Referee to Co-Pilot: Ad Verification in the Age of AI with Mark Zagorski, CEO of DoubleVerify

Watch on Youtube

Listen on Spotify

Ad verification is no longer just about blocking bad ads—it’s becoming a core input into how media is planned, optimized, and measured. In this episode of Signal & Noise, hosts Rio Longacre and Brett House sit down with Mark Zagorski, CEO of DoubleVerify, to unpack how verification is evolving in the age of AI.Mark brings a rare perspective, having led multiple AdTech companies across market cycles—from building and selling eXelate to Nielsen, to turning around Telaria, and now steering DoubleVerify as a public company. We discuss how that background shapes his philosophy as DV moves from being the industry’s “referee” to a more active co-pilot for marketers focused on outcomes.The conversation covers a wide range of timely topics: DV’s strong recent performance and what’s driving it, the shift from protection to performance and attention, and how acquisitions like Scibids AI and Rockerbox signal a tighter connection between media quality and business results. We also tackle some of the industry’s toughest debates—over-blocking and its impact on quality news, the role of AI in verification and optimization, CTV fraud and measurement gaps, sustainability and carbon signals, and how brands can invest confidently in credible journalism without sacrificing suitability.

This is a candid, operator-level discussion about trust, transparency, and accountability in modern media—and what it really takes to connect quality signals to outcomes in a rapidly changing ecosystem.

---

Read the full transcript bellow:

Brett House (00:01)

Welcome Mark Zagorski to Signal and Noise. It's great to have you on the podcast.

Mark Zagorski (00:05)

⁓ Thanks for having me Brett, good to be here. Hi Rio.

Rio (00:09)

Good to see you.

Brett House (00:10)

So for those that don't know Mark, Mark and I actually go way back to the beginnings of my career in ad tech at a company called, a little company called Exolate and sort of traversed the data marketplace, data management platform worlds before being purchased by Nielsen ⁓ and launching the Nielsen Marketing Cloud. And then Mark did a whole bunch of terrific things. People probably know him as the CEO of DoubleVerify, but in your own words, Mark, I'd love to have you share with our audience.

A little bit about your trajectory in MarTech, AdTech, and kind of what brought you to your current ⁓ five-year plus stint at DoubleVerify.

Mark Zagorski (00:46)

Yeah, yeah, no, I mean, look, I've been swimming around in this ⁓ internet ad pool since 1997, where I was with Modem Media, one of the first digital ad agencies. Some folks ⁓ who are long in the tooth may remember those folks, that crew. And ⁓ yeah, and most recently sold Exolate to Nielsen, ⁓ then ran a company called Talaria, small public company, which we merged with Rubicon to create Magnite and then.

Brett House (01:20)

yeah, formally tremor video, right?

Mark Zagorski (01:23)

formally tremor video,

Brett House (01:24)

Yeah.

Mark Zagorski (01:24)

It was Tremor, we sold off the ⁓ buy side and just focused on sell side, turned that to Talaria, and Talaria merged with Ruby, became Magnite, and then ⁓ I turned that over to Mr. Michael Barrett, who's done an amazing job with that business. And then ⁓ most recently in the last five years, IPO'd Double Verify, and that's where we are today. So it's been a really great kind of decade or so of selling companies, merging companies, and now IPO and companies.

Brett House (01:56)

That's awesome.

Rio (01:58)

Yeah. Well, your reputation definitely precedes you, Mark. We're thrilled to have you on, on the show and welcome everyone back to Signal and Noise. So again, this is Ria Longacre here with my cohost, Brett House. And as you can see, Mark Sygorski is on. We're going to talk into and dig into the state of ad verification. And I guess you could say Mark is really in the, CEO of DoubleVerify, he's right in the pilot seat, right? Mark, we're thrilled to have you here. And I just think this is such a very important. topic, especially like looking at ad verification and how this relates to things in age of AI, right? Anyone who's followed this space knows this has been a lot going on this year. DV, I posted some really good Q2 results. Stock bumped quite a bit. Looking at the numbers here, 20 % revenue growth, guidance went up, and the market notice shares went way up. And I wish I'd bought more stock before that. ⁓ Trajectory looks good.

