Episode 115: Brian O'Kelley on AI, Carbon, and the Scope3 pivot

Summary

Marketecture Podcast Summary: Episode 115 - Brian O'Kelley on AI, Carbon, and the Scope3 Pivot

Release Date: March 21, 2025
Hosts: Ari Paparo and Eric Franchi
Guest: Brian O'Kelley, CEO of Scope3

---

Introduction

In Episode 115 of the Marketecture Podcast, hosts Ari Paparo and Eric Franchi delve into an insightful conversation with Brian O'Kelley, the CEO of Scope3. Building on the momentum from their successful Market Live event in New York, the episode explores Scope3's strategic pivot from a carbon measurement and reduction ad tech company to an AI-driven data verification and brand safety powerhouse.

---

Scope3's Strategic Pivot: From Carbon Measurement to AI Ad Tech

Brian O'Kelley initiates the discussion by addressing Scope3's recent pivot. Originally focused on carbon measurement and reduction within the advertising industry, Scope3 has expanded its offerings to include advanced AI-driven brand safety tools.

Brian O'Kelley [03:47]: "It's definitely back to ad tech and it's kind of fun to come back to ad tech and everybody comes back to adtech."

This pivot signifies Scope3's evolution rather than a complete overhaul, integrating their core mission of sustainability with enhanced media quality solutions.

---

AI Integration in Brand Safety

A significant portion of the conversation centers on Scope3's new brand safety tool, which leverages artificial intelligence to provide granular and brand-specific content suitability assessments. Unlike traditional keyword and category-based systems, Scope3's AI can understand the context and nuances of content, offering more precise brand alignment.

Brian O'Kelley [05:11]: "You can actually just add, ask the question like, should Coke be next to this piece of content? Is that aligned with what the brand is trying to accomplish?"

Brian illustrates the tool's effectiveness by contrasting it with conventional methods that often result in over-blocking. For instance, a recipe mentioning white zinfandel at a Tupperware party might erroneously be flagged as alcohol-related, leading to widespread content blocks that don't consider individual brand preferences.

Brian O'Kelley [09:06]: "I built basically what we launched last week, which is a contextual analysis engine that could take this content and understand it for brands."

This AI-driven approach not only enhances brand safety but also reduces the inadvertent exclusion of valuable content.

---

Overcoming Traditional Taxonomies in Advertising

The discussion progresses to the limitations of traditional taxonomies in advertising. Brian argues that simplistic categorization fails to capture the complex and dynamic nature of modern content, advocating for AI-driven solutions that transcend conventional classification systems.

Brian O'Kelley [25:22]: "The world is really hard to put into simple buckets. And so the idea of yes, we need to do media plans, yes, they're hard, yes, they're important. I think there's a lot of examples where taxonomy is a useful way to talk, a useful way to communicate, but not a very effective way to operate."

By moving beyond rigid taxonomies, Scope3 aims to provide more flexible and accurate content assessments, aligning better with the nuanced needs of brands and advertisers.

---

Enhancing Ad Tech Platforms with AI

Brian elaborates on how Scope3's AI capabilities integrate with existing ad tech platforms, such as Amazon DSP and Google Ad Manager (GAM). Instead of building entirely new platforms like DSPs or SSPs, Scope3 focuses on embedding intelligence into the ad ecosystem to enhance decision-making processes.

Brian O'Kelley [15:52]: "We're giving this sort of unified intelligence layer where you can bring your intelligence one place and apply it to a whole bunch of different systems."

This strategy not only streamlines operations but also empowers advertisers and publishers with advanced tools to optimize their campaigns without overhauling their existing infrastructure.

---

Carbon Footprint Reduction through Media Quality

A pivotal theme in the interview is the correlation between media quality and carbon footprint reduction. Brian emphasizes that by improving media quality—through measures like reducing fraud and enhancing viewability—Scope3 effectively minimizes the environmental impact of advertising.

Brian O'Kelley [12:28]: "Anything related to media quality. If you make media quality better, you cut your carbon."

This dual focus on quality and sustainability positions Scope3 as a leader in transforming the ad industry's approach to environmental responsibility.

---

The Role of Humans in AI-Driven Ad Decisions

Addressing concerns about AI replacing human judgment, Brian asserts that Scope3's tools are designed to enhance, not eliminate, human decision-making. The platform offers complete transparency, allowing humans to oversee and validate AI-driven classifications and decisions.

Brian O'Kelley [22:50]: "You can see exactly what it's doing. So when we did this with Meredith, we actually looked at a lot of the classifications together and used it to improve the models."

This collaborative approach ensures that human expertise remains integral to the advertising process, fostering trust and reliability in AI systems.

---

Future Vision: Scope3 in Five Years

Looking ahead, Brian shares Scope3's ambitious vision for the next five years. The company aims to solidify its position as a leading platform that resolves media quality and brand safety challenges while significantly reducing the advertising industry's carbon footprint. Additionally, Scope3 plans to foster innovation in ad tech by providing a robust foundation for other businesses to build upon.

Brian O'Kelley [35:54]: "We're dramatically reducing the carbon footprint of the media and advertising industry by cleaning up supply chains and getting rid of a bunch of crap."

This forward-thinking strategy underscores Scope3's commitment to sustainability, quality, and technological advancement in the advertising landscape.

---

Conclusion

Episode 115 of the Marketecture Podcast offers a comprehensive look into Scope3's transformative journey under Brian O'Kelley's leadership. By integrating advanced AI into brand safety and media quality solutions, and maintaining a steadfast commitment to reducing carbon emissions, Scope3 is poised to reshape the future of ad tech. This episode is a must-listen for industry professionals seeking to understand the intersection of technology, sustainability, and advertising.

---

Notable Quotes:

• Brian O'Kelley [05:11]: "You can actually just add, ask the question like, should Coke be next to this piece of content? Is that aligned with what the brand is trying to accomplish?"

• Brian O'Kelley [12:28]: "Anything related to media quality. If you make media quality better, you cut your carbon."

• Brian O'Kelley [22:50]: "You can see exactly what it's doing. So when we did this with Meredith, we actually looked at a lot of the classifications together and used it to improve the models."

• Brian O'Kelley [35:54]: "We're dramatically reducing the carbon footprint of the media and advertising industry by cleaning up supply chains and getting rid of a bunch of crap."

---

For more insights and in-depth interviews, subscribe to the Marketecture Podcast every Friday and join their vibrant community at Marketecture.tv.

Transcript

[Sponsor Announcer]

This episode is brought to you by Zeta Global.

[Ari Paparo]

Do you know what it takes to transform marketing into a data driven profit center? Are you able to align the C.

[Eric Franchi]

Suite around your AI vision and strategy?

[Sponsor Announcer]

Zeta Global has the Playbook to help you get started.

[Eric Franchi]

Download Driving growth in the AI era today at ZetaCMO AI book again, that's Zeta CMO.

