they created with, you know, advertising apps and other games within the app experience? Can they take that model in transit? Have you seen any evidence that suggests that they are doing that? I mean, Apple haven't used their system. It seems like they're very successful. I think they're still probably early on and really branching out into non-endemic advertising. I'm not saying it's not happening, but I don't think it's at scale yet. But I think the data and the foundation they have are probably pretty good to do that. But then also the models that make the intent to convert or whatever it might be. Yeah, I'd just say on the gaming side, I mean, I used to game when I was young. I'm not so much in the weeds, but there are so many gaming environments that are just massive in terms of audience and attention, but greatly underserved and misunderstood. We're talking Fortnite, maps, Minecraft, re-herbers. That's a good part of my teenage boy's life. For, like, eighth grade until they developed social lives, right, non-online social lives. but it was an extraordinary amount of time that they spent within that playing these multiplayer games. And what's interesting is, you know, you kind of think that, okay, you're playing Fortnite, you bought Fortnite from Epic, they're running the ads, but no, there's all these external developers that make Fortnite apps, and in those maps you can run ads and do other brand integrations, and the attention there is insane, the amount of people going on there. Whereas Epic is fundamentally the publisher of the game. They allow these developers in to create content, essentially, right? Yeah. Same for Roblox. I mean, Roblox is kind of the platform, if you will. And then there's Roblox game developers. And those game developers inside their games can do certain things. And they attract a specific audience. And what's the reason for that? Was it a crowdsourcing to expand the – like to give the ability to deliver more advertising? Or was it just a lack of bandwidth and time and attention? I don't think that advertising was at the forefront of the idea of wanting to do this, but I think it's just kind of people, they wanted more content, they wanted more users, they wanted more engagement on their platform or their games that they built, and then these outside developers kind of came in and started doing that and really bolstered that ecosystem. I kind of claim as you want a good app store ecosystem. The only apps that are on there are from Apple. It's not very exciting. So you think that was a good decision that somebody like Epic Games made to allow this to happen? Or do you think that they gained and learned from what these developers were building, like, in their ecosystem, so they can then charge forward with their own advertising? Yeah, I mean, there's definitely been some of that, right? I mean, for example, let's take Roblox. I mean, they've been, over the years, definitely playing with their kind of monetization strategies and what they allow and what they know. and I know there was certain companies that did like years ago some preliminary advertising in Roblox and then they got shut down because Roblox started doing it themselves yeah right so they certain businesses just went out of business right because yeah we took it over right it's kind of like kind of an odd thing where you have I don't know let's other advertisers advertise on Facebook like right why would Facebook do that in the long run but they certainly need that content and that community so I don't really see that going away it's going to change yeah it probably is tied, I mean, just supposition would be that it's tied to kind of the community building and that they don't want to threaten the sanctity of that community building. Correct. It's very important. Yeah. And so what do you think the future of sort of in-game advertising, we've been talking about this for a long time, you're saying it doesn't, hasn't really increased above, what's the percentage? Five percent. Five percent of total media spending, maybe. You know, understanding the amount of time spent, which is a pretty important metric that certain age groups spend in-game certain demographics, right? probably largely male, largely teenies. Yeah, I mean, I don't even know if that's necessarily true anymore, depending on the game. I think that's, I mean, I don't want to out myself on my age, but we're kind of a little older here. It's only the boys that play games. I don't think that applies anymore. Like, you really have a niche. A pretty wide spectrum. It's very wide spectrum. So what do Epic Games and all these other players have to do to really develop? And in a sense, it's their own role. It's within their own gaming ecosystem and advertising model that can scale and grow and and keep consumers, keep gamers coming back because they're not, you know, is it product placement? Is it? There is, I think there's a few things here, and I'm actually working on some personal projects in this area to hopefully solve some of these problems with some other colleagues. Yeah. But I do think it's a fragmentation on both the inventory, the formats, and the data. Yeah. That is kind of hindered. And lastly, and this is actually an important one, kind of just a misunderstanding on the brand and agency side what gaming is right because if if you as an agency get let's say we have video budget right okay where does gaming fit in yeah is video really it's digital media but in game has video placements yeah is that video is that in experimental what is the form factor of this video is it like a spot is it ctv because he's landed on the Xbox and there's a video in it, you know? So I don't think they quite understand that yet. Sometimes, you know, you have these oh, this is my social budget. This is my and it doesn't, where do I fit it? Right? So I think that's a big one. But then, going back on the other side, you know, again, for example, identity, you know, while I still believe not everything should be tied to identifiers and hashed emails and whatnot, you know, there is the need on the broader gaming ecosystem to kind of coalesce, right, where you can then take someone that plays Roblox, and you know who that is when they log into Discord, and you know who that person is. There's a total disconnect. Right, there's a little disconnect, and then the formats, right? Yeah, because you're not going, like, cookies are following you around browser environments. That's not there. That's not there. You might be in a desktop client, or you might be on a console, or whatever it may be, right? And then the formats, right? Like, okay, I did a dynamic game. I don't know. I put a – A billboard. I put a Wu-Tang Clan Nike-branded store in a Fortnite map, and people are interacting. What is that on an impression level? How do you translate that to this is this CPM over here when I won banners, right? Yeah, exactly. And those are things that, you know, I think we want to solve. There needs to be a fundamental standardization and then a cross-platform, which I want to say. Cross-platform. That's the board of play. Some sort of identifier that can persistently follow players that, even if it's not within the same game, you can just get Epic Games game, right? It could be another type of game on another platform. Yeah, it reminds me a little bit of OpenAP on the kind of television audience side where competitors essentially came together, right? Yeah. did something so definitely working on that personally on some projects and then yeah i'm excited for that that's awesome yeah we haven't talked to a ton of people about gaming i think we had an earlier interview but but you mentioned the wu-tang clan yeah and you realized that you just planted seed in my heart it was actually not planned or delivered at all that little did people know that you had a hip-hop past i did yeah tell us about that in my old days in switzerland Yeah, first I started on actually writing graffiti, and then I had to lay low for a minute after certain things happened. Wow. And then I got into more rap, and I was assigned to a label in Switzerland. There's YouTube videos. Wow. DM me for the link. Yeah. It's on Spotify. You know, it's all good. And so what marked tech industry? Yeah, so I was always extremely interested in advertising and marketing. And when I was going to school, I always knew I wanted to end up somewhere here, but I didn't know anything about ad tech. I was teaching media economics at the University of Zurich before I moved. Is this where you did the graffiti? Was it the streets of Zurich? Streets of Basel. Streets of Basel, yeah. So then I moved to Florida right around when I turned 30, and I started at what was then called Intergy, now Playwire, one of the premier publisher-side