But in a lot of ways, verification isn't just a victory lap. mean, there's been a lot of, I think, questions in the industry about where things are going. A few weeks ago, we had ⁓ a guest in our show, Dr. Augustine Fu, who many know he's been a sharp critic of brand safety tactics, not specifically of DoubleVerify, but just in general. There's been argument for him, as well as for many publishers, that some blunt keyword blocking can drain revenue from quality news and maybe even sometimes misreal fraud. ⁓ we want to address all these concerns, talk about them and understand what DV is doing, where the changes are being made and why DV is leading the pack and how it relates to this recent good news. We're also dig into some of the Double Verify's recent moves, including the Cybus AI acquisition, which I thought was really cool and one of the, I think a harbinger of a lot of the &A activity we've seen recently. In addition to the auction time optimization, and as well as rocker box attribution. And what these bets say about time trust outcomes and the whole age of outcomes thing. ⁓ think industry never sits still, lots of change, a lot going on, especially with AI and really a lot to talk about. ⁓ welcome here, buckle up everyone and let's dive in.

Mark Zagorski (03:58)

Yeah.

Brett House (04:14)

Yeah.

Mark Zagorski (04:14)

You really kept that narrow, Rio. We're really gonna keep this. We're gonna wrap this up in like, I think we got like five minutes and we'll be good. Yeah.

Rio (04:17)

Well, you did say anything was good.

Brett House (04:18)

There's a broad set of topics. You said you were available until like 5 p.m. Eastern, right? ⁓ So, you know, you guys have been, I've heard you say the word referee, right? And you you and I having spent, had some stints at Nielsen, you certainly know what that sort of neutral arbiter role is. ⁓ As, know, and I love that analogy for DV. Where do you draw the line between sort of neutral arbiter and sort of business partner ⁓ with brand clients, with agency clients without compromising your independence as an arbiter.

Mark Zagorski (04:56)

Yeah, mean, look, we have...one main constituency who we serve and that is the advertiser. Over 90 % of our revenue comes from advertisers ⁓ and our goal and not our goal, our mission is to be an independent ⁓ partner to them to help them assess the quality and efficacy of their media placements. There are many platforms that do what we do, but they're also selling media buying and selling platforms that do what we do, but they have skin in the game, which is they have a vested interest in making sure that transaction happens. We don't. We have vested interest in making sure that the data we provide the advertiser is accurate and helps them make the best decision possible. So, you know, there is no line in which, you know, we feel like...we have to cross to be independent or good. We're already there. That is our mission. If our advertisers don't trust us that we're doing the right thing, then they won't spend with us. And that is key. Our independence, the veracity of the data that we have, and our focus on ensuring that they are getting the right information to make the right decision is essential to who we are as a company and our value prop to anyone who works with us.

Rio (06:18)

Well, results speak for themselves, that's for sure. ⁓ Looking back to the outstanding results last quarter, street notice, stock bumped, what do you attribute some of this success to?

Mark Zagorski (06:28)

So, I mean, look, we started off the year saying this is a year of transition for us, and it still is. We are investing in solutions for both CTV and for social platforms because we believe ⁓ more and more dollars are flowing there, and there's less and less transparency. So we know that this year was going to be one in which there was going to be some variability. We started the year off saying that. We benefited from some active behavior that we had not with new products which are still being developed but with kind of selling our vision to new customers and upselling those solutions to our current customers as well. We had some really strong strategic wins over the last several quarters. Microsoft, Kenview, Charter, all of these companies ⁓ came to start working with TV and we were able to upsell them some new solutions. So we benefited from our focus on ⁓selling the vision of where we're heading as a business, bringing new customers into the fold and taking advantage of some of the growth opportunities in upselling products that have arguably been around for a while, but have not had penetration with some of these new customers that we were just onboard. So we had a good first half of the year. I think we will continue to show strong growth ⁓ based on the evolution of our solutions into other venues over the next several quarters as well.