[Sponsor Announcer]

This episode is brought to you by Amazon Business. How can you free up your team from time consuming office tasks? Amazon Business empowers leaders to not only streamline purchasing, but better support their teams.

[Eric Franchi]

So they can focus on strategy and growth.

[Sponsor Announcer]

Free up your teams and focus on your future. Learn more about the technology, insights and Support [e-mail] Welcome to the MarketEcture podcast.

[Ari Paparo]

This is Ari Paparo and I'm here with Eric Franchi. So this was a big week for Market. We held our first event here in New York, Market Live on Monday. It was a huge success, over 400 attendees. We got great feedback. And from that event we have today's interview. So I sat on stage one on one with Brian O'Kelly, legendary Brian O'Kelly who has been making news with his new products, effectively his pivot of scope 3. From a carbon measurement and reduction ad tech company to a full on data verification algorithms AI ad tech company. Quite the pivot and we had quite the conversation. So you're going to hear it now. Eric, what did you think of the conversation? You saw it live, right?

[Eric Franchi]

Yeah, it was great. I mean, we're an investor in Scope three. I went to the event the week prior, so I was familiar with the material. But he did a lot of presenting the week prior and not a lot of answering questions. So I think this is the first time that he sat down and answered a bunch of questions, both on a product and a company level. And it was a really good conversation you led. And I came away with, you know, a bunch of new thoughts around their business. And I mean, it's a really exciting time.

[Ari Paparo]

And I have to mention, it includes a very unfortunate Irish accent that I do for St. Patrick's Day.

[Eric Franchi]

Yeah, that. It was St. Patrick's Day.

[Ari Paparo]

It was St. Patrick's day.

[Sponsor Announcer]

We recorded this.

[Eric Franchi]

I didn't even make that connection, but it was, it was an incredible Kieran O'Kane impression.

[Ari Paparo]

If you're listening, don't come after me, man. Brace yourself, brace yourself. It'll make sense when you hear it. Yeah, and I want to give advance notice. So marketecture has a new podcast coming out probably next week, is called the Brand Forum and We're doing in depth, one on one conversations between Mark Detric, co founder Jeremy Bloom and leading voices in the brand marketing area. So he's joined by Josh Palau of Pfizer and they are having great conversations. So look out for that. I think the first episode supposed to drop next week. We'll give you an update next week and where you can find it on podcasts. Nothing to do right now. It's coming soon. And with that, let's go to the Brian O'Kelly interview.

[Brian O'Kelly]

So please welcome to the stage scope three CEO Brian O'Kelly. We had a little logistical. I was told like three minutes ago no one could find you. And you were sitting right there. Yeah, there's some metaphor there. All right. So yeah, you reached out to me like a month ago and you're like, hey man, I have cool stuff cooking. And I looked at my stats and our podcast with you was our by.

[Ari Paparo]

Far top episode of last year.

[Brian O'Kelly]

And I was like, yeah, sure, let's do it. Incentives, you know, that's right. So last week you had a great event in the Upper west side. I attended. And you announced a lot of stuff. Like it was not a little stuff, it was a lot of stuff. And I'm not easily impressed, but I was impressed. So I want to start the conversation by going through the what? Right? So I think to make it simple, maybe it's not so simple. There were two products, there's one that's in production right now, that's a kind of a verification brand safety tool. And another that a little more roadmapy, a little more prototypey. That was a kind of an SPO on steroids kind of tool. Is that accurate to start?

[Sponsor Announcer]

I'll take it.

[Ari Paparo]

Okay, let's go first.

[Brian O'Kelly]

What is the brand safety tool do?

[Sponsor Announcer]

So Scope three Brand Standards is a reimagination of brand safety and suitability from a basic perspective that AI can understand content way better than really keywords and categories can. It's part of a broader thesis that we've always simplified things into categories because of both technical limitations and the fact that generally speaking, humans have done a lot of the trading and buying in programmatic. And so if you think about AI doing more and more of the actual. I guess it's not hands on keyboards. What is it called now?

[Brian O'Kelly]

Keyboards don't exist.

[Sponsor Announcer]

Yeah, whatever keyboards are in the AI context, you don't need that intermediate layer that if you think about understanding whether an article or post or video is really something that a brand wants to be next to, you can actually Just add, ask the question like, should Coke be next to this piece of content? Is that aligned with what the brand is trying to accomplish? It almost turns the brand itself into the category. So imagine that for a piece of content you get a yes for Coke, no for Dagio, maybe for Unilever. And so you actually get very granular, precise responses. And one great benefit of that is that imagine you have an article. We had Lindsey Van Kirk from Meredith on stage and she gave the example of a three bean dip recipe on Allrecipes where the author says, oh, this probably came from a Tupperware party where there was a bunch of people in the 80s drinking white zinfandel and it gets tagged as alcohol. And not only is it bad because a couple brands don't like alcohol related content, every single advertiser who says no to alcohol content immediately blocks this article. Whereas I think some advertisers might care and some might not. And so this nuance is what we can get with this bespoke, brand specific way of thinking about content.

[Brian O'Kelly]

Right. I think Jonah Goodhart gives the example of Time magazine Woman of the Year. Taylor Swift got banned because it had violent content because she said she was fighting for women's rights. There's really no better advertising problem than analyzing the text of a page. For AI, it is like the custom made problem, right?

[Sponsor Announcer]

Yeah, I mean it doesn't take a lot of effort to figure that one out. I think the other one is generating creatives from AI, which was.

[Brian O'Kelly]

Which I guess some people are investing.

[Sponsor Announcer]

A number of startups might be thinking about. But yeah, these are sort of obvious problem sets. And it's interesting to me that we didn't have more startups thinking about this problem. And the reason is because the easy part is the AI part. A lot of people think about like, hey, you know, we're going to make AI as a chatbot on top of a product. In this one, you have to build so much product around the AI, you have to think about. You mentioned verification. You have to have MRC accredited viewability and ivt. You have to have integrations into all these platforms. We actually had to buy a company to get 15 years of work in brand safety and verification integrations with all the major platforms. There's a whole lot of stuff you have to do just to get to the starting line and then you can use AI.

[Brian O'Kelly]

You launched a new product with integration into Meta. That's not an easy hurdle, right?

[Sponsor Announcer]

Yeah. Launched a product that took 15 years.

[Brian O'Kelly]

15 years that we just acquired. Would you Mind telling the story about the People magazine example on the plane? I thought that was really compelling.

[Sponsor Announcer]

When I was talking to this company, Adlux, which is this amazing French company we bought, I Talked to the CEO about all the work they'd done for 15 years to try to compete with DoubleVerify and Integral AD Science really hard as a French startup without a huge amount of resources to compete with these global companies. I said, don't you think that AI is fundamentally going to reinvent this business? He said, yes, but I don't have the resources to go figure it out. So I had a flight to Japan the next week. And so I downloaded about 50 articles from people.com and a couple LLMs on my laptop. And yeah, on the flight to Japan, I built basically what we launched last week, which is a contextual analysis engine that could take this content and understand it for brands. And to do that, I actually went to some big brands and I went to their websites and just copied the values and principles from their about us sections. I didn't type in, like, I don't want violent content. I wrote, I don't want check my ads to come after me for advertising here.