monetization firms, and then was there for 16 years and kind of built a lot of the ad ops, yield ops data practices over there. And now I'm solo as a consultant doing my own thing and talking to guys like you. Nice, nice. Well, thanks for joining us. This has been wonderful. We have quite a windy backdrop with Signal & Noise here. A nice one, though. But it is nice. We're on the roof on a little island right outside of Miami. And it's been a pleasure talking to you. Thanks for joining us today. And we'll have you on the pod for a full length. I'd love that. Anytime. All right, man. Thank you, Brad. Thank you, Rio. Hi, this is Rio Longo here with Signal & Noise podcast. And we are day one if possible. It's a beautiful day here with the Mad Connect Mansion, which is just off Miami Beach, and we're kicking off the first day. So we have a special guest joining us at Kickoff Possible, Nathan Lindbergh from VP of Global Brand Partnerships, Silver Wolf Ads. Thank you for coming. Wow, what a great time to be here. I mean, this is absolutely gorgeous, beautiful day in Miami. I mean, what else can we ask for? I think they get a lot of beautiful days, but I'm jealous. I'm jealous, especially living up in the Northeast. It's a very different picture down here in January. It's been a tough year for you guys. Don't need the snow coming to me. It's funny. I flew in from Denver. In Denver, it was 43 degrees and, like, cold rain. Yeah, you guys are very generous this year, passing all your snow on to us. I really appreciate that. We didn't get much. Really glad you could make it here. We're really excited to hear about you, about your company. I think it's an area that people don't realize how big it is and how important it is. So you can start off by telling listeners a little bit about your background. and love to understand the business and what makes it tick and the revenue model and all that. Yeah, I mean, I come from video games, so I've been selling video game advertising for over 20 years now. I started in print. I moved into digital and web. I worked at Amazon for a little while on the Twitch brand. I spent some time at Epic Games. I've been able to see kind of the burgeoning, what I would call, creator economy that exists within the video game space. It's a new area for video games, but as we've seen and we've talked about many times in your podcast, the creator economy is happening in every industry. It's happening everywhere, whether it's YouTube creators, whether it's Spotify artists. Like, everyone is finding ways to create their own content, and it's because there's just an insatiable appetite, right? The access to data, the access to information is so much faster than it was 10, 15, 20 years ago, right? Like, some of us still remember what that dial-up tone sounded like. Some of us remember that from the late 1900s. And now, just information is so fast and the audience, their appetite for content is insatiable. Industries can't keep up. And we've seen that transition happen with film and TV and now YouTube and Twitch. And now we're seeing it kind of move into video games. And that's where Overwolf is really trying to position itself as a platform where tools and apps and mods and the things that gamers want can be created by individual content creators to help them then monetize and have a more enjoyable gaming experience. So do you work with, like, gaming platforms that then have creators who are publishing games there, or do you work, like, just with creators themselves? How does that work? Yeah, we sit kind of in the middle. So our job is to kind of create a tools and services platform that's powered by advertising. and that tools and services platform allows you, if you're a really excited player of, let's say, League of Legends, you want to create an app or a website that helps people play League of Legends a little bit better, you can go and create that using Overwolf's suite of tools and services, and we have a monetization platform set up to help you then monetize that. And you get 50% to 70% of the ad revenue that comes into your platform. And so it creates a really strong value exchange for the consumer. It helps the game publishers because we obviously vet and moderate everything. Sure. And at the end of the day, people are having a more enjoyable gaming experience and more content is being created for the games they love. It's kind of a win-win-win for everybody. 50-70%, that's a good share. As someone who's worked at other platforms and services and see what the rev shares are for those. What are they, for example? I mean, I've certainly seen platforms where it's sub-20%. And so if you think about that in relation, yeah, It's a really great deal for creators, especially knowing that the advertising that sits in those ecosystems and services that you're using, which are free, having that ability to say, hey, like, this is wonderful. And the fact that I'm getting a tool and a service that I'm enjoying is for free. And then the person that has made this is getting a 50 to 70 percent rev share. That's awesome. Has AI changed that at all? Are there more creators now because of it? I'm imagining that's a case, correct? I say we see the efficiencies in the AI angle. Obviously, you're not able to create these apps and tools and services quite yet with AI technology, but certainly we're seeing people innovate and certainly optimize those experiences. I think that when you look at how AI has affected other platforms like social and stuff like that and the AI slop that you hear from people, a lot of that, gaming seems to be a bit of a strong moat around that in terms of preventing those opportunities. We're not seeing just the lawlessness of content creation. We're seeing innovation. We're seeing efficiency. But we're not seeing just the lawlessness of IP flowing everywhere and stuff like that. And we think that that's, you know, for brands especially who are concerned about that stuff, it's a place where you can kind of feel a little bit more comfortable about where your ads are running. I think many people don't appreciate how big of a business this is. Like, are there numbers you can give us in terms of number of creators, what types of revenue numbers some of these big creators and publishers are seeing? Yeah, we've set some aspirational goals for ourselves. Our CEO has this focus of trying to pay out a billion dollars to creators by 2030. So we have a strong focus in the North Star, and we've hit about $300 million last year in payouts to creators. Some of our creators are getting beer money for the weekend. Nothing wrong with beer money. Nothing wrong with it. Others are seeing generational wealth. Some of our app developers have sold their apps and their businesses for over 40 million euros. So we are seeing those opportunities. We are seeing this new creative economy that's happening. And it's happening in a different way than you would see on a YouTube or a Twitch or other kind of what I would call entertainment-based platforms, right? These are people who are creating the tools and services that you enjoy versus someone who gets on a microphone and talks and entertains that way. And I use the analogy of if you're in a coffee shop, and George Clooney walks in, you know George Clooney, you know who he looks like, you know who he is, right? If, you know, Banksy walks in, very famous artist, right? You know his work. But if you walk in. So recently no one even knew who he was, right? So that's the beauty of, I think, what our creator economy is. It's the tools and services, people. It's those behind the scenes that are making great works. It's a very different take on creator economy. But it's nonetheless valuable and it's nonetheless lucrative. I mean, these creators, as I said, are making anywhere from some beer money for the weekend all the way through to fully generational wealth. Where do the creators come from? Are these just normal people who are interested in it? Are these people who, like you have been in the industry for a while? What's your take on that? All walks of life. One of my favorites is a gentleman named Sandy. Sandy is a total girl dad. His daughter has epilepsy, has a hard time playing video games. He ended up making some mods that helped and allowed her to play in a much more tame environment. He used our platform to then sell them to other parents and families who also could use that. And now he has a business where he can stay home with his family, and he works in a small little studio, and he makes anywhere from $10,000 to $25,000 a month. Not a bad business. And he would tell you, first and foremost, that that was never in his, like, mindset coming out of school. You also see, you know, lifelong gamers who want to, you know, provide that value as well. So