Rio (07:58)

I've heard you talk a lot about moving from just brand safety, protection, and verification into more performance, outcomes, really looking at things more holistically. What solutions are you bringing or problems do you think you're solving for CMOs? And you mentioned the brands themselves that maybe even weren't solvable a few years ago.

Mark Zagorski (08:19)

Yeah, so you know the evolution from us being what is classically known as kind of a protection business to one that also kind of has performance as part of his mission is not as big of a stretch I think as some might believe. We started off by saying we're gonna help you take garbage out of the system, right? Ads that aren't viewable, fraud, ⁓ ads that show up in content or context that's not aligned with who you are as a brand. When you take that...

Brett House (08:47)

Yeah, it's an efficiency play. Yeah.

Mark Zagorski (08:48)

Yeah, when you take garbage out of the system, what's left is going to perform better.

Right. And I think that's the nature of who we've always been as a business. Now, fast forward to today, you know, we're helping you find media quality. But with the addition of side bids, we're helping you find that media quality, those impression quality, that impression cheaper. Right. And now with the addition of rocker box, we're going to help you prove that that impression or understand whether that impression actually drove a sale. Right. So if you think about it, it's the natural evolution of quality data, finding quality environments, finding that quality environment cheaper, and then proving that that quality environment or that impression actually drove a sale. So it's an expansion of our core, but we've always focused on efficiency, pulling garbage out of the system, and helping what's left work better.

Brett House (09:40)

The, yeah, side bids, that was an interesting acquisition. figure we could jump ahead to some of the questions we had about that, Rio, was, and that's, those are custom bidding models, right? Based on signals like viewability and attention and IVT and other, right?

Mark Zagorski (09:46)

Sure. Yep.

Rio (09:56)

Yeah, I almost felt like that acquisition kicked off the acquisition. We've seen a lot of them the last, let's say, year or so, especially in the AI space. I thought that was like a harbinger. Not a lot of people predicted it. I'd love to hear a little bit more about your rationale for making that move.

Mark Zagorski (10:13)

Yeah, I'll be direct. The initial rationale for that acquisition was more defensive than offensive. And what we saw was we were in the world of pre-bid ⁓ inventory being very binary. It was either viewable or not viewable. It was either brand safe or not safe. It was either fraud or not fraud. And we had this binary take on how our solutions were implemented. And what we...What we started theorizing is, well, viewability is binary, but when you start looking at the way IAB defines it and how other advertisers are defining it, there is actually a scale of viewability. what we've, like what the IAB has decided is kind of arbitrary, right? Anyone in the media knows it's kind of like there's this line of like, hey, the IAB says this is viewable or not viewable.

And I think some of our customers started feeling the same way, not around things like fraud or brand safety, but things like suitability and viewability, where there's a spectrum of, you know what, if I could get this ad significantly cheaper and it actually has ROI against it, maybe I'm willing to not look at viewability as binary, right? Maybe an ad that didn't meet the binary definition of viewability ⁓ actually still displayed long enough for someone to click on it so that it actually did do something. And if I can get that super cheap, then why wouldn't I buy it? So we started thinking like, you know,

Brett House (11:26)

Yeah. Yeah.

Mark Zagorski (11:45)

is this the way the world is going to go? Is all buying just going to be more on a spectrum, more on versus binary? And that was the initial thesis behind CyBiz, which is like, let's protect what we have so that we can...

Brett House (11:56)

Yeah, but it also moves you a little bit into the activation space. You and you have this auction time optimization, which I think it reminds me a little bit Rio of our conversation with Lou Pascales, was, as you can imagine, Mark, was a fascinating conversation. And he talked about the criticality of sort of signal, whether you're on the publisher side or the brand side.

Mark Zagorski (12:01)

Yeah. Yeah. Yeah.

