The Accuracy Crisis in Advertising: Scott McKinley on Bad Data and the Hidden Tax on CTV

20 hours ago
54 min read

In this episode of Signal & Noise, Brett House and Rio Longacre sit down with Scott McKinley, founder and CEO of Truthset, to explore the hidden data quality crisis inside Connected TV advertising. Scott explains how inaccurate identity matching and weak audience data are quietly wasting billions in media spend across the digital ecosystem.

Drawing on decades of experience across companies like Nielsen and eXelate, Scott argues that trust is the foundation of advertising markets. As platforms like Google, Meta Platforms, Amazon, and Netflix continue to dominate with authenticated audiences, publishers and broadcasters face growing pressure to improve the accuracy and reliability of their own data.

The conversation covers:

Why bad identity linkages create massive inefficiencies in CTV
How trust drives advertising economics
Why authenticated audiences command premium value
The future of measurement, standards, and governance in digital advertising
Why data accuracy will become a key competitive advantage for media companies

Watch the full episode and join the conversation.

#SignalAndNoise #Marketing #AI #Data #Media #AdTech #LiveRamp

🔑 What We Cover💡 Key Takeaways🎯 Why This Episode Matters Read the full transcript below.

Brett House (00:01) Hey everybody, welcome back to Signal of Noise. I'm Brett House joined by my co-host Rio Longacre. And today we have a phenomenal guest, Scott McKinley, ⁓ the founder and CEO of Truthset ⁓ and someone that spent 25 plus years. I don't think we're counting up to 30, right Scott? Cause I stopped counting at 25. know that. Digital advertising, identity, data strategy, analytics, you and I cross paths briefly. Rio (00:07) Hey there. Scott McKinley (00:12) Okay. and Brett House (00:29) at Nielsen, although you were there for, I was there for about five years. You were there for what? Eight years, something like that. And then you've been, you've been now on this entrepreneurial, entrepreneurial streak. You founded IDFI, you sold that and exited that business. And now you've founded Truthset. And for those that don't know, ⁓ Truthset's a really interesting company. Scott is a worker, one of my old New Star executives, great Galetta that was behind OneID, which was sort of our identity. Scott McKinley (00:33) Yep. Yep. Hmm. Brett House (00:58) machine learning based identity resolution solution. And ⁓ Scott is a former Olympic cyclist, which I like to say, I know I've mentioned that to you a couple of times, because I'm a cyclist, but I'm more of a weekend warrior. are a professional athlete. So is Greg Galletta. But welcome to the show, Scott. Thrilled to have you. ⁓ I know it's kind of connect paths and ⁓ impossible, you know, 15 meetings a day is a lot and we ended up connecting. But, Scott McKinley (01:16) Okay. Brett House (01:24) We'd love for you to introduce yourself to the audience and tell everybody a little bit about you and if I missed anything in your background. Scott McKinley (01:31) Thanks Brett and thanks Rio for having me on. Yeah, this should be a fun conversation. You know, it's funny about the 25 years thing. I think everybody just cuts it off at 25 because anything older than that, like it ages you into the Methuselah territory. What's scary about that, my first toe in the water in digital was in 1997, the year after I retired from Brett House (01:47) to obsolescence? Scott McKinley (01:59) for bike racing, I had a partner, we taught ourselves how to code and we were building websites for companies. And during that year, said to, I said, I look at my partner one time, said, I think we missed it. I think we were a little too late. To the internet in 1997. that's how it started. then I, ironically, this is a funny story, but we've time, we'll tell it. Rio (02:10) That was big business back then, right? Scott McKinley (02:28) I needed a job, like a real job, because I got married, my wife got pregnant. We realized, yeah, yeah, it makes you do things and you figure stuff out. And I picked up a ⁓ freelance gig with Cox Enterprises, which Cox at the time, yeah, well, Enterprises is the overall company, Communications is one of them. And they had done this thing just about, Rio (02:35) Yeah, that'll force you to get a real job fast, huh? Brett House (02:39) You have a mortgage. Rio (02:41) Yeah. Stuck Cox Communications, like same organization. Scott McKinley (02:57) six months earlier where they said, oh, the internet's here. So we're to treat this like another medium. Cox, that's what we do. Cox started in newspapers in the 1800s. Literally, started. Rio (03:08) They're way back. Yeah. And then they have they have radio stations. They have print media. Brett House (03:09) Yeah. Scott McKinley (03:12) This is what they, you get from their perspective, this is no big deal. It's another medium. What we do since the 1860s, when they bought up a bunch of newspapers in the Midwest, that was great. They built that up. created the first, you know, big sort of aggregated newspaper empire. And then what was next? Radio came along. What did they do? Cox radio is still huge. So they dominated radio. have tons of stations across the country still. And it was ⁓ television. Rio (03:41) Have a nice spun off the radio stations. I think they're just in like an MVPD, a cable company now. Scott McKinley (03:45) I don't know. I don't follow what they've done yet. But the point is every time a new medium from radio, newspapers, radio, television and cable, they just this is cool. They would allocate funds very smart. They would try to get as much leverage off their existing properties as they could to support the new the new media empire that they were going to build, which they invariably did. And so when the Internet came along, they're like, we're just going to allocate 100 million dollars to new media. And they cut that. They call it Cox Interactive Media. again, Peter Winter was hired to run it. He's a really good guy. But you know, this was a different animal. This is the animal that was going to eat all the other animals, basically. it took them a while. Brett House (04:23) Yeah. So the transformation playbook didn't quite apply to the new internet, right? Because we've heard of a similar story from Jessica Hoag, you know, is the chief get officer at Hearst. And she talks about this continuation of transformation and reinvention based on new media over a couple hundred years. But you're suggesting that it doesn't necessarily, it didn't apply quite as easily as they thought to the internet era. Scott McKinley (04:28) No. Yeah. Yeah. The best example, the best example was, you know what supports newspapers? Classifieds. Classifieds were the workhorse. I was in San Francisco, I was in San Francisco, they were done in Atlanta and Craigslist showed up and they're like, Scott, what's going on with this Craigslist thing? know, is it joke? It's right, why, we don't have to worry about this. And I was like, oh yeah, you have to worry about this. Rio (04:53) Craigslist killed that, yeah, mean, that went away. Brett House (04:54) Yeah. They're coming for your they're coming for your revenue Rio (05:12) 40 % of your revenue is gone, just snap your fingers, right? Scott McKinley (05:14) That's right. That's right. That's what happened. it's, when you stand back and look at it, Craigslist and the rest of the online, ⁓ classifieds players like those classifieds.com, there were some other ones eBay basically we used to be kind of like a classifieds at the start, but they turned, turned, they turned Craigslist turned a $24 billion industry into a $2 billion industry. Like the whole thing compressed. Brett House (05:31) Yep. Yeah, and yeah, like in overnight, like in terms of the history of Cox, like this was a drop in their history sort of ocean, right? Timeline, you know, that's. Rio (05:40) And it took them years too. Scott McKinley (05:49) But it Rio (05:49) Craigslist Scott McKinley (05:50) was. Rio (05:50) made it, initially it zero for them and they slowly built up into a small business with the same shitty website. I just thought that was fascinating, but you're right, it gutted the industry. Scott McKinley (05:58) It shrunk it. It took the whole pie and shrunk it. wasn't like they were going to eat the pie and take it away from Cox. They shrunk it. So no matter what you did, like that, that opportunity is gone and they made investments in classifieds.com cars.com and a bunch of other like sort of new media diversification classified strategies. None of them really, I mean, when you shrink a market, by the way, we'll get to this probably, but the kind of same thing's happening with Nielsen. Nielsen's been running on a $2 billion. Brett House (06:14) Yeah. Scott McKinley (06:27) TAM, which is the linear addressable market, which they measure obviously, the linear telecom market. And that's just gonna kind of go away. I mean, that one might not shrink, but it just gets eaten alive by a new set of technologies that you can't migrate to at all. Brett House (06:30) Yeah. Yeah. Yeah, and it's a golden goose. That's right. Rio (06:44) Well, Scott, yeah, people don't realize this about publishing. It's actually half the size it was like to your point in 2000, you look at today, like the whole publishing industry, like the overall revenue is half this, quite literally half the size, despite the economy, like probably more than doubling over that time. So it's remarkable the, know, the precipitous and ongoing drop and how it is. And yeah, it's wild, but Scott, mean, the reason why we're excited to have you on here is like, there's been a lot of talk about CTV. We had some great discussions at. possible about that ⁓ with your extensive experience in this industry and media. And I think really the key tension that the whole advertising business is built on this enormous confidence that we have, right? When you look at precision, precision advertising, targeting to do that, you need addressability or you need audience quality. need, yep. We're in the age of outcomes after all, you need identity graphs. Last couple of years enter clean rooms and now AI, right? AI driven optimization and targeting. ⁓ But I think under