it's a really mixed batch. We have over 187,000 creators across the world who are making apps, tools, and services. And that, I think, really, it's a very welcoming and kind of opportunistic area right now that we see more and more people coming. I think that AI makes it a lot easier then to create and to kind of maintain those situations, which makes it even more, I think, approachable from that perspective. How can you respond to people who might question, like, why do you need ads in these games? or like what's the whole point to people who maybe are not big on the ads things? Like what's the value proposition, both from a consumer point of view? Obviously, from a creator point of view, the monetization, I mean, it makes so much sense. So guys, people like the guy you subscribe can work from home, publish games, and get checks every month. But from a, let's say, a user perspective, it's the argument for ads. Well, I would say that the question is do you want to pay for it or do you want an advertiser to pay for it? And what we have seen is that when you have an ad environment that is non-intrusive, that is complementary to the experience, and it's giving value both to the creator who makes it and to the user who's using it, people are fine with it. Gamers in general, I hear this a lot from brands and agencies, oh, gamers don't like advertising. Gamers are people. They don't like bad advertising or irrelevant advertising. So if you create a situation where I'm doing something and you interrupt me and say, hey, pay attention to this, I feel like that's already a negative situation that you're creating through that advertising medium. So if you can figure out the right way to put the right message in front of the right audience, we've seen brand lip studies. We've seen a ton of quantitative results that show that our audience really appreciates the fact that they don't have to be nickel and dimed for services, and instead those advertising dollars go directly into the pockets of people who make their favorite tools and services. So it's a value exchange. I think it's a good point. I think people underappreciate just how many people use these games. Let's say you go on a bus, right? You see all these people playing their phones. Many of them are playing games, and most of those games, to your point, are ad-supported, and there's no way everyone's going to either once or once or once going to be able to pay for these. So the ad-supported model I think makes a ton of sense. So your role as a VP of Global Brand Partnerships, can you maybe explain what it entails? And, yeah, it's finding the right brands and advertisers to be in this space and to partner not only on the day-to-day media transactions we do, but create bigger and more kind of altruistic partnerships across the platform. So not just running ads in our IAB standard media areas on our website, but actually creating long-term campaigns where users can play and win and earn. We have a number of different ways that we help brands really get into the space and actually work with users from an engagement perspective. So imagine being able to play your favorite game, do a certain number of things, win a prize from a brand. That's a great experience for you. It's a great experience for the brand. And so our team is really focused on not only here in the U.S., but globally trying to find the right brands and agencies and partners to come in and play a bigger role in that ecosystem and help to drive more premium revenue to our creators. What types of brands are advertising and what sectors are you getting the most interest from right now? It's great. We see a lot of different partners. We just wrapped up a campaign with the folks in our block, which you would think, okay, taxes, gamers. But actually, one of the largest small business groups out there are content creators and creators building. Oh, that's interesting. Okay. So we actually saw we had a great campaign. We worked with them on promoting their different activations and stuff like that. And they came back to us and said that not only was it the best gaming partnership that they've done, but it was one of the best partnerships, period, that they did, which for me is really important because we want to get outside the idea of, oh, we're just good for gamers. We're great for brands across the board. We've seen some really great opportunities with travel and tourism, Universal Orlando Resorts we've worked with recently. USRs are always a favorite. There's McDonald's or Little Caesars. So we see a pretty good variety across the board in terms of what brands and agencies want to work with us. Entertainment's popular. And then obviously endemic game publishers are very curious about kind of how we can help them connect with their future audiences and grow their overall target there. So your payouts, the goal is to be at a billion, you said, by 2030? Yeah. Those are aggressive goals. And then what's, in order to pay out a billion, how big, if you're, just doing the math here, so a billion dollars and a few more paying out, let's say, 50% to 70%, right? So that's a pretty big ad number, right? It is. It is. And it's not just ads that help support our business. So we do have payments and other verticals in our business. So there are other areas that we're also generating revenue for. So across the board, we're trying to kind of build this kind of payout system. So whether it's our ads business, our tax kind of payments business, there's a whole suite of tools and services that we're trying to help to get to that billion-dollar number. So it's not just ads because that would make us a pretty big player in the ads business. But we're on our way, and I think that we're, you know, from an ad perspective, being able to put brands in an environment like video games is really valuable. And we're hearing that especially from kind of research companies like Adelaide. You talk about attention and the value of attention scores. and video games are very, I mean, if you have a child or a significant other that plays games, it takes many, many times to get their attention, right, because they're locked in and focused. And advertisers can't find that everywhere. And so, you know, this is an opportunity where if we can get it right with the correct kind of partnership between the ad and the consumer so that it's a pleasant experience for both sides and we can track it and we can put data on it and we can do those other things, I think that advertisers are really excited because they want to be in that environment, right? If you say to somebody, do you want to be in an environment where your audience is highly engaged and is going to see your ad and is going to have a much higher level of retention, everyone's going to say yes. You can tell them an awful lot about these audiences as well. Yeah, and it's a big thing that we're talking about this week at possible. This is the first time we've launched an ad product at an event, So we're excited to talk about GamerGrid, which is our new kind of first-party data management platform. We're combining it with ArcSpan and LiveRamp data to give advertisers a much better understanding of their audience in this high-attentioned environment. And I think that when we change that narrative, it ends up becoming a much more easy thing for someone to consume and for someone to buy. Because too often we're being told, you have to target gamers. That's a large group of people now. As you mentioned, everybody from people playing on mobile games, PC games, consoles, switches, a lot of people playing games these days. And to say one holistic group needs to be reached, it's not the case. And they can take their audience and enrich it with, let's say, on board and get a ramp ID, then match it up against yours, see the overlap and target audience specifically? Exactly. And they can buy it direct. They can buy it through programmatic direct. And this is a way for them to find their audience in our high-attention environment, which is video games. And I think that that's a much better way for this industry to work. I think that as someone who's spent so much time in video games, we have never made it easy for advertisers to buy our platform and our service. And it's the reason why we look at games like Grand Theft Auto 6, which are going to release later this year, largest entertainment franchise launch ever. Oh, it's a multi-billion dollar franchise, right? I mean, I don't think people appreciate these games launch, and they will do tens of millions of dollars of sales in the first few days. Yeah, and we still don't perform movies. And we're still on pace with podcasts in terms of revenue. And you look at TV and CTV, they're light years ahead of gaming. And we have to fix that. And