[Brian O'Kelly]

That's a good one. That's part of my prompt at analytics in parentheses, right?

[Sponsor Announcer]

No, you don't type that.

[Brian O'Kelly]

You don't even type.

[Sponsor Announcer]

That's too scary. You don't want Kristof. And then I put I don't want to get boycotted and these things the AI understood well. And then I also built a simple keyword search, and I actually had a bunch of keywords from some agencies and others, and I found that the keywords blocked five times more articles than my little AI attempt. And if you went down and looked at them, and we actually did this with thousands of articles later, the keywords are very inaccurate. They just don't find the right content. They miss obvious stuff. They get Taylor Swift, and anybody who goes after Taylor Swift is dead to me.

[Brian O'Kelly]

So risky business going after Taylor Swift.

[Sponsor Announcer]

Exactly. Exactly.

[Brian O'Kelly]

So sounds like quite a flight. You don't go coach to Japan, I'm guessing.

[Sponsor Announcer]

No.

[Brian O'Kelly]

All right. No.

[Sponsor Announcer]

The carbon footprint will be paid down by the amazing work we do.

[Ari Paparo]

All right.

[Brian O'Kelly]

All right. So I wanted to bring that topic up. We'll get to the second product in a second, but. All right, so scope 3 carbon measurement. Carbon tools for advertising. This new product, you can make an argument to do with carbon, but it seems like a pivot. And, you know, it's St. Patrick's Day, so I want to quote our friend Karen O'Kane who like the minute everyone knows, probably a lot of you don't know Karen. He's a very grumpy Irish guy. He's like me, if I was Irish and like angry. So. And he's like, so like the day you started Scope two years ago, however many years ago, he said something to me along the line of like, oh, fucking O Kelly. He is going carbon. That's a load of horseshit. He's gonna go back into ad tech. You just wait and so respond.

[Sponsor Announcer]

First off, I really wanna see his response to that impression. And Kieran, when you listen to this, I want to hear your ARRI impression.

[Brian O'Kelly]

I want to hear it.

[Sponsor Announcer]

I really want to see it. It's got to be posted. I will amplify. I can't wait. Well, a couple things. One is the idea of getting rid of fraud as a way to reduce carbon and anything that's waste in this industry, whether that's viewability, fraud, crappy supply chains made for advertising, all of that is total waste of both money and carbon. So my investigation into talking to Adlux was actually about IVT viewability. They had an attention product. The guy was like, we should go after all these sources of waste so that we can solve the carbon problem. And then I kind of get obsessed with the brand safety bit as well. So I think it's an evolution more than the pivot. I understand the feedback. This part, like the brand standards product is directly in line with what Scope three has been doing for the past three years. The other product is the crazy ad tech one and that one you can call a pivot if you want. It's definitely back to ad tech and it's kind of fun to come back to adtech and everybody comes back to adtech. But I do think that the brand standards bit is fully in line with our core mission.

[Brian O'Kelly]

Because you found in the approach to carbon that it was entirely correlated with quality, right?

[Sponsor Announcer]

100%. Anything related to media quality. If you make media quality better, you cut your carbon.

[Brian O'Kelly]

So is the carbon measurement product not just for you, but for the whole industry? Is that a real thing or is it taking a backseat?

[Sponsor Announcer]

I think that globally, right this second with the Trump administration, there's a new omnibus act in the EU that's pushing back their CSRD like the carbon measurement by two years and pulling it away from a lot of small and medium sized businesses. And I think there's a sigh of relief from a lot of CEOs who felt like it was really hard to do all the things to do making money and DEI and sustainability and I don't know what else they do, but they said it was hard. I think this is a backlash that will come back to the middle over the next three or four years. I do not think climate change is a joke. I do not think that we can all just hope somebody else solves it. Drill, baby, drill. Is not a good solution to our planet. But right this second, a pure play. Sustainability business is tough, and we're seeing a lot of that. Just because anyone betting on SEC regulations or things like that is not doing well right now.

[Brian O'Kelly]

Right? It's optional. It's a thing a brand can choose to spend time and money on. But the benefit used to be sort of regulatory or maybe financial, and now that's disappeared.

[Sponsor Announcer]

There was never regulatory requirements. There was always like maybe regulatory requirements. But Scope three, which is our name but also refers to value chain emissions, has never been regulated effectively. Like, we never had a single regulatory filing based on scope3.com, our company's data. What has changed is the idea that you can value signal that by doing sustainability work, your friends and kids will think you're great. We definitely do not see that as much. You see actually the opposite in the US you see people afraid to admit that they're environmentalists. But guess what? Nobody has changed in the last six months from loving this planet and loving the environment and caring about it to hating it. I've still not met a lot of planet haters. And if you are one, please come up to me later. If you don't like this planet, I can find you another one. Someone who wants a new planet, Mars, no atmosphere, have fun. But our business is based on money. It's always been based on saving people money and saving people carbon. And so the fundamental benefit of saying to someone, on average, when you work with scope three, you save 10, 15% of the money you spend on programmatic and 30 to 50% of your carbon. If you want to come do that because you save money, awesome. If you want to do it because you save carbon, awesome. I won't judge you either way, but you get both. And that doesn't change in any administration. Better supply chains, better media, better performance, better for the environment. That's all exactly the same idea.

[Brian O'Kelly]

All right, so the first product, the brand safety and product is using AI and it plugs into the buy side or the sell side, kind of like the competitive brand safety products. Let's talk about the second product, which takes a little bit more mind space to get around the way I interpreted the second product was AI driven. For now. It integrates into Index, it's on the sell side, it's a curation platform and it allows the AI to figure out what inventory should be provided to the buy side. Tell me why that is too simple or wrong?

[Sponsor Announcer]

You said to me on X that we ran out of names before we got to this one. And that's about right. We're calling it the agentic media platform. But the way I think of it is what if we had a media agent that could actually make a decision or give a signal on every single ad impression? Instead of building a DSP or an SSP or trying to do transactions, what if we could just insert that insight or decision into every impression, whether it's running programmatically or not? So just a little tip of the scales on every single impression? Yes. It could do SPO in the sense that we could look at an OpenRTB request and say, huh, that's a really long S chain, don't buy that one. That would be an SPO use case. But it could also do a lot of other things in the sense that we could do other media quality things. We could say, is it fraud? What's the viewability going to be? What's the attention potential of that impression? We can look at a whole bunch of different signals and make decisions on the impression anywhere in the supply chain.

[Brian O'Kelly]

Right. But what struck me was that if you take AI out, let's say there's no AI in this product, it would be one of the first standalone SPO products. Because what you're saying is you're enhancing the signals, you're curating, in a sense, this inventory on behalf of an advertiser and then they can use whatever DSP they want. And to my knowledge, there is a startup or two doing it, but there's no big industry player that is in that position.