that's his basic question, right? Like are the signals underneath any good? And I guess what you're saying is a lot of times the answer is no, and I don't disagree. And like, if that's the case, what is like modern marketing infrastructure really look like? And what is a, is there a more sophisticated way to move, to move bad data through this, know, or good data through the system faster to get, to get to those better targeting and better outcomes we're looking at? I think it's going to be a fun conversation. Scott McKinley (08:11) You want to get into it? Brett House (08:11) Totally. Yeah, yeah, yeah. mean, you've Rio (08:13) Let's do it. Brett House (08:14) been arguing about this data quality problem for quite a few years. mean, ever since I think you left Nielsen probably before that. And I remember a lot of the conversation, especially at Newstar, we were preparing for signal loss. It was cookie deprecation. And all of us spent countless cycles trying to solve, because our clients were saying, what's going to happen to this house of cards? Is it a house of cards that we've built? Scott McKinley (08:17) Yeah. Rio (08:30) and panels and articles. Brett House (08:36) What have you seen across the companies that you've run and then obviously TruSet that's convinced you that we're really unreliable data is kind of at the core of problem and it continues to be a problem. Scott McKinley (08:48) Yeah, so that is the genesis of the company. ⁓ One of my roles at Nielsen was to do diligence on potential acquisitions. So I had little team and we would go look at what they're doing and invariably, probably more often than not, they were obviously massively exaggerating their capabilities and stuff like that. But every time we got into the weeds, we would look at the data sources of the data they're producing or the methods they were using to assemble data for things like targeting or whatever the thing might be. And I was just horrified every single time. Like, I cannot believe that we've built this, you know, multi-billion dollar technological infrastructure that runs on really low octane fuel. You know, like that's, it's just a great analogy. And I actually, you know, I actually wanted to do this at Nielsen. I felt like, Brett House (09:35) Yeah. Yeah. Scott McKinley (09:42) The linear TV market was obviously aging out. It's gonna go away almost completely It's all gonna go to addressable and addressable kind of like that conversation we had about classifieds Addressable CTV is it see the people who measure the way you measure addressable CTV is never gonna enjoy the same economics that Nielsen enjoyed for measuring linear television Brett House (10:05) Yep. Scott McKinley (10:05) Because it was the word. was the only place he went. It was the currency. Okay. It had some governance with the MRC. But it was the way that buyers and sellers agreed to size an audience on television. And it was very helpful. The people right on Nielsen, I was there when we tried to make a transformation and we failed for a bunch of reasons that aren't entirely rational. I can explain those to you actually. ⁓ But. ⁓ Brett House (10:11) Yeah, the accreditation, right? Yeah. Scott McKinley (10:35) Where was I going with this? Was it, it, was it, it, was it, ⁓ Rio (10:40) Well, I guess, yeah, we're talking about Nielsen and measurement, but interject really quickly for the audience. So Nielsen is really providing a measurement currency so we can evaluate like how many households within a specific group are actually seeing a specific ad that's running the program during a specific day part. not everyone who's listening probably understands how ads are bought, right? So I think that'd be a good thing to dig into a little bit. And then we can maybe even go into CTV because I think that's a really cool discussion about how Scott McKinley (11:02) Okay. Brett House (11:10) Yeah, for sure. Rio (11:10) data works there and targeting an inventory. Scott McKinley (11:12) So picture the 50s, right? When, or whenever it was, when television starts getting mainstream and all these families are crowding around the console TV in the shag living room with the wood panel and they're watching the $64,000 question or ABC news. then, you know, think about those days. So all the marketers are seeing this happen, the big CPGs and they're thinking, how do I get these audiences? But there's no way to size them. So what happens? They all make primary research. Rio (11:33) Yep. Scott McKinley (11:42) They brag about how big their audiences are and nobody believes them. It's a lot like programmatic looks today and even CTV. Nobody really knows how big these things are. Nobody knows how many humans are seeing all these device IDs that are being bought and sold and traded. Nobody knows something bots. And so when Nielsen came along, there was chaos in the fifties because the audiences were clearly watching television. Brett House (11:45) Yeah. Yeah, yeah. Rio (11:50) Ha Brett House (11:59) Yeah. Scott McKinley (12:09) It was an incredible new medium where a brand had an opportunity to tell a sight, sound and story. 32nd vignette of the brand and how this is going to help your lives. was a gorgeous. And what happened? This is facts. When Nielsen, it took them 15 years to build a rating system. They never said it was going to be perfect. It was always an estimate. But when it went, when that system got asserted in the market to give buyers and sellers at least a directional view of how big these audiences were, the friction. Brett House (12:16) Yeah. Yeah. Scott McKinley (12:36) got pulled out of that market and that market started to grow like crazy. it was literally, can absolutely credit Nielsen for the linear television market becoming a $70 billion market very, very quickly. And... Brett House (12:49) Yeah, a really important part of it removing that friction point, right? To allow the economics and for people to have the confidence to actually make the purchases based on real estimated reach metrics. Scott McKinley (12:52) Yeah, like- Okay. Rio (12:58) Well, yeah, the buyer would know, would be able to say, we know what we're buying. We're buying like an advertisement in front of the specific audience, right? You can understand the value of it. can think about how much you want to spend on it. Brett House (13:06) Yep. Scott McKinley (13:07) The word we're circling here, the buyer has the confidence in buying, is trust. The word is trust. When there's trust in markets, when there's trust between buyers and sellers, when there's trust between two humans, life is easier. Things go better. You can look at a ⁓ bar chart of the top 100 countries in the world, and you can map their GDP per person. Brett House (13:15) Yeah. Yeah. Scott McKinley (13:37) against levels of corruption in the country. And it is exactly linear. Like India, huge population, super low GDP per person, highest corruption rates in the world. mean, it's so you can literally correlate everything. can trust is trust removes friction, lack of trust. It's all friction. And I would argue that right now in the digital space, we're in a lack of trust world, right? Nobody really knows. Rio (13:48) Barely functioning legal system, right? Yeah. So Nielsen Brett House (14:04) Yeah, yeah, we're in- Rio (14:04) was selling trust, right? I mean, that's really what they're selling with their model. They weren't saying it was perfect, but this is, Scott McKinley (14:07) They're so 100 Brett House (14:10) Yeah. And I Scott McKinley (14:10) % Brett House (14:11) think with the, especially with the open internet, right? The walled gardens, which a lot of brands are flocking to, there is, think, a level of trust because they can connect sort of an engagement action with some sort of point of sale purchase. They can show attribution. Whereas the open internet, mean, you mentioned it earlier, you sort of alluded to, we've had a ton of conversations with Dr. Fu and Mark Zygorski at Double Verify around a lot of the slop and unverified. Rio (14:22) And it can show attribution. Brett House (14:36) There's a big promise for reach, which is what linear television was promising, but it's not really proving out. We're realizing that there's actually more inventory being promised than actually exists in the ecosystem. You're promising reach, but it's not all real reach, and that's problematic. Scott McKinley (14:54) Yeah, I mean, you the reason why measurement is so important is because of that story we just told. I measurement itself was the thing for linear TV that allowed the market to establish some trust between buyers and sellers and grow. Right. And I think if you want, if you want to be, there's two ways to look at the current situation. You can look at it very, very negatively and say how bad it sucks and how much bots and fraud and everything else is flowing through the system. Or you can say that's okay. We're just behind the curve. We're just catching up. Eventually we'll have measurement. Things will run. The market will grow. There'll be trust between buyers and sellers. That's what happens. So my assertion, just coming back to the Nielsen thing and truth set, I thought, well, if Nielsen, because we kept buying companies and then messing them up. And I realized like, this is not the way to transform because, know, Nielsen takes orders. Right? Brett House (15:25) Yeah. Yeah. I was part of one of those companies. Excellate, excellate Nielsen marketing cloud that I had people, I had one story where, oh my God, I don't even know, can I tell this? guess we're an independent podcast, but where I had people, because we were trying to launch a business unit with Mark Zagorski and Damien Gavrache. You knew that whole play, right? We brought in some of Matt Krepsich's MTA solutions in the MIMM and started commercializing those. But there was a lot of pushback organizationally around launching this very successful high growth post-acquisition. Scott McKinley (15:57) Yep. Brett House (16:11) sort of marketing cloud, which was sort of next generation from Nielsen's perspective. And so they would have people like spy on our organization to ensure that we were kind of running by the playbook that they instituted. so I thought that was a little weird. Rio (16:26) That's