that is a big reason why making it easier to transact through the kind of ads and situations that Overwolf is building, IAB standard, trackable, first-party, third-party data layering, that's going to help the ad industry go, I'm comfortable with this. This I can actually transact for real. Well, once you have to fix this, maybe you can help fix podcasts next. Audio is totally under-monetized as well. It is. But I think it's a lot of problems in terms of fragmentation and lack of understanding, even comparing metrics across audio platforms is super challenging. I don't think gaming is quite as – I mean, gaming has a lot too. You look at game publishers all operate on their own infrastructures, and so having that consistent synergy is really difficult. It's the power of Overwolf to be across 1,500 games as opposed to, you know, a platform like a Roblox or a Fortnite, which is, you know, vertically integrated, but they don't talk to each other. And so to have a platform like Overwolf that sits over 1,500 games, you can have a much different strategy than if you're just working with one game publisher on one activation. Well, really exciting. I'm going to be looking out for those announcements here if possible. Love the pants, by the way. Oh, fantastic. So trying to keep that vibe, you know, you've got to make that Florida vibe work. Yeah, if we're coming to Florida, why not, right? Especially, you know, perfect weather like this and beautiful sunshine. So, well, good luck to your possible. Enjoy the next few days. Appreciate you coming by. And maybe we should check back in in a few months and do one of our feature long-form pods. I think that would be a lot of fun. Excellent. Good to see you. Thanks. Thanks so much, yeah. All right. So we are at Signal and Noise, day one in the mansion on Venetian Way. It's a beautiful place here. And we have Dave Rosno, who's the CMO of Permutive. That's right. Thank you very much for joining us. Great to be here. Yeah. And we actually hosted, well, Rio and Brett, the hosts of Signal and Noise, actually hosted Joe, who's from Permutive. I don't think the podcast has been released yet, so maybe we'll give some sneak previews along the way. Everything you want to know about Permutive in two podcasts, right there with Signal and Noise. Epic. Right. So you just joined, actually. You've not been at Permutiv in, what, three months? I'm about 10 weeks in so far. I'm loving it, and I still have enough of my old perspective to bring to the conversation. What was your old perspective out of interest? Well, you know, I've been on the buy side for a long time, so I'm pretty sensitive to what people are saying in the market and how to screen through it. And then in my previous role we have played a big role in popularizing curation So I hope we get to talk about that and you know where it was and where it is Absolutely. Yeah. Yeah. So the buy side has given you a really interesting lens. Obviously, I'm sure a big part of what attracted you to Permit to Permit to you was that buyer lens. What is your mission as you come in, your 10-week team? What have they given you and said, like, this is what the nut we want you to crack? You know, I think the most important, well, there's two things. There's what's important and there's what I love to do. Right. So what's important is how do you differentiate in the marketplace and how do you say it in a way that makes sense for the buy side, right? There is a lot of technology in that tech and mostly the buy side should not have to care about that. What they should care about is I can do something for my agency, for my brand, for my client better, more effective than I did yesterday. or make it easier. So I don't really care about your tech. I don't care if you were vibe coding the last five years in a basement and spent $100 million or if you've got a genius with an MBA in a data science, an MBA, a PhD in data science, and they've just come up with something yesterday. If it makes my life better on the buy side, that should be the story. So that's story number one. But how do we take Permutiv, which is the leader in enterprise publishing data support, and DMP. How do we take that and share what we've done for the buy side? So that's the job. You're not saying that sell side tech overcooks it, are you? With the teching... No, I think it's just like my sensitivity is having been in meetings, we're successful. This goes for everybody. This goes for branded trays at stadiums. Yeah. Sorry if you're watching. It was a good meeting and I love you. But like that, I have a very clear recollection. Like, you're not solving the world. Tell me how you helped me and put it in the right picture. The thing I like personally, so that's what I think is so important, you know, for Permuted and like why it's so important to explain what we're doing in a way that makes sense for the advertisers. What I love, however, and what really got me to be so excited to join is when you meet a company and you realize that there is this enormous amount of information and something both scaled and new that the world doesn't know about. Like, it's like you're a diver and you found an underwater city and you're like, nobody knows about this. Like, this is the kind of thing I want to make noise in the market about. And that's what's happened to Permit. This decade of working with publishers and creating data that the buy side is unaware of the depth there. Right. Okay. So this is interesting. Give me a little teaser, I guess, as to what some of those things that have kind of gone unearthed or those gems. that you've come in and gone, like, well, we should be definitely talking more about this. Absolutely. You know, to cross over to a little bit of what we were going to talk about anyway. So, you know, there's this whole notion of first-party data signals, right? Is that important in the market? Is it overhyped? The truth is that in a world that is increasingly working with the support of AI, so AI augmented, Real truth at the key is like the most important thing. Yeah. I think we all learned some new vernacular in the last two or three years is all this money has gone into AI and LLMs about training data. Right. Like the whole world for some reason now knows everything about training data. If you have an AI, let's say, augmented system, which is where I think we are right now in our journey as an industry. If you have an AI augmented system and you don't have the best source of truth, which is actually what consumers are doing, then the whole thing falls apart. So what Permutiv has done, which is very different than what I've seen in other places, is create a much more robust first-party data set on behalf of publishers. And so instead of the six key areas that typically are used to transact for programmatic, which is kind of what we were using in the rest of the digital ecosystem, it's magnitudes greater. It is thousands and thousands and thousands of pieces of information that are being tracked over time that help publishers be better publishers. but that entire much more robust look at the consumer has not been pulled over to the buy side. So this notion that you had like six elements versus this high definition picture of your audience, so you can actually buy your specific audience, better approximate the brief, drive better performance because you are buying the right people, like all of that is going to be net new and I think something that we're going to be hearing a lot about. Okay. So that's the underwater city. Yeah, this is really fascinating. Do you feel like that is currently going on a tap that the Biceh just doesn't know about those publisher-direct audiences? Well, I hope it's not just me. Yeah. I hope I'm not the only one. But, yeah, this notion of a key value pair and this notion of how much information is possible to grab on behalf that publishers have been doing. And maybe this is special because of Permutiv. I'm still new because we've been a SaaS technology with a ton of funding that's been doing this for a decade, and we've just been going and going and going. So we're already supporting the majority of enterprise publishers around the world. But I don't think BuySide is aware that there is this treasure trove of first-party data that will better allow them to say, this is who I want in market, let me buy that specific person, and then optimize based on what you're telling me are the signals back. And what do you think that's a root cause of? Is that the fact that they've been used to buying from the parties, or is it that they've been used to buying contextual? What do you think buyers do that just causes them to have that blind spot? Oh, this is not a buyer's thing. This just hasn't been available on the buy style. So this is a company