[Sponsor Announcer]

Yeah, that's fair. And I think curation is one way of thinking about it in the sense that you're picking the inventory you want. But I think because we call curation something else, I think for me it's more like abstracting intelligence away from all the ad tech platforms and creating kind of a universal or unified intelligence layer where you can bring your intelligence one place and apply it to a whole bunch of different systems. That's the sort of vision.

[Brian O'Kelly]

Yeah, that makes sense. And that is an advantage of curation. So how does the AI play into this? So we're training an agent on a brand's behalf or an agency's Behalf to do this for them. Is that a good way of thinking about it?

[Sponsor Announcer]

I think that's where we'll go. I think in the medium term we will have true agents that are learning on every single ad interaction with a person and getting smarter. Where we are now is actually looking at different rule sets and data and learning to basically push back in rules because we can't really run inference on every single impression. And it would be environmentally crazy right this second to try to run a full AI inference in every impression. So what we're doing basically is saying, okay, how about audience? Let's look at all the audiences we could buy offline. Let's say when this set of audiences come in, what would you do? Would you buy it? Would you not buy it? Would you bid more? Would you bid less? So we're giving this sort of. We're almost simulating what an AI would do in real time and caching that response. We can do that for context. So we can say for this piece of content, this video, this ad, whatever, what would you do? And then when it actually shows up, we know what we're going to do. So we're getting the benefit of those called a custom algorithm or an AI in the loop agent, without having to incur all of that cost, both money and carbon, on every single impression.

[Brian O'Kelly]

Yeah, I was going to ask that about latency on the brand safety product. So you're using AI to classify the suitability of a given URL, but it has got to be cached at some level, right?

[Sponsor Announcer]

Absolutely, yeah. There's no point doing everything. But if you think about this, we can get through the entire inference process in about 100 milliseconds.

[Brian O'Kelly]

Right.

[Sponsor Announcer]

And we're doing this because we believe that all content will be fully dynamic. That the idea of a URL sitting around for years before someone changes the content, realistically, everything's going to be more like a chat. Your content, my content, your Internet, my Internet. They're going to look totally different. If we can't come back with a completely dynamic understanding of content and how it maps to different brands in the blink of an eye, we're not going to be able to do any kind of brand safety or suitability at scale in the future.

[Brian O'Kelly]

Yeah. Once again, it's a perfect use case for AI. It's just engineering to figure out how to do it fast, cheap, and get it into the ecosystem.

[Ari Paparo]

Yeah.

[Sponsor Announcer]

And just take that the next logical step. Like if you say to an AI, well, I'm having this great chat, you know what brands Might be really appropriate for this chat. I mean, that's really easy, right?

[Brian O'Kelly]

And so doing that on a static webpage is also actually easy. It's just a question about how do you bring it into the fore.

[Sponsor Announcer]

Yeah, that's a special case. Like, to me, the static web page is the easiest, easiest environment in the world. And if you can't do it on an easy case, which I would say is a challenge for some of the existing players, wait till we make it hard. I think we should be building for the hard case and the scale case and the future case and working back to where we are, even if that means some of these things seem like they're a little bit overkill. Someone said, why would you run inference in every ad? It's like, well, because all the content is going to be inference. And so if we can't actually handle dynamic content in the future, we're not going to be able to do anything.

[Brian O'Kelly]

So I want to talk about those that actually have their hands on the keyboard, the humans involved. It feels as though humans have sort of a, are maybe hesitant in some cases to hand over some of these decisions to an AI. In the end, if you have an existing solution, let's just say, and it is doing whatever job it's doing, good, bad, ugly, and then you switch to an AI and something bad happens, you have a lot of risk, you can lose your job, you could lose your client, things like that. Do the humans just become trusting of the AI, or do you think you're going to have to do A B tests and splits and compare and prove it over and over again?

[Sponsor Announcer]

I mean, if you think about this in the brand safety context, the A B tests are fairly easy because we have 15 years of, let's call them B tests, which is probably generous, more like F tests of how the current solutions actually have failed to do extremely basic brand safety tasks for a very, very long time. And those are not humans. You can't blame humans that they put in bad keywords or didn't categorize properly. What we're talking about is one cool thing about the Scope 3 product is every single article, every single video, it's all completely transparent. You can see exactly what it's doing. And actually for the publisher, you can see how every brand model has classified each piece of inventory. So we're not trusting that it does the right thing. There's no black box. You can actually see it. So when we did this with Meredith, we actually looked at a lot of the classifications together and used it to improve the models and we actually prompted differently in some cases to get better outcomes. So I think you're saying, well, humans are critical to understand whether the models and prompts and systems are working. And we aren't doing that in the current system. We should have been doing this for the past 20 years and we haven't been. So I feel like this is an opportunity to get human intelligence into the machine and system in ways that we haven't before.

[Brian O'Kelly]

Yeah, I doubt that many users of a given brand verification system could tell you what the error rate is, both alpha and beta type errors. It's not a well known thing. So when it comes to humans, humans have sort of intuitive ways of looking at information, taxonomies being one of them. Probably everyone in media has at some point said, well, this is a good ad for sports and this is a different ad for news. And you've been writing about this. You wrote just a blog post today on the subject, or something along the lines of taxonomies being dead and advertising. So I'd love to hear your thesis about taxonomies being dead. And then the second half is, how are the humans going to react?

[Sponsor Announcer]

Well, a couple years ago we were doing a project with a large holding company and we were trying to help them understand the carbon footprint of all their different media. And we asked them for the taxonomy of channels like, you know, search, social, display, video, and they had failed to have a single taxonomy. They could not decide whether online video was part of video TV was it display, was CTV and BVOD the same thing. What about ott? The number of different people we asked for a simple taxonomy of channels and the number of actual channel taxonomies we got back were just insane. And so if we can't come up with a taxonomy for something as simple as which channel is this thing? I think you can see the problem. The world is really hard to put into simple buckets. And so the idea of yes, we need to do media plans, yes, they're hard, yes, they're important. I think there's a lot of examples where taxonomy is a useful way to talk, a useful way to communicate, but not a very effective way to operate. And so to your second question, how will humans respond? We will continue to simplify. Humans are not the most brilliant intelligent form we will ever come up with, but it's all we got. And it's one reason we have things like unconscious bias, because it's really easy for us to put people into different buckets or anything into buckets. That's what we're good at because maybe we're tribal and social in ways, you know, you see rustling in the grass and you assume it's a lion, not, I don't know, a kid wandering around with a bunch of bubble gum in his mouth. That would be very strange. We wouldn't expect that. And so we will always be taxonomy people. We will always be ontological, we will always live in this sort of strange shadow reality. But it doesn't mean that's the only way we can operate in media.