funny. So they're worried you were gonna, you're somehow gonna sabotage or undercut their Brett House (16:30) Yeah, like people would shadow you, you know, to make sure that you weren't going out making bold statements about ad tech, martech, et cetera, that we're going to upset. And Scott, you probably ran into some of this and you have your own, right? Cause you, you. Scott McKinley (16:34) Yeah. Well, the truth is I wrote the strategy for the acquisition of Exelate. I was the architect, but I was coming out of a Mexico offsite where I was trying to buy two MTA companies, either Convertible or Adametry. And they both got acquired within 24 hours. AOLBot, Convertible for $107 million. Brett House (16:46) Yeah. Okay. Uh-huh. Yeah. Scott McKinley (17:07) And then Google bought a Dometri. And so that I, and Haskell said to me, what, what do we do? And I'm like, you know what? Fuck it. I'm like, what is, what is, what is MTA? It is a database that can collect or a data infrastructure that can collect exposure and transaction data. And it's a set of analytics on top. And I said, we've already learned, we've already learned at Nielsen that our analytics are probably better than the analytics of these little companies. That's what we do with analytics. We do media analytics and we should be able to do the analytics part. In that case, all we really need is a DMP. So that was what started that, okay, great, let's go buy a DMP. But my strategy for the thing, which is what got it done, was a plan to essentially use the DMP to platform Nielsen's data, right? Like my vision for the thing was, which I sold out, my vision wasn't, let's let X-Lake keep doing what they're doing. My vision was, holy smokes, Nielsen has global... Brett House (17:41) The data management platform, yeah. Yep. Scott McKinley (18:04) media exposure data, global, and as global buy data with the RMS retail measurement systems, let's bring together, we were trying to bring together what we called watch and buy, you connecting exposures to transactions for a long time, but there was so much friction. To me, this was the thing that was going to bring watch and buy data together. And then we would move into the, it was a great. Brett House (18:08) Yeah. Yep. Yeah. Rio (18:24) Yeah, for those who don't know, DMP, a data management platform, you've described it as a cookie farm, it's collecting cookie data, right? So combining that with the other signals in head, mean, that makes a lot of sense. I can see why there'd be a sound strategy at the time. Brett House (18:35) Yeah, yeah. Well, there was was kind of a breakdown with household versus persons level data. Can you talk about that a little bit? Right. Just in terms of how do you actually fuse those? Right. Because a lot of the Catalina solutions, the buy solutions were all at the household level. The television data and even digital ad ratings was still was still at a household level. And you're trying to get down to a person's level. Scott McKinley (18:47) haha Rio (18:56) This is even back with back with when it was with linear, right? We're not having gotten to see TV yet. So it sounds like that was an issue even back then with the Nielsen data, right? Scott McKinley (19:01) Everything's your household. Brett House (19:01) Yeah. Scott McKinley (19:04) Well, that's, mean, sticking with the X-Late story, know, the, when I, when we evaluated X-Late, I said amazing pipes, amazing connectivity, amazing. There's so much bleeding edge technology in this data management platform, which is exactly what Nielsen needed. But, and I, and I really actually dismissed all the cookie data that was in there. They had no sense of identity to your point. They had no sense of a person ID or a household ID. They just treated a cookie as this must be a person. Brett House (19:28) Yep. Yep. Scott McKinley (19:33) How do we classify it and sell it? That was, and I was. Brett House (19:35) Yeah, you should have seen how they profiled me. I think I was a single female living alone in New York City when I ran my profile analysis. I'm like, they've got me completely wrong, right? Scott McKinley (19:45) But you gotta, we've got to digress for one second here, not throw all these companies completely into the bus because they, know, the cookies were reliable for a long time. Like device IDs were very reliable for a long fricking time. Retargeting is still reliable. So, you know, and then you can't blame them if their buyers are saying, I want scale over accuracy. Like, hey Mark, and all you guys over there Exhalate. Brett House (20:00) Yeah. Yep. Scott McKinley (20:13) When you have a chance to give me a hundred million cookies that might be something versus 20 million that you know are something, I want the hundred million. I want the hundred million. want the hundred million. you know, it's yeah, when I looked at it, was like, I don't want any of that data. Like Nielsen is about truth. We're going to use this to tell the truth about what's happening. We're trying to migrate that into a personal level idea and a household level idea. Like you said, there was even budget that I argued for. said, I don't want to acquire this company unless there's X million dollars set aside to build an identity graph to put in it. Because to your point, Brad, there is no sense of person or household, which is what I can see was absolutely going to be needed to be competitive and dominant, even just in measurement, you going forward. that was, but the story here that's really relevant is like this starts to explain why the internet quickly became such a ⁓ mess. I wouldn't even call it signal and noise. I'd call it static. It's like. There's no value to it. it just, it's just a, know, like right now there are 20 billion hexadecimal values floating around in these buying systems. These representing devices, which is supposed to represent people. Supposedly this is, this is not even signal and noise. is static. You know what I mean? Noise I would characterize static is you don't even know if it's real or not. Noise is okay. It represents something, but we really have no idea what it represents. And signal is it's something that Brett House (21:29) Yeah. Scott McKinley (21:39) probably likely maybe represents a human being that's gonna have a kid and a wallet in the car or a trip they gotta take. So you really want the signal. And it's like, if you look at, so look at the 20 years of development, everybody built these either noise machines or noise processing machines, one or the other, right? Like you just had to like, like if you were on the buy side, sorry, if you were on like the, need to create audiences to sell. I'm gonna create noise. created, they just spat out noise and then the buy side would buy all the noise and it's either just cover their eyes and hold their nose and go on their merry way. Brett House (22:07) Yeah. Yeah, noise, yeah, noise kind of sort of behind the segmentation schemas and the target audience groups. Rio (22:20) Yeah, and if the CPMs are low enough, to your point, scale is sometimes, you know, there's little precision, but we have scale anyway. It doesn't even matter, right? So, I I saw ⁓ a lot of that. Brett House (22:28) Yep. Scott McKinley (22:28) Well, Rio, you're tapping into yet another endemic problem with the whole system, which is the biggest buyers, the biggest CPGs. They were and still do use MMMs. Like they tried all these little experiments. They tried NCS with Nielsen. They tried Cercana. But they lost confidence in all of these very precise measurement systems. And they ended up going back to Mix. What does Mix do? Mix looks at what Mix just looks at how heavy was I here over what period of time and how can I see my sales go up and down relative to that media weight. Brett House (23:06) Yeah, to fuel like strategic decisions around budget allocation at the channel level and at a very high sort of aggregate level, right? Scott McKinley (23:14) Totally. So, and their system is not sensitive. NMMs are not sensitive. And what they do is they say, ⁓ we noticed when we did a time series regression, when we look back over the last year, we heavied up in Phoenix and sales went up. So I get a credit, the sales with that heavy up. And that's great. Except that what that does is it creates a KPI for efficiency. At that point, if that's your, if those are your calculations, all I want to do is drive down the cost of all that media weight, right? And that's what they do. I don't care if it's good or bad. just want weight. Right. Well, I do care, but I can't really make Rio (23:46) I just need to spend my budget too. mean, that's always part of it, right? It's funny, I work with a couple of clients, a couple of clients now, and I've seen the MMM get created, right, by the strategy team. And then, and then it goes over to totally different team who's actually going to buy the media. And they work with the agency to create the plan based on that. And then, you know, there's a lag, the translation of that into actual like tactics and where the buys happen, which platforms, which, I mean, it's just to say it's not a science is an understatement in my opinion, much of Scott McKinley (23:49) See what- Brett House (24:14) Well, the activation from the insights in that six month window is maybe less scientific than it needs to be. It's not directly tied to your buying platforms. Yeah. Rio (24:21) Well, I'm talking just translating the MMM into an actual media buying plan. Forget about even like looking at the outcomes and tying back the measurement to like how you reconfigure and adjust that. I mean, think that's a whole other thing. Scott McKinley (24:35) point is that if that's the way you're measuring and the resulting action items of what you should do to make it better is push down media cost, this is bad. This literally created a lot of the mess we're in right now because what that does, motivates all your partners to optimize for scale with no caring about quality whatsoever. It also grinds down the publishers, which actually works against you as a major brand because you need those publishers to survive. Rio (25:04) You do. Scott McKinley (25:05) to attract people who can rent their time and attention for seven seconds. Brett House (25:09) Yeah, they need