that has traditionally been working with publishers, and publishers needed this, by the way. We have accrued 182,000 segments. There's no reason to have a first-party deterministic data set like that, unless it's in the service of publishers meeting advertiser needs. So this is just something that's new to the marketplace. Sometimes something new comes up. We've seen it. it creates a step change in what's possible in the digital ecosystem. And so I just think we've just started talking to the buy side in about the last year or so. And so this is just one of those things where we're going to be sharing it and figuring out the solutions and it'll be rolling out. Fantastic. Okay. This is really interesting. And does this segue into curation for you? If you want it to. Yeah. It really touches every part of the digital ecosystem. Right. So let's talk a little bit about duration, what you believe it has been up until now. Okay. What is it destined to be on? What potential has it got? Because I have heard lots of different lenses on it. It's the new ad network. It's a new way of doing this. It takes margin. There's lots of different lenses that you hear online. Sure. So help me break it down. Let me define it like this. I've had this conversation a lot because in my former role, we did a lot to popularize curation. Yeah. So I think it's really important to keep in mind that there's two kinds of curation. There's regular curation and good curation. Good curation is how it started. Yeah. Right? Which is the offering should be you have a unique data set. You're combining that with inventory into multi-pub PMPs. Yeah. You're activating that on the supply side. So part one, unique data set. Part two, combining it with Multipub, taking that data, combining it with Multipub, activating it in PMPs on the supply side. Three, you're optimizing it in real time. This is a huge upgrade for the whole system because you're taking the data, which is your source, your insight, and you're taking it out of a 30 to 45-day review cycle and bringing it into real time. That's just an upgrade for the whole system. That kind of jump ahead. However, the word curation is the regular word, and in some cases it means added network, like we're curating inventory. By the way, good idea to curate inventory. Yeah. But when you talk about why curation became a trend, it's those three things together that's the good version that I think is so important. Okay. I like that. I also heard different lenses on it as well, and one of them that stuck out to me was the idea of a chef in a kitchen, knowing what they're doing, being a well-trained chef, and having access to all of the ingredients in one place, as opposed to what we had been doing for a long time, which is different parts of the ecosystem having different ingredients, like the SSB having inventory, the DSB having data, Yeah. And never the two short meet until the bid happens. Whereas with curation, it was, you know, it was interesting to me to think about the fact that you could prepackage and predict stuff better. Yeah, the chef metaphor, I love, first off, I love a good metaphor. I mean, you've never met anybody who likes good metaphor. If I look down at my notes, they're full of metaphors. I worry sometimes that a chef metaphor might be more like the inventory, and I hear so many people talking about curating the inventory. What you really want is that unique first-party data set activated on the supply side to get the real time. Bringing real time into data is a game changer. One of the things that we've been working on is this organization called Atria that we support. So this is the number of publishers in Europe who said, we want to get together and create our own curation solution. And there have been a lot of discussion in curation. Is this net positive for publishers? A lot of the SSTD shared data that showed that it was incremental value, that curation was actually driving incremental revenue and incremental value for publishers. But I think I love the Atria example because, one, it shows that publishers are leaning into it, and there's no better signal that curation can be positive for pubs than pubs saying, we want to take our destiny into our own hands, and this is the route. And then, two, I love it because it's a permutative tech underneath it powering it. Right. Okay. So let's rewind it a little bit because I want to unpack what's something really important you said, publishers taking their destiny into their own hands. Sure. So help explain what Atria is to me as if I'm a complete noob and haven't got any idea what it is. Sure. So Atria is a group of six publishers in Europe that have gotten together. They're all from your clients and said, we want to create our own closed curation solution so we can bring more scale across our audiences, but take advantage of all the things that good curation allows for. So on our pipes, on permitted infrastructure, we've helped them put that together and support it. And it just launched, and it's really off to a great start. Fantastic. And the mission at hand is to help publishers take control. Yeah, and I think it's also about scale, you know. So, you know, publishers are in this situation where the buy side is, their expectations are being set by companies with spectacular technology at the walled gardens, you know, And they have scale. So certainly the first-party data solution that publishers have is much more robust, and the buy side has more opportunity to capitalize on that. They have a lot more control when they're working with a publisher. So how do those publishers work together to bring together more scale to make that equation more attractive for advertisers and agencies? So that, I think, is the thinking behind it. Love that. Okay. And let's now switch our attention to the most buzzwordly buzzword of them all, which is AI, of course. Let's talk about how that – how do you feel that intersects with what Permitiv's mission is? Okay. And obviously that it poses an existential threat to the publishing industry. Okay. And what does – so what does your narrative change to be when it comes to a publisher? You know, we did most anti-ad tech thing ever when it comes to AI affirmative, which was we built it and we didn't say anything. Where normally you tell the entire industry what you're thinking and planning. Yeah. And then you go ahead and build it. So we've already built agents into our technology. They're already supporting hundreds of millions of dollars of publisher revenue. And what we've seen is it's just augmenting the entire process. So I'll give you a couple of examples. One is the amount of time it takes to respond to an RFP can go from days to literally 15 minutes. And you know if you're a publisher selling in, that is the game changer. It means you can take on a lot more business. It means you can pitch more. But most importantly, you can work in a whole other pacing and see what the outcomes are. The other thing that is amazing is it helps you better unlock your first-party data. So there's a premium publisher group that we work with in Europe that, because they have so much more insight and can work so much more quickly with the data, they're able to win clients' brands that they didn't think they could even compete with, because they could say, oh, actually, you would never come to us, but our audience for what you're looking for is bigger than the classic leaders in the category. So it's allowing them to increase the TAM. And if, you know, you're on the revenue side and you say, hey, our DMP and their agentic agents are helping us increase our TAM, not just win our share of the pie, that's a pretty good day. Yeah. Yeah. Is that just on merit of being fresher and being sort of quicker to the table on pitches? It's the ability to number crunch in a way that no one could before. And I think, and we'll see if when you have your interview with Joe, if it lines up, there's this weight that comes with our history as a SaaS company. But not that it was built with this in mind. But when you're a SaaS company, you're that much closer to being able to add AI into the mix because everything has to work automatically anyway. Where that might be more difficult as a managed service company. Yeah. So I think we saw this immediate opportunity to make our processes better, what our solutions were for clients better. And it has turned into, with AI, this just upgrade. So it used to be if you wanted to figure out how to meet an RFP, you might want to talk to the person who knew your segmentation or your taxonomy the best and, like, really work through it. Now the agents are great at that. It's like having your best librarian who's been at the company forever and knows all of your thousands of segments and can pull