[Brian O'Kelly]

Yeah, I get that. But there are. Here's a counterexample. So Privacy Sandbox. There are topics Product used to be called Flock. When the first version of it put consumers into anonymous cohorts that had no names. The cohorts were just a string of letters and numbers. And maybe they meant sports enthusiasts, maybe they didn't. And people basically freaked out, how am I going to use this data? Scientists love it, but ordinary people, practitioners, thought it was terrible. They're like, ah. So the new version of topics has sort of more human readable names, ieb, taxonomy like names. But that's actually kind of illogical. If you're an AI talking to other AIs, the string of letters and numbers should be sufficient. So do you think the advertising community can deal with, like, just trust the AI? It's not sports, it's a string of letters and numbers. And you don't need to worry about what it is.

[Sponsor Announcer]

Just following your logic. You know, Privacy Sandbox is crushing it.

[Eric Franchi]

Right?

[Brian O'Kelly]

Well, they're a wonderful sponsor of our event, so I. Oh, excuse me, I'm sorry.

[Sponsor Announcer]

Privacy Sandbox is freaking awesome.

[Brian O'Kelly]

Thank you.

[Sponsor Announcer]

Privacy Sandbox.

[Brian O'Kelly]

Well, I was just making the point about the reaction to that particular technology.

[Sponsor Announcer]

I know, but we're stupid. Humans are dumb. We are neither. One was the issue at hand. Right. The issue was, was Privacy Sandbox integrated deeply into systems. And if you think about all the ways that those topic names or letter number combinations flowed into the systems we used to buy, had that been integrated deeply into Coke I for magic AI trafficking and it drove 25% better performance. Everyone would be thrilled and celebrating Privacy Sandbox. And all I'm saying is the depth of integration and all the other things that make systems work or not is what's going to make something like that succeed. And when pundits, and I'm not a pundit, but many of you may be, write stupid things like, I can't handle these letters and numbers and Google changes it for us and it's still bad, who do you blame?

[Brian O'Kelly]

All right, I'm going to do a little Bach personality question. So I noticed that when you built AppNexus after right media. Right Media had a lot of problems. I think it didn't clear payments. You had to have every network, had to work with each other, all this sort of stuff. And then you built AppNexus that really.

[Ari Paparo]

Solved all those problems.

[Brian O'Kelly]

Then AppNexus had a really tough go at it to be the execution layer. You were fighting the trade desk and you're fighting Google and ad serving and all that stuff. I noticed your new ad tech products do everything besides transact. Is there a pattern of some kind here?

[Sponsor Announcer]

I don't know. I think my goal is always to build new value wherever possible. I feel like a lot of the transactional pipes are solved. If you look at something that has. All IB Tech Lab does is try to standardize transaction pipes, that just feels like a well solved problem that doesn't need a lot of innovation. Right. If you think about where we haven't seen successful scale curation has not been solved. Custom algorithms have not been solved. We do not have scalable ways to push intelligence into the ad tech stack. So I don't see any reason to build another dsp, another ssp. We may need another ad server, but I don't really want to build an ad server. And it's not clear what medium you'd be building it for at this point.

[Ari Paparo]

Are you a bidder on GAM if.

[Brian O'Kelly]

It goes up for auction?

[Sponsor Announcer]

No, because GAM's a web ad server and I'm, I'm a seller of web.

[Brian O'Kelly]

Now Freewheel, we could get together on that one.

[Sponsor Announcer]

Let's do it.

[Brian O'Kelly]

Yeah, let's do it.

[Sponsor Announcer]

But I just think there's not enough white space in those areas for innovation. If you think about it just from a purely entrepreneurial perspective, you want a space that's massive, but where you can build a defensible moat. I would ask all of our friendly startups, when you thought about your businesses and you're thinking, I want to go after a trillion dollar space. The biggest, juiciest, sexiest, hardest opportunity that I can actually build into. Why did none of you build an SSP or a dsp? I mean, come on, guys, those are the big, juicy, fun businesses and they're all looking. I know you built a dsp. Everyone's built a dsp, but everyone sold them too.

[Brian O'Kelly]

It's a really good business.

[Sponsor Announcer]

I'm just saying. It's just that's not where the exciting opportunities are. You've Got to find Blue Sky. So I think for me, that's where, yes, Right Media was the first ad exchange. And so we didn't know all the things Right Media sold to Yahoo. Yahoo didn't fix it. We had to fix it. But I would look at AppNexus as a right media 2.0, whereas I look at scope 3 as completely blue Sky. We're trying to do things no one's ever done before. So it's more like right media. In 10 or 15 years, we'll be here and you'll be telling me all the things I did wrong at SCOPE three. And I'll probably be like, well, fine. But I'm doing Scope three, Nexus four. It's like trying to fix all the problems I created with scope 3. This is the only thing I know how to do. Make problems, fix problems.

[Brian O'Kelly]

I'm picturing us in a picture of you and I in a nursing home being like, I remember that time we did.

[Sponsor Announcer]

Yeah, but we'll own Freewheel. Okay? We'll have Freewheel and each other. It will be a great nursing home.

[Brian O'Kelly]

All right, so you brought up custom algorithms. So how soon does your Gentex AI start bidding?

[Sponsor Announcer]

I mean, I don't think we'll ever bid like a dsp, but in terms of influencing bidding decisions very quickly, I think that's a hugely interesting opportunity. And so we announced last week that we're going to be integrating into Amazon dsp. And one of the first use cases is to take some of this insight and not just bid back yes or no, but actually bid back with a propensity or some kind of weight to say, hey, good impression, bad impression. I think that's hugely interesting. Imagine a custom algorithms platform that is completely independent of where you execute. You can run it in curation, you can run it in a dsp, and my favorite use case, you can run it in gam. We can actually push those decisions into GAM as key value pairs. And so for the first time, you'll be able to run custom algorithms on an ad server. I think that's incredibly cool, and it's why we don't need to build an ad server, because we can actually make GAM smart by plugging into the integration framework.

[Brian O'Kelly]

Or do you have a name for that header bidding or something like that?

[Sponsor Announcer]

Like that. I like that. I'm as God's love, they do it.

[Brian O'Kelly]

All right. There was one little demo you gave last week where I was like, what the fuck is that? You said something like, okay, the publisher has a floor price and maybe it's too high and maybe it's too low.

[Ari Paparo]

We don't know.

[Brian O'Kelly]

So me as a AI agent, I'm going to offer a different floor price. That'll be a soft floor.

[Ari Paparo]

I know I got that wrong because.

[Brian O'Kelly]

I didn't understand it. So please tell us about this strange floor price product you have.

[Sponsor Announcer]

Oh, my gosh.

[Brian O'Kelly]

Okay, dumb it down because, I mean, I really have no idea what it was.