Brett House (12:14)

like harnessing and capturing the signal, forget about just first party data, we're talking about third party data signal, and it's critical for success, for relevance, for content optimization, ad optimization, auction, real time optimization, is that really the goal is to increase the amount of signal that you're capturing? ⁓

Mark Zagorski (12:19)

Yeah. I think it's to not to increase the amount of signal, but to increase the usability of our signals in more applications. you know, I mentioned viewability as one, but for the vast majority of customers of CyBids, they're using their own signals, right? So whether it's someone like Diageo who has their own... ⁓

Brett House (12:39)

Yeah. Yeah.

Mark Zagorski (12:53)

you know, their own KPIs. The idea on my side is it's a custom KPI engine, right? And it builds an algo based on whatever that advertiser wants to focus on. It could be ⁓ a DV metric like attention, but it could be entirely their own metric. It could be sales, could be, you know, clicks, it could be anything. It could be viewability. But the idea there is... ⁓

Brett House (13:17)

So that algorithm allows you to suck in data from the client, whether it's their own first party data or part of their data ecosystem, which could be third party data partners through a cleanroom type thing. So you can bring all that in there and then customize to the variables that are driving the most impact. Is that?

Mark Zagorski (13:24)

Yep.

Rio (13:31)

Has that been combined as well with DV

Mark Zagorski (13:32)

right now.

Rio (13:33)

signals? mean, that sounds interesting.

Mark Zagorski (13:35)

Yeah, yeah, and if you kind of play this out, think about Rockerbox now.

Now we have RockerBox, which is pulling in attribution signals and all kinds of other transactional data that can also be optimized again. you're right. There's accumulation of more signal that we're getting through RockerBox. And then there's the application of that signal in more places through silence. And it's cool. So I think it makes a really powerful package for DV to talk to customers around.

Brett House (13:41)

Yeah, yeah. Yep. Yep.

And it sounds like it's a real diversification play. It's like limit your risk in terms of where you played and how you're being defined in market today, like competing against IAS, and broaden the scope of what your business actually does to support brand efficacy and effectiveness, right?

Mark Zagorski (14:24)

Yeah, know, look, think Rio, you mentioned earlier, it's kind of like what problems are you trying to solve? And I think, you know, CMOs and advertisers are always trying to solve the quality problem and they're trying to solve the performance problem, right? You know, is my media in the right place so that I can protect my brand and does it actually work? And if they can do that through a single platform and a single customer relationship, they'd much rather to do that than work with 15 other different places. We also...

Occupy a pretty unique spot in the fact that we're not just open web, right? We're not just CTV or just social we see data and transactions across Open web across mobile web across mobile app across social across CTV across audio Like and we do so and there we know reason why we can be there is because we're agnostic and we're not competing against any of those guys We don't compete against meta. We don't compete against Netflix. We don't compete against trade desk and that puts us in a really unique and agnostic position.

Rio (15:27)

Are these acquisitions changing who you're selling to to a certain extent, like agency versus brand, CMO? mean, has it changed that dialogue at all?

Mark Zagorski (15:37)

I think it's brought more folks into the discussion. When you start talking about optimization, now you're pulling traders in and media buyers in. Brand safety in many cases

Brett House (15:42)

Yeah, more functional folks from different areas.

Mark Zagorski (15:54)

was a discussion at the, sometimes at the CMO level, sometimes at the media level, and in some cases they actually had specific brand safety leads, right? You know, the brands had their own brand safety folks. This now has now brought in performance folks, analytics folks, ⁓ a big, it does.

Brett House (16:11)

Yep. Which changes the nature of the conversation too, right? You have to get a lot more technical, a lot faster with those data science gurus that are running sort of.