to be able to effectively monetize their properties. And if you're going against that, you're really going against the open web. And then you're forced into these walled garden type of ecosystems to do all your buying, right? Scott McKinley (25:19) In my, in my first start, my second startup, which was a mixed sort of a advanced mixed modeling company meant to be digitally oriented, meant to try to solve this problem. Uh, we had a client, I it's a top five CPG. Like I shouldn't say who it was, but they, we did a study for them. Um, this is, there was barely even any video inventory. This is probably, uh, 2005, maybe 2006. mean, there were some. inventory coming online, Brightroll was there, all the rich media stuff is there. It was some video online. And so they wanted to know, how do I compare the ROI, the ROAS, the effectiveness, efficiency of online video versus television versus print? These are the three things that they bought. So we did this very robust, very robust, and I would say accurate study. ⁓ One fault of mine is I always. Brett House (25:53) Yep. Yeah. Scott McKinley (26:13) the truth even if it's not in my own interest sometimes but we told them the truth. Brett House (26:18) There's no obfuscation, there's no obfuscation with Scott McKinley, right? Scott McKinley (26:22) None, none, even if it puts me out of a job. And we told them, we showed them the online video at that time was very performant. was right up there. It might've been better at the time because the CPMs were low and everything else. And they said publicly, we're going to allocate 20 % of our budget, which was over a billion dollars to online video. They actually did a press release on this bragging about how modern they were. First of all, there was not that much inventory on the whole internet. They could have bought out the whole internet for $200 million. And they had a second stipulation. This is back to the MMM driving efficiencies instead of effectiveness. ⁓ They said, but we have a rule for it. And I'm what's the rule? They're like, we're telling everybody we're never going to spend more than a $3 CPM for this video. And I was like, dude, like. Rio (26:55) Right. So we're going to spend all this money and we're to be at a super low rate. I mean, he could have bought all the inventory 50 times. That's just crazy. Yeah. Scott McKinley (27:17) But the point was, and then at the time I talked to Todd Sassordati who ran Brightroll. It turned out this is the same, we shared this client and he named it. I'm still not gonna name who it is. And he said, Scott, says your story is so right. He said, I would tell them, I can't tell you guys that we're delivering a single video to a single actual human being because of this low price point. Like, hey, your kids use it. Brett House (27:38) Yeah. It's kind of the, you nailed the hit and all that. This was how many years ago was this? Scott McKinley (27:43) This is 2005, so whatever that is. Rio (27:45) Nothing's Brett House (27:46) Wow. And nothing's changed. We hear this problem all Rio (27:46) changed. Scott McKinley (27:48) Well, Brett House (27:49) the time, right? You're reaching CTV audiences on a mobile phone. Scott McKinley (27:50) I said like. Rio (27:52) So funny. Yeah. Scott McKinley (27:54) What's worse than that, it's like, hey, do your kids have Napster? You know if it loads really slowly and you scroll to the bottom, there's a bunch of stuff down there? Those are pixels with thousands of video calls. That's where your ads are getting delivered, Like literally not to people. But the whole point of this whole story, I think it's all relevant. It's that the industry evolved in a direction to optimize for scale. Rio (28:02) Yeah, well that's the video. Brett House (28:10) Yeah. Yeah. Scott McKinley (28:22) no sense of accuracy. Everyone assumed that the early performance of cookies was just going to carry through. No one paid attention as this stuff degraded. Nobody really, everyone turned a blind eye when all these fraudsters, whether they did it deliberately or just through evolution, started selling things that were not humans. They were not humans. Like these are just not people. And then a funny thing happened in around 2010. So Facebook showed up, Mark Sygorski never sat down and said, Brett House (28:28) Yep. Scott McKinley (28:52) I'm brilliant. I'm going to predict that if I can get somebody to give me their email address, I can build a media empire. He didn't say that. He said, I'm to build a ⁓ book face thing, Facebook thing, so that we can all laugh at each other. But in order to do that, you have to identify yourself. And that accidental hack is what allowed him to build the most powerful media empire the planet's ever seen by requiring people to authenticate, to say, here's my email address. Because what that did is it gave Rio (29:14) Mark Zuckerberg, yeah, yeah. Brett House (29:19) Yeah. Scott McKinley (29:22) Facebook control of the consumer. It gave Facebook a durable person level identity that's owned by the person that's not likely to change or be traded that they can hook data to. And that is the, that's the playbook. mean, the punchline of this whole conversation is why is, especially CTV and these big networks, why are they not just looking at the walled gardens and saying, that's the playbook that wins? The playbook that wins Google. Facebook, know, Netflix, Amazon. Authentication is the fricking hack. Like get it through your heads. Right now, and I did a study like I said, Nielsen privately, the average CPM at the time for an unauthenticated cookie based audience was like a buck 50 or a buck or something like that. Maybe even lower. That's the whole open internet average. Brett House (29:53) Yeah. Authentication with persistent identifiers, right? Which is in the in. Rio (29:55) Well... Brett House (30:02) Yeah. Yeah, kind of like a remnant inventory CPM. Scott McKinley (30:21) The average inventory for that same CPM for that same audience on Facebook or on Google and place where there was authentication was $8 or $10. And that's like the difference. Yeah, the difference. so everyone's like, can't believe they ran away with 70 cents of every media dollar goes to the, it's because they have human audiences that actually can see the ads and maybe respond. Like everything else is. Brett House (30:31) Yeah, it's almost 10x. Yeah. Rio (30:31) Huge difference, yeah. And it's no, it's no wonder why they actually convert, right? And you can actually prove outcomes and, know, it's wild. but Scott, let's turn to CTV. I mean, this is such, you know, it's growing, it's growing like crazy. think Brett, covered this couple of episodes, 18 % this year continues to, so like, and it'd be good for the audience to be able to hear about like, like how does C how is inventory actually bought? Obviously if you go to the walled gardens, you go to like, like Amazon to buy their own and operated Amazon prime video. You're logged in, they know who you are, but a lot of CTV inventories actually bought through, you Scott McKinley (30:48) I 100%. Yeah. Brett House (30:57) yeah. Rio (31:15) I'm debatable as an open webinar, but you can go to the trade desk. You can even buy like, like audience extensions, offsite inventory and Amazon DSP for, for Netflix or other places. Let's talk about like, how is that bought? Cause you have, you have the inventory itself, right? Then you have the data, the audience super important. And maybe we can walk through like, how is that bought? And like, how is, what is the role of data and how is important to have good data and how is that data? A lot of times not any good. Cause I think that's part of your thesis, right? Scott McKinley (31:44) Well, when CTV started to become real and everyone could see this is where things were going, despite a lot of people, you know, poo-pooing it, YouTube is not a TV channel. know, like, what? mean, remember, no one ever watched YouTube on their TVs, right? Like, that's almost the moment, you know. Brett House (31:57) Hahaha Rio (31:58) You're right. Brett House (32:02) Yeah, yeah, you still hear Rio (32:02) I do now. Brett House (32:04) denials around that today. I've heard denials with conversations with pretty senior execs and I'm like, it is television. I don't care what you call it. It's still people are watching 60 % of all television consumptions on YouTube on the big screen. Scott McKinley (32:15) Dude, a hundred percent. it's, you know, this is one of the problems that traditional TV networks have is they think their content is just extra precious because it costs a lot of money and attracts a lot of viewership. And I've had this argument forever about, you know, the idea of attention and outcomes and the value of media. Just because you spent, you know, a billion dollars making Landman and a lot of people tune in and watch it. Brett House (32:28) Yep. Scott McKinley (32:45) It doesn't mean that that ad is going to... You put a pharma ad in front of me for a... Put it in front of me 25 times in two hours for a condition that I will never have. That ad, despite you getting the media right and having great creative, is not going to perform. And I could be later that night on freaking Instagram, lazily, boringly doom scrolling. And I see a black and white shitty, poorly produced creative for a mountain bike. Rio (32:47) I love that show by the way. Scott McKinley (33:15) that I happen to be in the market for, I'm going to stop and I might even click. So you can't, you can't, they have this, this preciousness, which you understand they're defending their legacy. Like what we do is make great stories that people like to watch. Brett House (33:15) Yep. Long-form Hollywood content. It's a Hollywood economy right versus that Scott McKinley (33:32) Yeah, then, but the point is you guys are missing the data piece, right? You're getting crushed by Instagram in a terrible media environment because they know who I am every single time I log on and they can do it. Brett House (33:43) with shitty ads, you know, yeah. Rio (33:45) Right. Scott McKinley (33:46) But they're going to work because the three-legged stool of advertising is the audience. Like I made a product for somebody, it's an audience. Then the next one is the creative. What am I going