[Sponsor Announcer]

Okay, so we all know that deals have floor prices. You could have a fixed floor where you just say, I get 17 bucks, you pay 17 bucks, that's it. Or 17 can actually be a true floor where if you bid more, you could actually end up paying more than the floor price. Right. And you could have a competition where different people pay above that. Now, I guess the question is, could you have a PMP with a floor of a penny? Well, of course, that's just an open auction. And so somewhere in that range we have value creation. So if I advertise a deal at a penny floor, can I set a soft floor where I say, you can bid whatever you want, but I'm only going to accept bids above $5, $10.

[Ari Paparo]

Whatever disclosed to the buyer or not could be.

[Sponsor Announcer]

Often they're not.

[Brian O'Kelly]

Often they're not.

[Sponsor Announcer]

If you look at auction theory, if you go to an auction, there may be a stated reserve price. When you go and buy, I don't know, a $10,000 painting, they may say the reserve price is 5,000, but they may actually not take it. That may just be a hint of what they think it's worth or what they want you to pay. There's a lot of games being played in our industry and elsewhere in auctions to try to signal to buyers what things are worth. All I'm saying is that we've always had a lot of these soft floor and other floor price capabilities, whether it's in the ad server, in the ssp, it seems to me that it's reasonable to make that something that we could decision. If I have a reason to believe that impressions are worth more than the floor price, shouldn't I be able to push back on a DSP that's trying to aggressively bid shade? Because really what we're talking about is what's the distribution of value? If I'm a buyer and I'm willing to pay $50 and the publisher is willing to take pretty much anything, what's the right clearing price? Is it $30? Is it $25? Is it $20? If I have an insight that could tell the publisher that there's willingness to pay from the buyer. I should probably go push that price up as high as I possibly can. It's just another way of thinking about yield management in a world where we have disparate buyers and sellers to push back against some of the bid shading algorithms that we're seeing from the buy side that are hurting publisher yield.

[Brian O'Kelly]

Right, so the product is a. APPLAUSE wow. The product is a publisher facing AI driven floor adjustment. Is that one way to put it?

[Sponsor Announcer]

It could be.

[Brian O'Kelly]

It could be.

[Sponsor Announcer]

I mean, the whole point here is it's not really clear who we serve anymore. Right. You've got so many different intermediaries playing in these transactions. If you're a curator and you're trying to make margin, this will make you more margin. If you're a publisher trying to run direct publisher curated deals, this will make you more money. If you're a buyer trying to go find the lowest possible price for inventory, then you're going to go find yourself falling into some pretty low quality invent. Unless you have media quality floors, which is one of the features of this platform where we can actually build these complex trade offs between quality, price and scale. What we're trying to do is give the tools in the middle of the transaction. The SSP is the heart of the transaction environment to actually make these choices. Do you want more margin? Do you want lower prices? Do you want more scale? We need these tools.

[Brian O'Kelly]

All right, let's finish it up. What's Scope three in five years.

[Sponsor Announcer]

Wow. I'm hoping that we are similar to what we are today. Post last Thursday, which is a platform that is solving once and for all some of the industry's media quality and brand safety problems in a way that's beneficial for both publishers and advertisers. That's really important. Two is that we're dramatically reducing the carbon footprint of the media and advertising industry by cleaning up supply chains and getting rid of a bunch of crap. And third is we're enabling a huge amount of innovation in ad tech that we haven't seen in a decade because we've built this platform for people to build innovative businesses on top of the Scope 3 platform.

[Ari Paparo]

All right, let's call it there.

[Brian O'Kelly]

Everyone give a big round of applause. Brian O'Kelly.

[Eric Franchi]

All right, we're back everybody, with the refresh. So you heard the Brian O'Kelly Ari Paparo interview from Architecture Live. It was epic. So we're here now a couple days later with the news of the week. So we have to start with the marketexture live conference because it was the biggest news of the week. What you got, Ari? What are you thinking? Are you feeling good?

[Brian O'Kelly]

Are you feeling okay?

[Ari Paparo]

Couldn't have been better. We sold out. We had over 400 people actually show up. There was great vibe, great energy. We got feedback that was extremely positive on our post event survey. It was our first event. We're certainly going to be doing more of them. And what we want to do is not just have another event, but make it a event that has its own kind of Persona and has its own feel to it, which is very hands on, no nonsense, detail oriented. Not as technical as maybe the IAB Tech Lab or pre bids events, but not as kind of high level and strategic and financial as some of the other events like the excellent Ad Tech Economic Forum that was this week. Also a lot of the same people, but different feel different, different take.

[Eric Franchi]

Yeah, we'll get to the Ad Tech Econ forum in a second. So I spent the entire day at Architecture Live. It was amazing, man. Hats off to you and the rest of the team. My impression was that where a lot of conferences, oftentimes, you know, it's as much about the networking and the time you spend in the lobby and having meetings outside of the content and certainly that that was happening to a great extent. The sessions were packed, so much so that my partner Josie, Joe Zawadzki, who is usually a lobby lurker.

[Ari Paparo]

He is.

[Eric Franchi]

I thought he left because I couldn't find him. And then he appears at the end of the day. I was like, I thought you left.

[Sponsor Announcer]

Where'd you go?

[Eric Franchi]

He's like, I was in sessions the entire time. So like, hats off to you with the programming. What was, I mean, besides the Bach interview, which again was really good, what were some of your highlights?

[Ari Paparo]

Yeah, so I think the Bach interview, everyone was really anticipating the session that you ran with Brett Wilson. The Startup AI showcase was interesting. We had seven companies up, giving very short pitches. Some of them came in a little flat, but some of them really killed it. So I just love seeing new products, even if, even if I'm not an investor or advisor or anything, just seeing the new products. What did you think about that session?

[Eric Franchi]

Yeah, I thought it was good. I think, you know, just like with being, you know, reflective, I think we could give the startups in the next one or should we do the next one more guidance on, you know, do a product demo or film a product demo, you know, versus, you know, hey, you got three minutes to do whatever you want. But you know, I see this stuff every day. The impact that it had on the rest of the audience who don't see these things every day was really tremendous. So I think it's a nod to just having more innovation, more showcases at these events, because this is something that people don't see.

[Ari Paparo]

Absolutely. And I think we didn't do them a great service by giving them enough guidance on what we wanted to see. So a couple of them came in with really nice demos and a couple of them were more like sales pitches, and the sales pitches fell a little flat. The other thing, the other conversation that made a lot of waves was the Perplexity conversation. So we had Taz Patel, who's the head of advertising there on stage and then got covered in Ad Exchanger, and I think Morning Brew also covered it. So we made a little news with his statements.

[Eric Franchi]

So I mentioned this when you first announced him. So I've known Taz for a long time. He mentioned in 2005, slinging pop unders.

[Brian O'Kelly]

Yeah.

[Ari Paparo]

Was that for you?