Mark Zagorski (16:19)

a lot more technical, but also a lot more bottom line driven too. Like show me what you're gonna save me. Show me what results you can help prove. And I think that's, again, think that's good. We wanna have those discussions with advertisers because, and we've said this before, it's kind of like you grudgingly pay for insurance. I'm sure neither of you, right, neither of you love paying for insurance. No one does, but you love like,

Brett House (16:22)

Yeah. I think I think I've heard you mention that before in public too. Yeah

Mark Zagorski (16:48)

for performance. Like, this is great, I want to work with you, I want that to happen.

Brett House (16:52)

You love when the insurance actually kicks in to support you in a time of need. You crashed your... Why am I paying this amount? Where's the value of this? Yeah. So for our audience, because we define side bids for the audience for rocker box. And this is sort of close to heart for probably the last five to seven years of my life at New Star TransUnion. Heavy duty MTA. We did obviously some of that in MMM and some of that at...

Mark Zagorski (16:56)

Yeah, it's good when it works, but when it's not even, I don't, why am I paying? This is, this is fire. Yeah.

Brett House (17:17)

Nielsen with visual IQ. We are commercializing that as part of the Nielsen marketing cloud as well as their home-built CPG MTA solution. But just for the audience and tell me if I'm wrong, Mark, know, Rockerbox is sort of a consolidated solution. It combines the capabilities of MTA, classical multi-touch attribution and marketing mix modeling and incrementality testing, right, to give you sort of a single source of truth. Now that's kind of a...

Mark Zagorski (17:19)

Yep. Yes. Yes.

Brett House (17:44)

positioning statement and I apologize for that but can you break that down for the audience in terms of⁓ the benefit, because I know the full capabilities of these types and they can get very heavy and very deeply algorithmic and very predictive and complex and hard to implement. ⁓ Can you kind of talk us like how the how rocker box sells into brands and agencies and how do you see it in the ecosystem compared with both the open source models like the Meridians and the the Robins from Metta as well as like the big players like TransUnion and IRI and ⁓

Mark Zagorski (18:10)

Yeah.

Brett House (18:17)

you know, there's a bunch of players in that space. Econometrics and... ⁓

Rio (18:20)

It's interesting too, Brett, you bring this up and I'm curious to hear what you have to say about it, Mark. Cause I felt like MMTA attribution in general, lots of questions about it. These are coming up, are people gaming attribution and with signal loss, like is there a future for MTA? And I saw a lot of brands start to really shift focus to more MMM, right? And maybe we should, which I think had kind of fallen off for a while, right? So I thought that was kind of refreshing. And think end of last year, beginning of this, incrementality, incrementality, questions about it for retail media.

Brett House (18:47)

yeah.

Rio (18:49)

How do we do this? I even did an episode on it on another podcast a few months ago, like just talking about like, what is it? Right? So, yeah, definitely an important topic. Love to hear what you have to say.

Mark Zagorski (18:56)

Yeah. Yeah, look, there is no lack of complexity in our space. once anyone feels like they get their arms around it, we just throw another thing in there and make it more confusing for advertisers. I mean, we're not even talking about like, know, ⁓ large language model fueled chat bots. And, know, when they get ad supported, how we're going to figure out how we attribute things to them. Yeah, they will all be ad supported. You can mark this. There is no doubt in my mind they will be ad supported.

Rio (19:10)

Yeah.

Brett House (19:11)

Yeah. Yeah, when, not if, right? When they get ad supported.

Mark Zagorski (19:29)

all CTV's ads.

Rio (19:30)

Well, perplexity just lost their head of ads, right? I Taz Patel left, is, ⁓ think it says more about the chaos there than their goals.