to say to that audience? And the third one is the media. Where can I rent seven seconds or 30 seconds of attention from the human being that is that person, tell my story to that person. That's the, that's the, you pull out any one of those legs, the stool doesn't stand up, right? In the old days, in the old days, there was no concept of audience. It was all. Mass targeting, like every product that got advertised on television was to a mom who could walk down the street and buy this Swiffer mop in her corner store or her Sears or her JCPenney. And all that's changed. Like there was only Asian gender. That's what Nielsen could report on because it was hard to go beyond that. That's how it ran. So all the emphasis went to creative. This is where we got the mad men golden out, you the golden era of advertising. And it went to the medium, this incredible medium of television. And that's all you needed. And it was amazing. Brett House (34:38) Yep. Scott McKinley (34:44) Now it's all different. I think people have forgotten that just how important it is to get audience data right. There's a ton of audience data, but it's more static than it is signal or noise. Brett House (34:56) Yeah, and you guys have said that 40 % of every dollar spent on CTV programmatic is wasted because of bad data. And I think you calculated it, it was over 7 billion, 7.4, something like that, in lost CTV dollars. So clearly there's some friction in this ecosystem. What's the solve for that? How do we? Scott McKinley (35:03) It's more than that. Well, let's go back. Yeah, let's go. It comes back to Rio's question. Like how, how do people buy, buy a CTV? And when CTV started to get maturity and everyone could really sort of treat it like a real channel, there was a question in the industry, which was, well, what kind of a channel is it? Is it its own channel? Is it going to behave like linear TV? Are we going to buy it, sell it and measure it? Like we do linear TV, are we going to buy it, it and measure it? Like we do programmatic and it went the programmatic route, which I'm going to say. was the dangerous path it's evicted. Rio (35:45) What percentage biddable today, Scott, of CTV? Is the majority biddable? Scott McKinley (35:50) I don't know for sure, I would bet that ⁓ a lot of it is the majority is biddable. But here's the problem. When it went to the programmatic route, everyone was like, okay, well, how do we get the volume that we need? How do we create a transactable thing that we can buy and sell that represents a television that's supposed to be hanging in somebody's home? Right. And the thing is, so here's here's the, here's the Rio (35:55) I it probably is. I think I it is. Brett House (35:56) Yeah. Scott McKinley (36:19) supply chain of identity and data for CTV. It starts with a CTV. There is a CTV ID. It's set by the OEMs. It sits on the device. You can't see it. You can't get it. It's a device ID for the CTV, but it's not, at, right. The next one goes, right. It's hard to get the CTV ID. has to be, I think, acquired only from the OEMs who set it. Yeah, I don't think you can see it. And it's, you won't have the scale ever. Rio (36:29) So it's a device ID, correct, basically? Brett House (36:33) Yeah, an independent of IP address. This is is device specific to the television. Scott McKinley (36:48) Even if you got all the, you got all the OEMs or half the OEMs to give you half the CTV IDs, you would have big gaps in your viable inventory. And that's never sufficient, right? Brett House (36:56) Yeah. Rio (36:56) Will they share those with the like automatic, the ACR aggregators as well? like, is it, cause I know that a big part of the, like the, the screens themselves, they don't really make money on them. Or they, like, I think sometimes they lose money on them. Right. And they're making their money through, selling or syndicating all the data they're collecting. Right. So the device ID, that'd be one of the signals they're collecting. They're probably making money off of, right. The manufacturers. Scott McKinley (37:19) Yeah, but it's not, it's not prevalent and scaled enough to be the tradable identity. The tradable identity is IP address. So what you do is you can, can, that you try to connect the CTV with an IP address. Great. Now the, both those things are on the device at the same time. If you can get the CTV or whatever, you end up with the IP address. That's the thing that people are buying and selling the IP address then gets, then gets connected to a cookie or a email address or a postal address. Brett House (37:29) Again. Scott McKinley (37:49) through a variety of problems. Brett House (37:51) And those connections kind of suck, right? Like IP address to postal is super low percentage matter, right? Scott McKinley (37:56) So we did a sweet. Rio (37:56) Yeah, because IP is very broad, right? It could be several houses, could be a block. I mean, or if you're through VPN, it could be another state entirely or another country. Brett House (38:00) Yeah. Scott McKinley (38:04) So we did a study, a commission study last year by a SIEM, the industry body, which is doing really great things around this place. That's Coalition for Innovation in Media Measurement. Nice job, Very unwieldy to say all the letters. That was meant to shine a light on this and understand how accurate or inaccurate Brett House (38:11) The Coalition of Innovation Media Measurement, right? for NvG, yeah. Scott McKinley (38:28) are the linkages between IP addresses and the other things that you use to build audiences. Like IP address doesn't have a demographic. You need to understand who is in the home. So you need to link it to something where there is a demographic, like an email address or a postal, or you could do a cookie or whatever. We found that the ⁓ average among a whole pile of well-known providers of this linkage data, the average was, I believe, 13 % accurate when linking a IP address. Brett House (38:40) Yep. Rio (38:55) Wow. Scott McKinley (38:57) IP address to an email address. And then 16 % for postal address, I believe is what the stat was. And so you guys stand back. You're trying to support and embrace and leverage this new media channel that promises everything. Full site sound and motion delivery of brand stories into somebody's living room. What? This is incredible, much more measurable. Brett House (38:57) to to to to to to to to to to Yeah, okay. Yeah. Automated ad delivery that's switchable and you can adjust the creative you can you can do all sorts of things flexibly Yeah, that you couldn't do in linear. Yeah Scott McKinley (39:24) It's incredible, except that your guess about who those people are behind the screen is 87 % wrong to start. Like, that's how. Brett House (39:35) Yeah. Rio (39:36) It's wrong almost, not some of the time, almost all the time. Scott McKinley (39:39) It's wrong almost all the time. It's worse than cookies were, for sure. Probably worse than cookies are today. It's just really like we jumped from the frying pan cookies that didn't work into the fire IP addresses. They're not going to work. Now I will say, I just got to throw in here that averages are always very dangerous because they smooth everything out. And the assumption is that everybody is 13%. No, there are extremely accurate pools of IP linkage data. Roku, for example, they see the IP. Brett House (39:50) Yeah. Yeah. Yeah. Scott McKinley (40:07) Comcast, they're the ones who set the IP address, right? So there are pockets of sources of purely deterministic, no probabilistic guesses about the connection of a given IP address and devices connected to it, to the home and the people and everything else. But the goal really here, go ahead, yeah. Brett House (40:25) But yeah, and that supports either direct to publisher or direct to Roku type provider purchase, but it doesn't support ⁓ a cross ecosystem purchase from a media buying perspective. Like you would do with linear, right? Scott McKinley (40:40) Yeah, it's a mess. Exactly, it's a mess. And there's another problem happening at the same time. We're always out there advocating, just like we talked about earlier, Hey TV, know, to use an example again, the ability for Paramount to monetize 30 seconds of my time and attention during Landman, which is a show that I really, really love. I mean, if they don't... If I, if I'm authenticated and they ignore my authentication, let's come back to that for a second. And instead, instead just take my IP and put it into another bunch of IPs and sell it off to somebody because it's got the scale. I'm to see an ad that's not relevant to me. So they are not going to get the yield from that inventory that they should get. Take that same program. Now they've licensed it to Amazon. Now they've licensed it to Netflix or now whatever. They have my identity. Brett House (41:12) You Yeah. Scott McKinley (41:35) The chance, they're gonna monetize it 20 times or higher better than an unarmed. Rio (41:39) So these are all the other places where you can buy inventory, in other words, and like they may have bought that that IP address to supposedly link to you or like and then so to your point, like people are buying these audiences that are supposedly like highly accurate, but they may not be. Scott McKinley (41:54) Well, they aren't. It's just, it's not about. But like, I want to help save linear, ⁓ traditional television. I want, I think it's unhealthy for the walled gardens to control the narrative and the advertising around. Brett House (41:55) May not, think you're definitively saying may or not. Rio (41:56) They are not. Brett House (42:11) Wait, you want to save traditional television or you want to save the model that allowed for like broad reach across the entire ecosystem regardless of Scott McKinley (42:17) I want NBC and CBS and all these companies to survive. I think it's really healthy to have diversity. It's good for journalism. It's good for free speech. like you, it really is good for democracy. you know, that's like a people think like, really we're not digging wells here in India, but it is good for democracy. Like you cannot have the cultural narrative be controlled by a couple of companies, right? And it's bad for brands. Brett House (42:25) Yes, it's good for democracy. Yeah, well, it's