[Eric Franchi]

I think it was, yeah. Way back when he was a seller at interclick, which in 2005 was fiercely competitive with Undertone. And that was the market. Like, yeah, Interclick, Undertone and specific media. Like, we were just flooding the the web with these pop unders. And it's incredible to see the career that he's had. And I'm so excited with what he's working on at Perplexity. And interestingly, there's a juxtaposition, maybe we'll talk about it later with some recent comments by, you know, Sam Altman about his views on advertising. But to see them leaned in is like, great, because we are truly in this, like, you know, entering the post web era and Perplexity is, you know, going for it. And I thought the Innovid rebrand, Innovid official brand, that was a cool way to kick off the day. Like, you got an exclusive.

[Ari Paparo]

Yeah, Innovid. For those who aren't paying attention. MediaOcean had purchased innovative closed maybe a month ago. And at our conference, they announced that the surviving brand name would be Innovid. So Flash Talking is going to be retired as a brand name. And they talked about their integration. And we had Bill Wise and the president on stage and they gave us some good insights and we had a funny video to launch it. That was a fun thing to do.

[Eric Franchi]

Yeah, agreed. One more question on Architecture Live, then we'll move on. What are your takeaways from the Bak interview?

[Ari Paparo]

Right, so Brian O'Kelly, he announced his products the week before at his own conference. And when he announced his products, obviously he was spinning it the way he wanted to. So I think I was the first critical interview that had been done. The number one takeaway was when I pushed him on whether it was a real pivot away from carbon. And he said, yes, it was. And that I appreciate the honesty because effectively what he said was that he's still an environmentalist, environmentalism really matters. But that the market for carbon seemed pretty lackluster, that there was an imperative to measure and reduce carbon in the advertising industry, and he is pursuing it as a secondary metric, you might think, against quality. Because I think we have said this, and he said this, that quality and carbon are highly correlated because the fewer hops, the fewer badass bed websites you're on, the less carbon you use. So it's a pivot, and I don't see any other way to dress that up.

[Eric Franchi]

Yeah, no, I think that's fair. And credit to him for addressing it as such. Let's move on. Before we talk about the news of the week, just shout out to Tom Triskari and Rob Beeler for an excellent event, the Ad Tech Economic Forum. Yesterday, you were there in the morning, and then you seemed to take off.

[Ari Paparo]

I bailed around 2pm I. I'm just really busy. I got things to do, man. These podcasts don't make themselves well.

[Eric Franchi]

I stayed for the entire day, and at the end of the day, they announced an award for. They use the term influencer, which I thought was. I don't know, I. Influencer.

[Ari Paparo]

Did you win it or did I win it?

[Eric Franchi]

I won.

[Ari Paparo]

You won it. Oh, my God.

[Eric Franchi]

So, my man, Ari, you are the ad tech influencer. But I have.

[Ari Paparo]

Oh, my God. So for the podcast audience, Eric is holding up this giant crystal ball. It looks like it came right out of Donald Trump's coffee table. It's gold, it's a bull, and it's made of crystal, and it's like two feet wide. And this is gonna be decorating the Franchi household until Mrs. Franci throws in the garbage by accident.

[Brian O'Kelly]

Congratulations.

[Ari Paparo]

This is the first time seeing it. I'm very excited for you.

[Eric Franchi]

Thank you. I'm going to change my LinkedIn bio and my Twitter handle. I just wanted to discuss it with you first, given that you are the ad tech influencer.

[Brian O'Kelly]

Wow.

[Ari Paparo]

Okay. Well, I think we're gonna have to mix it up a little bit. I think we're going to have to. I'll be a guest on your podcast moving forward. Who's the co host.

[Brian O'Kelly]

Who's the host?

[Ari Paparo]

Who's the co host?

[Eric Franchi]

Oh man. Anyway, congratulations.

[Ari Paparo]

Is this the highlight of your year? I mean it must be, you know.

[Eric Franchi]

It was, you know they, I've helped them in the background just, you know, promote the conference and you know, kind of tickets and stuff like that. And I work at the intersection of ads and finance. So you know, in my little world this was, this was. I guess I'm kind of a guy.

[Ari Paparo]

Very impressive. Okay, thank you.

[Eric Franchi]

Okay, let's move on. So April 1st is April Fool's day. It is also the day that the Netflix ad suite NAS is going live. So again, Ad Exchanger had some coverage here from some comments Nicole made and they are launching NAS with their in house ad server. They built magnite as the SSP and integrations with DV360, the Trade Desk and Xander. So this is probably going to be a big moment for Netflix. What you think?

[Ari Paparo]

Yeah, I think we had previously covered that it was available in Canada and I made some joke about my girlfriend in Canada is the old thing and the Netflix people didn't appreciate that I got a little DM on that subject. But I think this is hopefully going to be a non issue for them really. It's moving from one ad server to another which is hopefully something nobody notices. And I don't think there's new news here because DV360 and the trade desk and Xander all can currently bid on Netflix. So I'm excited for them. Hopefully the work that's been done by John Whitacom and team will pay off.

[Eric Franchi]

The real question is are they going to dial this up in terms of just more ads, more impressions and will this become a bigger part of their biggest business and a bigger part of the the story that's the thing to watch over the course of the year.

[Ari Paparo]

Yeah, they've been pretty conservative it seems with the amount of ad load and the amount of ads shown to different tiers of subscriptions. Unlike Amazon, they did not put ads by default for most consumers. Netflix is obviously pushing more and more into live events. They just had the John Mulaney special that was like Live from LA every week. It appears that they're very consciously moving one step at a time towards more live events, making sure it all works, making sure the ads work. So I think that that's a really big part of this.

[Eric Franchi]

Yeah, that's a really good point. There's some controversy around UFC and it not really going very well with Disney, ESPN and you know, there's some Rumors that Netflix is going to pick that up, which will be huge.

[Ari Paparo]

Well, the thing about live events is that you can insert ads even for the paid subscribers. Right. Because there's breaks and there's nothing to put in them. So there's no problem with putting ads in. Whereas on, for the on demand stuff, the paid subscribers probably revolt if they had ads.

[Eric Franchi]

Yeah. Speaking of. And on the other side of the continuum, Ars Technica, I believe, had an article this week about Roku. And on the completely other side of the spectrum, apparently Roku is experimenting with having autoplay video ads before the main screen lights up. So you don't have a Roku tv, right?

[Ari Paparo]

I don't, but I've used them here and there.

[Eric Franchi]

Yeah. So you turn on the TV and the first thing you see is, you know, just the home screen.

[Ari Paparo]

Right.

[Eric Franchi]

So apparently they're experimenting and I give them credit for just experimenting with formats like this, like running an ad after you turn on the tv, before you get to the main interface to start to make your selections. It's a time of experimentation. I don't, I don't hate it.

[Ari Paparo]

Sure, it'll be interesting to show. I think one difficult challenge here is to measure if it causes sort of device churn over time.

[Eric Franchi]

Of course.

[Ari Paparo]

Right. Because I mean, maybe you live with it because you already have the device and then when it's time to upgrade or get a new tv, you say, ah, I don't really want that ad laden thing. I'm going to switch to an Apple TV or some other device. So that's probably got to be what they're thinking about.