that together really quickly. It's like it's having the person you always want responding to the RFP always on at your fingertips. I love it. And it's, you know, given, like, marrying that with, something else important you said, which was real-time data, which I also think as an industry we've spent a long time not having access to real-time data, not saying that. I think that's been a real thing in our industry where we talk about data, but only recently is it becoming really, really, really fresh. And so those things combined, shout out that things are becoming so much more efficient. That's exactly right. And the thing that is going to continuously change is we're going to realize what we can cut down lag in a lot of places. Because, let's face it, if you are talking to the buy side and the original upgrade was taking the recency of the data or the optimization from like 30 to 45 days in real time, that's huge. Now we're going to keep finding, as we have better technology, there are all these little spots that can unlock more opportunity for the buy side. You can get faster over here. And, you know, in any other industry, shaving seconds, shaving days, shaving weeks, whatever it is, provides a huge competitive advantage. So I think you're going to see smart tech companies, smart agencies, smart brands, like sniffing out who's on the cutting edge here because things are changing again. And, like, if they weren't leading before, here's our opportunity to have a real competitive advantage for the next two or three years. Yeah. Yeah, I totally see that as well. The next couple of years are filled with huge leaps. Yeah. And I love the terminology there, smart agencies, smart brands. Yeah. It's sort of the discerning between who's using it right and not. So I think that will be the difference. Well, we all had a pass before on, like, in the beginning of the industry. We all had a pass on, like, how quickly we changed. and then we realize, wait a second, technology might drive a real competitive advantage very quickly. And so I think, you know, there are now groups that are looking for that moment where if they jump in first and they can adopt some change, there's going to be a real advantage. Fantastic. That's a really good place to leave us. But thank you so much, Dave Rosner. I really appreciate your time. Wonderful to be here. I'm so excited. Excellent. Thank you. Bye. Hi. Phil Longacre here for Signal & Noise, day number one for Popsville. Here with a special guest, Eric Zubin from Emmett Advisory. Good to see you as always. Good to see you, Rio. So I know you came here to talk about quality, didn't you? How did you know? Well, your reputation for C2, and the other episode we did together did really well. We got a lot of good feedback about your discussion about quality and outcomes and all that. So maybe start there. Like, why are you so doggedly pursuing quality? Why does it matter? And why should people care? I think the story is, for those that don't know, I spent most of my career at Google. I was there for 13 years. And what I really saw was sort of that race to the bottom for the industry. And it wasn't just on price. It was on quality as well. And what that ultimately, the reason for all of that was buyers just didn't care that much. They didn't insert the quality controls. They started chasing cookies and users and attribution and kind of ignored the media quality that plays a really important role in also driving marketing effectiveness. Do you think brands care? Do you think buyers care? I know that's a loaded question. It's tough because it's not a clear yes, no. There's generally no binaries in this world. There are buyers that care. They are, I would say, the minority. Many will say they care. How many have actually put those things into practice? They are in the minority, but I do think that most enterprise buyers, as they learn and they understand that their money is going, not just the amount of money that's going towards lower quality media, but they're overpaying for that, that's when they become aware and realize that. So MFA is a perfect example. Do buyers care about MFA? I mean, I think they should. The outrage that you see on the Internet makes you think that they do. Whether they've done anything about it, what they've done that's actually systemic and sustainable, those are sort of open-ended. We don't know yet exactly if we've solved MFA or we've sort of pushed it. MFA is an interesting one because I remember, like, with clients, they would tell you they don't want MFA, but if you pull out all the MFA sites, CPMs, they go up, right? I mean, just naturally because you're paying low CPMs for inventory, right? So I always thought that was an interesting, like, if you really care, you should be willing to pay more for quality, right? Yeah, that's exactly the story. There's definitely the paying for quality. It's moving away from cost and thinking about value. And I think one of the important stories, a narrative that I talk about a lot is it's not just – I don't think low-quality media is that, right? You just have to pay the right amount for it. And the question is, how much of your budget that you're assigning towards this low-fall need is, can you be effective if you're spending too much of your ads on MFA sites where there's 20 ads around it, or on CTE, you're spending 40% of your budget overnight, can you be as effective? I'm not saying you shouldn't spend there, and I'm not saying you shouldn't. Maybe it is worth the same, but it's probably worth less to you. And so it's more of that allocation of budget, and it's moving away from this concept of average CPMs, because that's really, I think, one of the biggest scourges. if you were to reapplicate from a lower quality but cheaper media towards higher quality media, even if you kept everything at the same price, in theory that would be a good thing, but your average CPMs would go up. And that's often what the procurement teams are looking at, and that's sort of that red flag that they see. You're taking the time and the price maybe for low quality, low CPM, like within a brand's, let's say, media plan. Absolutely. The concept is called the mere exposure effect. And so there's sort of a lot of research around this to show that just for Coca-Cola to serve you this tiny little ad that you see over and over again, that's enough for them to sort of be in the back of your mind at least, not top of your mind, but back of your mind for when you're thirsty, you're going to think about their brand. And so there is value for that low-quality media, but it has to be cheap enough to be worth it. So it's like building the right plan to account for, okay, we're going to spend some on high CPM, high quality CTV. We're going to spend some on, you know, like this low quality, you know, long tail stuff that's going to be more ubiquitous, but knowing what they're paying for and planning for that. Yeah, and it is sort of campaign specific. And so that's easy for Coca-Cola because everyone knows their red iconic brand. They're going to buy it, everything else. But if you're a brand new marketer or a brand new product that needs to get out to market, the low quality media probably won't do the trick. You need to sort of like at least first get people to understand what your product is and who you are, and then you can do sort of more of that repetition through lower-quality exposures just to make sure they're top of mind for when those people do enter the market. What's the angle for the publishers here? Because obviously focusing on quality as a former planner and buyer, I mean, it should be helping the premium publishers, right? Because there's a focus on quality and getting your ads in front of the right places with the right publications with great content, with great audience data. I mean, that should be, I would think, of course, in quality is going to be better for them, correct? Absolutely, yeah. I mean, there's a lot of people that have been sort of screwed over in this sort of race to the bottom, and quality publishers probably more than anybody else. And so we've seen money move from where it should have gone, higher quality publishers, towards that sort of lower quality MFA. as buyers shift to higher quality, it's going to go towards publishers that have fewer ads on the page, that have more attentive audiences, maybe more loyal audiences, maybe audiences that have given them their data, and so they can use that as well to sort of make even more value out of the impressions that they have. You've talked a lot in your writing about vanity metrics and how you don't like them, you think it's forced. I think the obsession with vanity metrics has made brands and made buyers focus on the wrong things. Can you explain that? Yeah, so