like publisher balkanization is what it is. And you end up having all of these separate worlds that are completely corporatized in Behind Walled Gardens. Rio (42:51) believe in York Times, they struggled so much to monetize any of their inventory. now selling, like they make most of the money in crossword puzzles. mean, everyone's talking about how they're doing well and they are, congrats to them. But the thing is like they're monetizing all these games. They found different models, right? I think it's almost like they've given up on selling ads to certain extent. I don't know. Scott McKinley (43:09) Yeah, in their case, you also have just the tidal wave of really good user generated content. just I'm sorry, very few people are going to sit and read an article in the New York Times. For every person that's going to sit and read. Brett House (43:24) How life has changed since we started, right? When it was like newspapers, you know? Scott McKinley (43:27) I mean, for every human being that's going to read a physical newspaper or even go online and read a long form news story on New York Times, for every one person there, there are one million people happily scrolling away on TikTok and Instagram. Let's just face it, the consumption, the nature of media consumption has changed and will probably never. I sat down and tried to force myself to read a book the other day. I couldn't even read it. Brett House (43:44) Yeah. I know, Rio (43:55) But you don't just say, this is okay, but this is interesting. Brett House (43:55) Scott, was like, this a problem with me? Or is this a me problem? Rio (43:59) So I was having dinner with a bunch. Scott McKinley (43:59) I was like, how many pages do I have to get through till I get a dopamine hit, damn it. Come on. Brett House (44:04) This is not stimulating enough. Rio (44:06) Well, okay. But people will listen to those same people who will go to scroll TikTok. You're right. Will not read a long form article. They will listen to a two hour podcast. I think that it, so this is almost like the dumbbell effect where people prefer short form. agree with you. It's harder. ⁓ but I think there is a place for long for, but I just think the way people consume content has changed. And I think a lot of these, like, I, you're watching Lam, I love Lam, man. There's a lot of shows I watch and I watch them all the time, right. When they come out. Right. But, but yeah, but I think that. I haven't read a physical book in years, but I do read ebooks. So I think content consumption patterns have changed. Brett House (44:40) Yeah. So, so signal of noise is like the Clift Notes for the, the MarTech, DataTech, AI, right? We're like the dummy proof, just listen to us. You don't have to read anything. But, but so back to your point about saving what you said was sort of traditional television, or at least the model behind traditional television, which is sort of an open model that's supportive of all publishers that are, that are within that medium. How, do you see that? What are the next steps? How do you see that happening? Is this a full Xerund? Scott McKinley (44:44) Hahaha. Well, what can't, if we keep going down this path of pretending that an IP address represents a certain kind of person or it, will, I'll tell you what happened. Those companies will never get yield off their inventory because it won't perform because of the poorly targeted. Okay. So the money is going to, they will never get the yield. The next step will be the death now, which is already starting of. of traditional television, which is those companies will realize this in an effort to make revenue. They will start licensing their content to their competitors who can monetize it 20 times better. They, if they take half the take rate off of land man shown on prime with an ad tier, they're going to still make five times more than they would try to monetize that themselves. That's the nature of it. So I think we are, if we don't, if we continue down this path, we already hit the iceberg. Brett House (45:53) Yep. Scott McKinley (46:00) It's just rearranging deck chairs. It's not, it's going to all end up in the gardens. And I think that's really, really unhealthy. You're already seeing the consolidation. Like it's, it's crazy. The alternative is staring us right in the face, which is use that playbook guys, authenticate your users, authenticate your use, mandated, mandated. I was told by somebody it's funny. So we do serve some of these big TV folks. ⁓ we're trying, we're doing really good things to try to make this transition. And someone senior told me, Scott, you know, I ran that up the flagpole. Like, you you told me you should have mandatory authentication wherever you can get it for these reasons. And he says, I agree. And he ran up the flagpole. says, we're not going to do it. I said, why in the world, you know, what's going to happen? He's like, it's a, they said it's a privacy compliance issue. We don't have the infrastructure to create, to manage Opt-in. Yeah. And they're like, they're more afraid. Rio (46:46) What'd say? to manage consents, to, okay, yeah, that's a cop out. Brett House (46:57) Is that, yeah, that sounds like a cop out like a double opt in process is not. Scott McKinley (46:59) I don't, it's not a, I wouldn't say it's a cop out. I would say it's letting legal liability get in the way of existential survival. Like that's just a legal problem. You just got to, and a technology problem too, just solve it. You don't just say. Brett House (47:10) Yeah. Rio (47:14) Tools are out there, right? You can go to OneTrust if you just want to do cookie consent management or if it's a bigger consent management play there, there are consent management platforms. You can, I don't know, can have it stored in your CD. I don't know. There's no ways to do it. Scott McKinley (47:17) Yeah. Totally. Totally. Yeah, so maybe it was a little bit. I'm sure it was a little bit of a cop out. People just don't really want to change and they're bummed out and their careers are going to be over before this really hits. So who really cares? Rio (47:35) Well, yeah, Brett House (47:35) Yeah, well, and don't they... Rio (47:36) the legal department, they want to prove that they show their valuable rights. So they like to say no to things a lot. So I don't know. Scott McKinley (47:43) But Brett House (47:43) Do you think they do that Scott McKinley (47:43) here's- Brett House (47:44) like the back of the napkin math and sort of think, well, hey, how many, what percentage of our total audience reach back to the audience reach topic is willing to authenticate? And if it's a sub segment, let's say 10, 20, 30%, we're going to lose all of this extensible reach that we are promising our advertisers. So, you know, so then it's this downward spiral of like, well, then there's less revenue because we have less reach, right? So we're not going to go down that path. Is that part of the calculus? Great. Scott McKinley (48:08) Yes, yes. So let's go back from one second to newspapers and around this authentication thing. No, no. So I was only 10 years ago, I was at Ideafy, that company under ALC, which was a sub. And ⁓ I went down and talked to about a thousand, a conglomerate that operates, runs data for like a thousand publishers. Some very, very big ones that you definitely have heard of. ⁓ Brett House (48:12) We're going way back, okay? Scott McKinley (48:36) And I made this authentication speech and they, and I basically said, I trotted out that, stat that the monetization rate for the same audience in a cookie world versus an authenticated world is like 10 X, you know, really. And they, they said to me, Scott, ⁓ we hear what you're saying, but if we forced, if we put an off wall up, not even a, I said, don't put a paywall. You just want, you just want them to authenticate. That's the beginning of. Brett House (48:49) Yeah. Rio (49:03) Just ask them to give their email, basically. That's it. Scott McKinley (49:05) That's the beginning of success, you know, and they said, if we put an auth wall up, we've calculated that we're going to lose 15 % of our get ready for this page views, right? This is like. Brett House (49:17) my god. Yeah, e-e-e-e-Pageuse, Rio (49:19) So, think Brett House (49:19) what a solid metric that is! Scott McKinley (49:21) You Rio (49:22) about it, they were worried of 15%, not knowing this AI tsunami is going to hit them where they're going to lose 40%. We're talking to pubs, they lost 25%, 30%, 40 % of all referral traffic's already gone. Scott McKinley (49:28) Well, what... Also, I'm like, that's great because you're to monetize the remaining 85 % 10 times better than like the outcome is still an order of magnitude higher than you're earning today. And you're going out of business, guys. So that's that mentality. coming back to see to think about that, come back to CTV, their pushback was privacy. Maybe it was a cop out. Maybe not. Even since then. This is about a year and a half ago when I had this conversation. They've started realizing this. Like I set my father in law up on Pluto the other day. Rio (49:40) All right, 10 times. Brett House (49:46) Yeah. Scott McKinley (50:01) And it was, were very aggressive and made it very easy to authenticate, right? There was a good value proposition. So they're trying, but here's the other problem. The other problem is if I'm a seat, a streamer, a big, big program, a big TV network, and I've got my linear, I know that's dying out. That's bothering me out. I've got my streaming and a lot of it's authenticated. Most of it's authenticated. Maybe I have a fast property where I don't, I don't know if I want to, they should authenticate, but they're for some reason they're not. Brett House (50:28) Yeah. Scott McKinley (50:28) But the point is that Paramount Plus or NBC or Peacock, you are giving them your billing address, your email address, all these things. So there's a portion of that entire audience at a Paramount or NBC or WBD that is authenticated. You know what? They don't sell that as authenticated inventory because the buyers come along and they say, I want a hundred of the things. And they're like, we have 20 of the things that are really, really accurate because we have authentication. Do you want those? Literally