[Eric Franchi]

Yeah, that's a great point. But I do think that it's your tv, right? You know, it's a five to seven year commitment just taking it on and off the wall. It's such a hassle. So I think they'd need to be like super aggressive.

[Ari Paparo]

Yeah, well, Roku devices aren't always hard tied to the tv. They dongle consumers and they also have consumers who have TVs with multiple kind of navigation options. There are some devices that are default Roku and probably consumers would stick with it.

[Eric Franchi]

Yeah, yeah. Okay, that's fair. Hey, back to just LLMs and perplexity. So this week's interview on strategy was with Sam Altman, the founder of OpenAI. Obviously OpenAI is ChatGPT and it's a long interview as, as these things go. And there's actually an advertising section, so I don't know if you had a chance to read it. We'll certainly put it in the in the newsletter and he's non committal but hoping at least his what he's saying is hoping that, you know, they don't serve ads. I'll pull a couple quotes. Ben says, from my perspective, when you talk about serving billions of users and being a consumer tech company, that means advertising. Do you disagree? Sam says, I hope not. I'm not opposed if there's a good reason to do it. I'm not dogmatic about this, but we have a great business selling subscriptions. It's just interesting given how leaned in perplexity is that ChatGPT OpenAI is on the other side of this. What do you think? Do you think this is just inevitable?

[Ari Paparo]

It really depends on consumer behavior because the perplexity bet is that the end result is something closer to what we currently think of as search, where you get an answer and you want to follow up with multiple sources and links. Whereas the OpenAI bet is that the answer is the answer and you never have to go anywhere else. And depending on those sort of modalities you might have different opportunities for ads. In the perplexity case, it is more advertising friendly because the consumer is looking to go elsewhere and it would also probably lend itself to more of a click model. Whereas the OpenAI model, if you're getting a complete answer, you may want to have ads in that answer or some other way of doing that. Also, OpenAI has just got a huge head start at getting consumers to pay. So if you're getting $20 a month, you don't really need to push on the advertising envelope very much.

[Eric Franchi]

Yeah, that's a great point. I will say that at the Ad Tech Econ forum yesterday, this idea of we're being post open web, what do things look like five, ten years from now dominated I would say more than half of the conversations. Everybody's thinking about it. Publishers, obviously, venture investors, everybody knows that it's coming. And you know, outside of cool things going on with companies like open ads and firsthand we've yet to really see what this might look like. So it's I'm fascinated with this whole idea.

[Ari Paparo]

Yeah, there was a chart. I don't know if it was Brian Weiser's chart or based on his data, but there was this chart showing the breakdown how you get from a $1 trillion ad market down to biddable programmatic at something like 40 billion or something which is 4%. And it was kind of interesting how go from remove linear, remove social, remove search, remove everything else and you end up with open Web at something like 80 billion. And then of the 80 billion, only half is biddable. So it's a big number. 40 billion is a big tam already. But it's, it's maxed out or declining. And if you have an amazing mousetrap, but turns out there are no mice left, it's not as exciting.

[Eric Franchi]

Yeah, yeah, that's a really good point. It's incredible. Brian also was probably the most hated person this week and I saw him at the event yesterday and you know, he's running around prognosticating doom. He revised his U.S. ad spending forecast just with the political climate, the tariffs. Basically he Revised it down 20%. So he says non political ad spend is going to be growing at 3.6% for the year, step down from 4.5% which is significant. We should have him on the pod. He is talking like cataclysmic events here, which is. It's unclear because these are scenarios. This is not a given.

[Ari Paparo]

Yeah, but no one knows better than him. Maybe there's someone who knows better than him, but it seems like no one knows better than him the shape of the overall advertising spend and how it breaks out among the giants. And there's increasing worries there's going to be a recession and recessions hit advertising first. So I think we have to clearly look at that. The other thing, and there's some Twitter discussion about this, is when there's a economic disruption like a big recession or something, there often are long term effects, not just short term effects where advertisers or consumers change their behavior. I mean, if you just look at the growth of streaming because of COVID that's one example of sort of a macro trend. But a recession could make advertisers think hard about their media mix and reduce on legacy channels, never to come back. So that's something to keep your eye on.

[Eric Franchi]

Yeah, that's fascinating. The last real recession was the 2000 financial crisis.

[Ari Paparo]

Yeah.

[Eric Franchi]

Were there any like media or consumer changes that affected media and marketing you can think back on? Because that was also in the middle of the rise of digital and right before mobile. So it's a little hard to draw a correlation on that one.

[Ari Paparo]

Well, yeah, I think you're right. I think that was an accelerant to social against traditional. That was the real golden age for Facebook. They were growing astronomically year over year. It may have just been inevitable anyway though.

[Eric Franchi]

Yeah, good point. One more, then we'll call it here, what's going on with DV360 adding household targeting, you added this one last night?

[Sponsor Announcer]

Yeah.

[Ari Paparo]

So this was not even a press release, just a blog post by Google, but people noted it. So DV360, the DSP owned by Google, has traditionally been behind in CTV. And the main reason they've been behind is because of restrictions on their ability to use certain data points like IP addresses on a tv. You don't have cookies and you may not have a login unless you're on YouTube. So IP address is sort of the lingua franca of CTV. And a couple months ago, Google put out a sort of statement saying they were relaxing the restrictions on IP addresses. And everyone said, oh, that's interesting. That's probably CTV related. And now the other shoe has dropped where they actually announced that they're adding household targeting to DV360 and they explicitly say they're using IP addresses. So good for them. To be honest, a lot of people have obscure stuff like that. And this is a little bit of a table stakes feature. But for DV360 to have the table stakes puts them in a good position because they've been so far behind on this stuff. And this allows you to do, for example, like CTV retargeting and data targeting, cross device targeting and a lot of other things that previously were difficult on DV360 outside of YouTube. So they're arming up a little bit to get a little more serious about ctv.

[Eric Franchi]

Makes perfect sense. All right, why don't we call it there? You deserve some rest after this big week, big lead up to it, man.

[Ari Paparo]

No rest for the weary. I appreciate everyone who made it to the event and hopefully we'll keep up the high quality standards as we do more of them. So thank you very much for that.

[Eric Franchi]

Absolutely. Until then, we will see you next week, everybody. Bye.

[Sponsor Announcer]

Bye.

[Ari Paparo]

Thanks. Bye.

[Sponsor Announcer]

Thank you for subscribing to Market. New interviews are added every week at Marketing and your favorite podcasting app.

[Ari Paparo]

Thank you for listening to the marketecture podcast. New episodes come out every Friday and an insightful vendor interview is published each Monday. You can subscribe to our library of hundreds of executive interviews at marketecture tv. You can also sign up for free for our weekly newsletter with my original strategic insights on the week's news at News Market tv. And if you're feeling social, we operate a vibrant Slack community that you can apply to [e-mail].