we're in a very dynamic ecosystem, and so nothing is static. Nothing remains good. When viewability first came out, it was a really valuable metric because it chopped off all of this inventory that wasn't valuable at all. And I wouldn't say it's a vanity metric entirely. It's not broadly, but in isolation, it's a vanity metric. So visibility would be like one example. It is an example, yeah. Completed views is another one, and that's the one that I think is a little bit easier to understand. And when video, when online video came out, most video ads were sort of the large full screen sound on ads. And so a completed view was a pretty good proxy that somebody watched it and was sort of attentive. When outstream video came along and you had all these small muted video players playing on the bottom below the fold or whatever, they completed two. And so completed views, viewability or, you know, complete straight, that became a vanity metric. It's easy to gain, is your point? It became, it was gamed. Now, what I like, I wouldn't throw that metric away. What I would say is you have to add other metrics on top. Quality needs to be multidimensional. So just the fact that it completed is enough. Completed plus it was a large ad plus the sound was on, that's a really telling set of signals that tell you that that was of the highest quality of the type of ads you can get. Turning to video, CTV, a lot of budget going. I think I was reading the other day it's going to grow 18% this year, which is pretty remarkable. And then obviously traditional linear television is shrinking, but CTV growing a lot. I think there was initially this perception that it was all high quality, the CPMs are higher, but there's been a lot. I think you talk about gaming. There's been a lot of that, especially CTV, and there's been debates about what is real CTV inventory in-stream versus out-stream. The viewers who are not deep in the space might not know a lot of these terms. Maybe just walk us through that real quick. What is real CTV inventory, in-stream versus out-stream, and what do you consider quality? Yeah, so I don't want to be the one to determine what is CTV because I do think that it's a little bit in the eye of the beholder of the consumer and of the advertiser. So each one of them might decide that feels like CTV to me, and we should consider both of those. I think one of the issues we have is that we have, like, sort of this broad category. So we do have a category problem with CTV and online video where, like, who knows? They are these channels that, as I tell it, they're not a monolith. There's really high quality. The most premium thing that you can think of is CTV and all of it. And then you have some of the lowest quality, some home screen, the small muted videos that are autoplay that also consider themselves in those categories. And so we have to think about what are those quality metrics that will determine the relative value of different impressions or different buckets of media within each of those channels. We talked to Dr. Fu a couple of weeks ago on this pod, and I even asked him, like, role of CTV at the web. His opinion is CTV is not part of open web. It is a separate thing. And opinion on that, because it sounds like what you're saying is a lot of online video that just like maybe it is CTV, right? even if it's not like on a premium, let's say it's not Amazon Prime Video, it's not Hulu, it's not like what we can definitely consider TTV. It sounds like you're saying there's a gray area of stuff that some brands, some advertisers may be fine with it and may consider it TTV. Yeah, absolutely. And I think it would be foolish for somebody to just, when you draw an arbitrary line, that's sort of the issue with the binaries. You're, first of all, too often valuing everything that's above that line as equal. But then you're saying everything below that line is worth zero. And that's not true. There is value even in those small little muted video ads. And so this is like moving away from binaries on the sort of Dr. Fu comment. I disagree. I actually think it's more helpful to consider CTV as part of the open web. The way that I've looked at the open web, I haven't seen this framing in a while that I've put together, but it's this idea of like the openness of digitally transactable media. And so when there is a buying platform that is exclusively able to access certain content to what they own, right? Google, only Google, only alphabets, their DSP can access YouTube. So YouTube is an operated inventory for Google. Exactly. And it's exclusive to them. That's not the open web. Video. So the walled gardens and everything else is the open web. And so CTV should be part of the open web. We don't have to think about the open web as like what's the browser, you know. Jesse Fu was saying how he believes buyers should go to the platforms themselves, so go to Amazon DSP, go for Amazon Prime Video, go to DV360 for YouTube, go to, you know, whose platform for it. He was saying go to them, buyers should go to them, not work through a DSP, buy direct. That was his thesis. I had never heard that before. Oh, actually, it reminded me of what he was saying that I've always been a fan of, too. There is a very finite amount of human time and attention. But we have pretended like it is infinite because there are infinite number of impressions. There's always, you can always create... Not human impressions, necessarily. Not necessarily human impressions, but maybe there are, but then there's 20 of them on a page. But the amount of time and attention that people can spend on any of those ads is finite. And so we've seen all of these things grow. The time and attention that people have and spend on this media has not necessarily grown. That imbalance of sort of supply and demand, and I believe when we finally recognize it as such, right, when people realize that they're buying 20 ads on a page instead of the one that they should be focusing on, when we deal with that supply scarcity that actually exists, the price is going to go way up because the demand is still going to be really high. So think about the most premium, CTV, actual people watching sound on video. There's not that much. Not only so much, but it's a limited amount. When buyers actually concentrate their demand there, the prices are going to go up, and most of that supply is not going to be transacted in this sort of programmatic DSP, at least not in a biddable, right? It's going to be a seller's market where they can say, hey, I've got something that almost nobody else has. I am going to keep it reserved because if you don't buy it from me up front and a guaranteed and pay me the top dollar for it, somebody else will. The Super Bowl gets sold out every single year. So looking at biddable versus non-biddable, I know in CTV now the majority is biddable. I'm not mistaken. I don't think so. No? I don't think so. What is the split? Do you have any idea, roughly? Who was I listening to recently? It'll come to me. Tatari. Tatari's CEO. He was at Market Tech and he's been at a couple of other events. He's saying that the vast majority of what they're seeing, like 90% of CTV, is not transacted in a biddable way. It might be running through programmatic pipes, but it's not biddable. Like, those are set guaranteed prices of how much is going to be filled. Maybe there's a little bit of intelligence of things moving around, a little bit of optimization, but it's not a per-impression sort of priced media. Right. There's no open exchange or it's not – I mean, even some P&Ps are still biddable, too. They're still a little bit biddable. I think most of it is almost closer to a reservation. It's closer to an I.O. deal. Maybe there's a little bit of intelligence. That's like going back to the future, right? That's like how things used to be, right, when you're buying linear television, right? You're going to the upfronts. I think up to 95% used to be transacted through upfronts, and there was a scatter market, which was more expensive, interestingly enough. It was not a race. To your point, there was only limited inventory. And that's the advantage to buyers that buy in advance, right? You're going to pay more to go into the scatter. And so you want to buy it. You want to reserve it. You want to hold on to it. That's how markets that are non-commoditized, right? You have a differentiated market and people will buy. There's always going to be buyers that want the best stuff and they're willing to pay more for it. They're going to pay a premium. They want that early access. And so that's what we need to sort of return to because we've treated media as this undifferentiated commodity. And now we're starting to realize there are qualitative differences. There's scarcity on the