the buyers are saying, I don't even care if 20 % authenticated. I just want IP addresses for the whole hundred things that I want to buy. it is a buyer issue. That's right. you can't blame, like you gotta be careful about blaming people who are doing things that look irrational or not in their best interest because it's coming down, it's the pyramid, it's the brands, it's the agencies and the KPIs were still running on these stupid reach KPIs. Brett House (51:07) So that's a buyer ⁓ issue, right? They're asking for the wrong thing. Yeah. Scott McKinley (51:27) Look, I was one of our several CPG clients. was at one of them the other day and they ⁓ told me that they've decided this is a hundred year old company. They've decided they realized actually that they have to stop applying the old calculus, everything, the economics of linear, what they did in linear, which they leveraged as a competitive advantage. They're so good at it. They were kept trying to use linear tools and tricks to move into addressable. And they finally realized these things don't play. Nothing is the same. We need to break our, literally break our models and create a whole new approach to use addressable, you know, for a competitive advantage for marketing. And they said, we don't know how to do it. It doesn't work for us. We want scale and it's really hard to get in CTV. It's the measurement, everything's messier. So that tells you how. Brett House (52:18) Yep. Scott McKinley (52:23) different linear and CTV are. just come, I want to close out this little topic with what streamers need to do is look at what the walled gardens are doing and say, it's a, that's the playbook. Authenticate your users, own your users. Stop, stop outsourcing everything to third party. Stop rolling. ⁓ Totally, totally. Brett House (52:37) Yep. Yeah, you're disintermediating yourselves right by doing that. Rio (52:48) But let's say this happens, open web versus logged in web. think that's the term I've heard thrown around a lot. So does open web just become a place for not only unauthenticated, but low quality, undifferentiated traffic versus this? If publishers did this, if they had some kind of authentication, you're right. They could have up to 10 times higher CPM. That's a huge difference. But if they all leave, what happens to the open web then? You know, does it still exist? Because you have the walled gardens, the logged in web. mean, it sounds like open web just becomes, it's a race to the bottom then, right? Scott McKinley (53:24) It already is a race to the bottom. you do the, if you do the circle, the overlapping circles of where it's, it's, it's less, I think it's about 17. Okay. think last time I checked, which was, think, Jones, Chris Cana Jones media puts out a thing that kind of maps the different channels and stuff like that. if last time I checked, think the unauthenticated open web was 17 % and losing a couple of points every single, there's no where that's going except down. Right. If you look at it, somebody just put out a stack rank. Brett House (53:49) Yeah. Scott McKinley (53:54) of where the money is going into the big media properties. there's a couple. So it's always the big guys, right? It's the obviously the Google's met what all those guys are the top. The only like open web things on there are like Pinterest, Snap, because these are little gardens. These are little gardens. They're heavily authenticated, mostly authenticated. They're building audiences that are sustainable and durable and they actually represent people. They're doing a much better job of of acting like a walled garden than the open web is. If you're a site that's just running on cookies or IPs, mean, I don't, what are you doing? I mean, it's a race to the bottom. Brett House (54:30) Yeah, and there's a big push towards, and we talked about this with the permitive, right, with Joe Root, right? And they're pushing towards just premium inventory. And to be premium, you need to have authenticated audiences. Even as they're smaller volumes of audiences, they're much more. Rio (54:46) And direct sell deals, like they're really sour and open exchange, right? completely. They're saying if you have, if you have the robust sell side signals, right. Which, which include all the things you just mentioned, the audience details, the, the, contextual signals, all of that, right? I mean, you, you can, you can command a premium and you should be supporting these publishers as they work directly. That was their thesis anyway. Brett House (54:49) Yeah, direct deals. Yep. Yeah, and how underutilized the data that they have, that publishers have, ⁓ is in the advertising sort of bidstreaming ecosystem. Once they push it through the DSP pipes, there's a narrowing of the pipes. All that rich consumer, they own the audience, is not being utilized. And so it affects yields, to your point earlier about Scott, because it's not going to perform, because you're taking a sub-segment of the data that publishers already have. And yet, some of them are still hesitating with going towards that authenticity. Scott McKinley (55:26) Yeah. Rio (55:33) Well, but Scott, where Scott McKinley (55:33) What's? Rio (55:34) do you see things going though? Just quick question. Like, so where do you do like with, if this happens, let's say, you know, publishers are looking, we have to get authenticated audiences. We're going to just force everyone to go through this path. Even if we lose two thirds, it doesn't matter. We're going to be 10 X on that, that one third. If that starts happening, like just the way inventories bought and sold may be changed. mean, cause again, Permutiv's thesis was it'll be more direct sold instead of majority being programmatic and like say most of that being biddable. You you could have more direct sold these. I mean, get ready to PMPs are still, you're still an auction, like does the format of these media, like the way media is bought and sold change and like, does this impact the way that you mentioned buyers? Does this change your behaviors? Do you think? Scott McKinley (56:19) I think that the trend for buyers who actually pay attention to what's happening and doing their best to assess value to different ways of buying in different channels, I think they will absolutely gravitate to where the inventory is trustworthy and credible. Okay, that's the first thing. then that usually correlates with there's some kind of ⁓ the chances of me reaching humans on that particular channel are much higher because they're authenticated. Like I think that's what's that because it's, you know, it's Brett House (56:35) Yep. Yep. Scott McKinley (56:48) It's just abundantly clear that throwing a dollar into a plot and a DSP who tells you they're going to buy a bunch of CTV is you might get three, five cents of value out of that. Really. It's, kind of that bad. I mean, on average. So I think that's what's going to happen. I think, I think the authentication story is inevitable. think either like even these publishers you're talking about, like the sort of race to the bottom. Like, ⁓ if you're a publisher that produces. actually good. You're not an MFA, you're not made for advertising site, you're not a bullshit recipe site that you've got to wait through 9,000. Lady, I don't want to hear where you were born. I want to see the recipe for this. It's an excuse to put in more ads. I think that website programmers, content producers who produce good content can easily earn authentication and they don't realize it. Brett House (57:23) Yep. Yeah. AI, AI generated was you say? Yeah. Rio (57:25) You're right. Scott McKinley (57:47) Like they can earn it. And if they can't, your stuff is so garbage. Sorry, you're, you're going to float around in the bottom and not, know, I don't know what, why people ever buy your inventory, but I think that everybody should try harder to create that value exchange with the user that earns the right to authenticate and delivers a good quality experience. You know, I'm a cyclist. used to, ⁓ I always watched the old cycling races and there's one company out of Australia called cycling news, which is the global source for everything pro cycling. Brett House (58:14) Yeah. Scott McKinley (58:15) And to Brett House (58:15) Yeah. Scott McKinley (58:16) me, that was like a litmus test. was like, during all this time where I'm realizing that authentication is the path forward for media, I was like, I cannot believe if they put an auth wall up and saying, hey, just we want to collect your email address to serve you better and give you personal, whatever it is, I would a hundred percent do it. And it took years and years years years to do it. And finally they did it. that's what for things looking forward for the open internet. Rio (58:35) No brainer, yeah. Scott McKinley (58:45) And even television, anything that's not in a walled garden already, for those enterprises to hope to compete in the new media world, they must authenticate their users. There is no other path forward. That's it. That's it. And I think once, as that increases, and I will say that again, the traditional TV networks, they've learned this and they're trying really hard to authenticate. The minute we get authenticated inventory, in the non-walled garden media world, now I can buy scale now. And it doesn't really matter if you buy direct or intermediary buying strategies are really about like how much scale do you want? Do I really want to do 14 contracts for IOs and do with 14? Sometimes you just want to buy it all that's available. That will work if what I can buy is authenticated. that's one thing that we're doing at Truthset. We're literally walking the walk around this whole subject. We're saying, Brett House (59:27) Yep. Yeah. Scott McKinley (59:44) The internet is a mess. It's not good for anybody. There's no one knows what they're getting. No one knows how many people are behind these IDs. No one knows how much fraud or so we want to, true set saying our approach to this is instead of starting with all this static, which might be some noise, which might, you might be able to squeeze some signal out of, which is the state of things. Let's start with all the people. Let's, let's recognize that there are 300 million adults or whatever the number is. Let's do our best to understand those people. Let's build up from that. Let's build audiences of the 12 million women today who might be, who are actually pregnant. Let's try to start from certainty. Like what does census say this audience is, know, instead of like, Hey, I'll sell you six million, you know, left-handed soccer moms in Kansas city, a DMA that only has 1.2 million people in it. Like that was the world. That's still the world that we live in. Brett House (1:00:32) Yeah Yeah. That's the probabilistic world, right? It's kind of a, it's like a model-based guesswork, right? Scott McKinley (1:00:42) Right. So what we're saying is, look, do that if you want, if you want people, if you want, like marketing works when a message reaches a human being who's got kids, a wallet, a car, a commute, a trip they want to take, not a device. A device is a... So let's start there. Let's allow a brand to build an audience with the certainty that this is the, these are the humans that I want to reach. And now let's push that out in the ecosystem. Ideally doing one-to-one. Rio (1:00:57) than just an actual human being, right? Brett House (1:00:59) Yeah. Scott McKinley (1:01:11) PIA matches with different destinations to pull out that 20 % of authenticated inventory, pull out another 18 % here, 12 % here, and slowly amass audiences that are big enough to justify a decent buy. That's what we're trying to do. We're trying to make it person-based, truly person-based. I think that digital and CTV should look like direct mail eventually, like eminently targetable, eminently measurable, like really, really certain. And also prices are going to go up. This is a world of, we're moving to a world of scarcity. The 20 billion buyable IDs floating around, very few of which are actually humans, much less the right humans, is gonna shrink. And it's gonna be a really, really good thing, except everyone's gonna struggle about prices going up, because prices will go up. Rio (1:01:54) Well, it's interesting to say, cause that's been Dr. Fu, we have had him on this podcast a couple of times. That's always been his argument is that so like you look at the amount of ad spend that's gone into digital, it's gone up by, you know, 10 X or something like that, right. Over time. And the inventory hasn't grown that much yet. The CPMs continue to collapse. Right. So there's just a lot of garbage inventory out there. if, if, if really, it's really, because it is a marketplace. If anything, CPM should have gone way up because you have more money chasing, not that much more inventory. So that's a good point you make. Brett House (1:02:27) Yeah. Well, and I love your point about the direct mail and, know, that digital is going to go the way of direct mail. mean, in the sense of like the authentication and the one-to-one reach that you're getting with direct mail. Do we still have Scott on the line? Rio (1:02:42) Yeah, the Gumiho lost him. Brett House (1:02:43) Well, hey, for everybody that, oh, there we are. He's back a little bit of a delay. Well, I, and I think we were just, were just, did you hear us? You, heard us, I'm assuming. Um, but no, we, we're yeah. Yeah. So Scott, this was in a great conversation. all a little bit over our next meetings. Yeah. Should we do quick hits or should we, uh, yeah, we'll do some quick hits and then we'll, we'll wrap up. um, this is probably just ask you a couple, a couple of quick questions. I got the first one. Uh, so what part of the identity ecosystem is. Scott McKinley (1:02:44) I'm back. ⁓ Rio (1:02:45) I'll be back. Scott McKinley (1:02:52) Yeah, I it. Rio (1:02:57) Do you to do quick hits or are we late? Brett House (1:03:12) most overdue for a reckoning. Scott McKinley (1:03:16) A focus on the quality of these linkages, which is the connection of one ID to another. People completely misunderstand that even if I have a quote, deterministic ID, someone gave me an email address, that's great. You can't do anything with that until you connect that email address to something else, to whatever it might be, to an IP address, ⁓ right? And those linkages. Everybody for some reason just assumes that those linkages are good. They're not good. Like the IP address linkage, we talked about 13 to 60 % average accuracy. I mean, you're throwing away potentially 83, 82 % of your dollars. So I think it's two answers. One is it's the linkages. People don't understand that data moves through these hops between identity spaces. And each time you lose a lot of fidelity, you lose records and nobody really pays attention to how important that is. The other part of the identity thing that I would highlight is what we've talked about this whole time, which is you need to authenticate your users. We're not 20 years ago where you were scared to handle PII. Like you should do that because look at the market caps of the other media companies who are doing that. They're worth hundreds of billions of dollars. Brett House (1:04:34) And they're not running into any legal issues, right? Usually. Rio (1:04:36) No. Brett House (1:04:40) To be determined. Rio (1:04:40) Or if they are, it's like a small potatoes relative to the revenue they're generating from it, right? Yeah, Google's had fines for privacy, but you look at how much money they're making, it's nothing. Scott McKinley (1:04:45) Yeah, I mean, you still have to meet you. Yeah, it's back to that conversation about that TV network. Like, would you rather just, you know, I don't want to do it and then go out of business or like figure out how to do it, figure out how to handle the legal liability, the technical challenge, which to your point, Rio, there are tons of tools that are built to do this. So that's what I'd say about identity. Rio (1:05:08) Scott, what's a bigger problem and what's more corrosive to the system? ⁓ Bad source data, bad modeling or just perverse bad incentives? Scott McKinley (1:05:17) Ooh, I mean, all three of those are kind of the same thing. I mean, they all work together. You have perverse incentives, which cause you to stretch your model to produce an outcome or a audience that you want. And then that turns into the first thing, which is bad audience data. So literally what you said unintentionally, it's daisy chain. It's the perverse incentives that drive the misuse of Brett House (1:05:33) Yep. Scott McKinley (1:05:46) technological tools, modeling, et cetera, which produce data that's not actually accurate and reliable and will produce a result that you expect. Brett House (1:05:54) So who's got the power to change, ⁓ to enforce standards to change this? Does it start with the brands that are actually doing the media buying? Scott McKinley (1:06:01) Well, you said two words that modify the answer a little bit. Yes. The brands are responsible for demanding transparency, ⁓ clarity, all this kind and they're starting to do it. Right. And it used to be that their agencies weren't so hot on that because once they dialed in a model, finally they can survive. They don't want to change anything at all. They don't want to do that. So, but all that's changing. think the agencies are stepping up. I think everybody is realizing that things have to change. And to your point, yes, it has to come. from the top. can't blame the data providers. You can't blame these intermediaries really, because they're just doing what they're asked to do. The other words you said though were enforced standards, which is quite interesting. I really do believe, I was at Nielsen, I came into Nielsen as a skeptic of their business. Like how are you making all this money off this panel? What the hell? And then I slowly did a 180 and realized how important independent measurement is and independent governance. And these things kind of go together. And what, what this new world needs, we talked about this at the beginning of the pod here, you really, this new world, this new messy, crazy, heavily obfuscated, very valuable world of CTV and addressable CTV, addressable television. It needs, it needs independent bodies, whether it's governing bodies, like a new version of the MRC. Brett House (1:07:03) Yeah. Good job. Scott McKinley (1:07:30) Or another Nielsen that maybe doesn't have the, maybe operates differently. It can't be Nielsen because that would mean they have to change their economic model because that will just not work. Maybe it would, I don't know, but we need independent. It's like when you go to a, you drive your car up to a gas pump to fill it up, if there's no octane rating on that pump sticker, you're not going to put gas in your car because it might be watered down because that week the... owner wanted to buy a boat and he watered down all his gas for a while. What that octane rating does is it gives the consumer confidence that an independent entity has said this gas is not going to break your car and it's going to give you choice about what grade of gas you want to buy depending on your use case, BMW or a rental car. So that same thing needs to exist in this new data ecosystem where buyers and sellers can agree If I'm to buy an identity for bretthouse.gmail.com, I need the confidence that he's actually a male, he's actually married or not, he actually has a cat, all these different things about him. So there needs to be some independent entity that helps bring trust into this data-driven ecosystem, much like Nielsen brought trust into the linear television market. Brett House (1:08:47) I think that's a great way to end it. Trust has carried carried through this entire ⁓ episode and Scott, Hey, thanks for the time. It was awesome to connect deeply. I'm yeah, I'm glad we could do this versus a 20 minute at possible because you know, it allows you just to really unpack a lot of these pretty heady topics. This was, this was a ton of fun and thanks for, thanks for joining us. And everybody that made it this far, deb, deb, deb.signalannoyce.ai. You can find us on YouTube, Spotify and Apple podcasts, as well as our newsletter. Rio (1:08:49) It is. This is great conversation. This is super fun. Scott McKinley (1:09:06) I appreciate it guys. Really fun. Brett House (1:09:17) which is going far and wide now and we will see you next time. Rio (1:09:21) forward to it. Thanks, everyone. Scott McKinley (1:09:22) Thanks guys.