really good stuff. And that's going to be transacted differently than sort of this. What we had is that remnant programmatic space. That's interesting. So what South Pethesis is saying, if you want quality, you should be able to understand the difference in quality and not quality. But also, like, brands should be willing to pay for it. And they should understand, like, there's a limited quality inventory. It should cost a premium. And you should factor that into your media plan. Yeah, that's the story. You have to factor that in. Don't pay for all media the same, right? Not all created equal. And so once you have that basic premise, there's really small things to do. Like on CTV, the example that I always show, most ESPs in that sort of programmatic or even non-programmatic, non-bittable environments, a lot of spend happens in the middle of the night, right? Because budgets need to get spent, and so they're spending 20%, 30%, 40% of those CTV budgets in the middle of the night. And those are mostly not people watching them, I guess, correct? Well, no, they might be. CTV is a one-to-one, so it's a little bit different than linear TV. So I'm not saying it's worth zero. it's probably worth less. But the problem is that buyers are paying the same amount for every single hour of the day. So at 2 a.m. and 2 p.m. and 8 p.m., they're paying the same CPM, which doesn't compute for most advertisers anything. No, prime time, not only is it like more prime glamorous, but also there's probably more people in the room. When you look at data parts, it would be different rates for normal television. It's a little bit different than linear, which isn't bought on it per impression. So I acknowledge it's probably not going to be a one-to-one match. But most advertisers would look at this and say, I don't want to spend 30% of my budget overnight. And I certainly don't want to pay the same CPMs at 2 a.m. as I would to reach a user who might just be more receptive to the message at 2 p.m. or 8 p.m. Or there's going to be more people in the room. And so that little trick and that really, like, obvious example, I think is enough to open people's eyes and say, oh, what other quality signals are we undervaluing? So you've been on the quality baton wagon now for at least a year, right? It's been about 10 years. But, yeah, like, you know, really up and at them for probably closer to two, but it was our attention for at least five plus years before that. Do you think – are you noticing a difference? Absolutely. I mean, people are talking about it more. I mean, the paper that I'm about to release now that was commissioned by STEM, the Coalition for Innovation and Media Measurement, that came about because they noticed. Everyone's talking about quality. It's a buzzword, but no one defined what it means. And it's hard to sort of, like, move together as an industry without this definition. And so Sim decided we're going to move forward. We're going to help define what quality means that really helps their constituents in the CTV world, both advertisers and publishers. They've been a little bit stuck. They, in my opinion, was not the right approach to sort of make the fight about premium, right? They're like, no, we're premium and we're better. I don't like premium particularly because it's not just binary. And it's still subjective, but then you say like, oh, you're what, right? Exactly. And it's a quality that is actually a little bit more granular. And so you can have high quality and medium quality and low quality, and low quality is fine, but at the right place. And so it adds a little bit more of that granularity and the nuance that I think is needed to tell that story. Because if I'm looking at YouTube, and I tell the story, like, YouTube is gigantic, and it has some really high-quality stuff, right? Like, no doubt about it. You can find anything on YouTube. I mean, you can find some incredible content. Incredible, and to the point, I think YouTube's still at, right? It's their, like, top 1%, 5%. almost any single user and any single advertiser would look at that and say, that is as high quality as any other media that I can possibly find. But that's their top 1% to 5%. The vast majority, there's a lot of medium-quality stuff, the stuff that we watch every day, but there's also people just playing music videos in the background on their TV. There's also lots of kids' content, and now probably even more AI, whatever slot that's playing there. There's a lot of low-quality stuff, which I'm not saying don't buy and don't pay any amount of money for. I'm just saying don't spend 30, 40 percent of your budget on that and be like, oh, this is my YouTube. That average CPM, the average completed view or whatever else that you're looking at, that is not telling the story. And, you know, a little bit of deviation, but just because they can do and this applies to Amazon, this applies to everyone else, because you can do some attribution from video down to a search or to a purchase is not proof of the totality of the effectiveness. Even if you sort of ignore the fact that attribution can be gained, it doesn't tell the full story of the short and long-term value of people being exposed to ads, remembering that brand, buying it later, sometimes in a way that you have no way to actually track and prove, or in a more of a probabilistic model where it's like, you know, me and my wife, she saw the ad. I made the purchase. They're going to attribute that to my household, but I didn't see the ad. And so, you know, George Box was a famous mathematician. He has a famous quote. All models are wrong. Some are useful. I like that. It's a great quote. And I think that's sort of where we've gone wrong because we have been – our world has been modeled. We've pretended that we're taking everything in the models. Exactly. And we haven't sort of said, oh, well, they're not perfect. And so if you put that little margin of error, then you actually – you're forced to sort of prove it and triangulate and confirm it in different ways. I think that's something important. I think a lot like marketing really is about it's rare that there's a right or wrong answer. OK, well, how right is this answer? How wrong is this answer? Or this approach? Let's try it out. Let's test it. Let's test it. Let's test it. I mean, I think people. And with these platforms, it's easy to test. I think a lot of what happens now is all it's implement this and let's forget it. Right. Because it's because it's hard. So maybe maybe I change. Not fully. It might help in some ways, but it could hurt, too. Right. If the incentives aren't aligned and then we're still measuring things, assigning too much certainty to those models, that could be really detrimental. And so it really depends on what checks we have. The one thing, sort of going back to your point earlier saying, like, we're going back in time. I talk about the quality controls that we had in the broadcaster, and there were multiple quality controls. But ultimately, maybe the most important one is that we knew that there was no deterministic, 100% certainty. Everything was probabilistic. Everything was statistically significant. We looked at the short and the long term. It wasn't 100% perfectly effective, but on average across populations, we knew that it was. We got into this precision era where we sort of got that. We're like, oh, actually, the attribution model. Precision, right? We thought we were in precision. Exactly. So we followed it. It wasn't accurate, but it was precise. And we followed that. Even though it was modeled. We had these models for attribution, for MTA, and all these other things. Which I think to a time, they probably worked well. But once they scaled and sort of got out of control without the quality controls, it got worse. So now we're going a little bit back to the future, and we have to reembrace this idea that most advertising, you can't know specifically, precisely how it – or every single ad and how it affected every single person. You have to look at probabilistic. You have to look at populations. You have to understand that there's models involved. Everything just gets statistically significant. There's ranges of probabilities, and you look over the short and the long term together. Well, this has been a great discussion. we'd love to get you back on the pod for a long form 60, 90 minute discussion. It's an incredible topic and the previous episode we put out continues to get a lot of views. So thanks again for coming by and I'm sure I'll see you in the next couple of days. Yeah, I'd love to chat. When the quality paper comes out, I think you'll appreciate a lot of the frameworks and principles and then there'll be another great discussion from there. I just love it. Awesome. Thanks, bro. Bye.