Mark Zagorski (19:37)

Yes. Yeah. But, but, you know, winding back to, to rocker box, think, you know, you're exactly right. Their focus and our, you know, our focus is trying to ensure that we can pull enough signals and run the right types of modeling that helps tie, you know, the the exposure to the outcome and do so in a way that is proprietary, unique to them, clear, and actually focuses on what the advertiser is interested in, And I think that is unique to RockerBox ⁓ where I think the connective tissue starts to make sense. this is something which is pretty, this is a data point which I actually just heard recently makes total sense is that we found that almost 30 % of ads that were being considered as being driving attribution were actually not viewable. exposure does not exposure and credit, i.e. cookie bombing from that ad was, yeah, was driving attribution. So, you know, to me that shows like, hey,

Brett House (20:50)

yeah, and they do the same thing in mobile like mobile mobile device bombing basically. Yeah

Mark Zagorski (21:00)

There's a desperate need to start with quality data, i.e. stuff that DV can assess is viewable or not fraud, and pull that into the model as well so that we can make what Rockerbox is doing even better than anybody else out there because we're doing that of, that filtering first. So Rockerbox sells directly to customers. They don't sell to agencies. They sell directly to mostly performance-driven advertisers. guys.

Brett House (21:27)

Oh, that's interesting. So in performance driven advertisers, like it doesn't matter the vertical necessarily. Some verticals are more performance driven than others. Is it more mid market or are you? Yeah.

Rio (21:35)

Maybe more lower funnel.

Mark Zagorski (21:37)

lower funnel. So retailers⁓ like Lowe's hotels, ⁓ Urban Alfreders, ⁓ Warby Parker, Weight Watchers, people that are really driving online transactions. That's kind of their specialty is that they because they've been able to really come up with a ⁓ model that shows true ROI on spend and then obviously apply that to the media mix model like where we can attribute revenue or attribute sales to this. Now

Brett House (21:42)

DTC.

Mark Zagorski (22:07)

this is where you should be shifting your dollars.

Brett House (22:08)

Yeah, exactly. Strategic budget allocation changes and shifts. Have you got, and I'd be interested to hear because one of the challenges that we certainly faced at New Star ⁓ was there was always a data hygiene and data unification issue, especially with the big brands. Like we always talked about how GM, who was a major client of ours, had 60 different data warehouses. So like the first six months of our contract with an MTA, MMM sort of hybrid was combining that data, getting that cleansed and enriched, et cetera, before you could actually begin starting to kind of power the models and getting more predictive and prescriptive in your... And so implementation times could go for 10 months, a year before they even see any results, outputs. I mean, is Rockerbox struggle with that or is it a different approach?

Mark Zagorski (22:54)

I mean, look, from my perspective, it is a new universe, right? Because I'm coming from implementations that are much faster, much lighter. know, we control the engagements in many ways and how we're pulling data in from third parties. It's based on just ad impression data, right? So that, yes, the ramp time is still considerably longer for a Rockerbox customer than it

Brett House (23:04)

Yeah. Yeah.

Mark Zagorski (23:23)

for like a straight DB verification customer. And the complexities of pulling that data in are not small. But I do feel like ⁓ this is a crew that's been around for a long time. They're digital first folks. These are not like legacy MMM guys from, know. ⁓

Brett House (23:31)

Yeah. From the old Nielsen days, they got rid of that business, Yeah.

Rio (23:47)

Yeah.

Mark Zagorski (23:48)

Axiom days, you know, like the, know, so, so they're pretty adept and they know, and, most of the customers that were, they're working with are digital first and digital transaction first companies, right? So the, so the data that they're pulling in is not as heavyweight as, as maybe what, you know, GM is like, you know, GM is looking at autos. They're, they're, they don't have online transactions per se.

Brett House (24:00)

Yeah, yeah. Yeah, something like a Bank of America or a GM.

Mark Zagorski (24:17)

know, Urban Outfitters and Warby Parker and a lot of these other folks, mean, like, a vast majority of their transactions are online. So the lift isn't as heavy, and I think it has to do with the kind of customers we've targeted as well.

Rio (24:29)

Yeah, plus GM, have the three tiers. It's a global company, right? They have dozens of agencies. I've worked in those big MMM projects. They could take six months to a year just gathering the data, aggregating it, building the models. I agree, really not easy. And maybe AI could speed some of that up, right, in terms of the data collection, normalization, et cetera. On the subject of AI, I'd love to hear, I don't know how much you can share, Mark, but love to find out about how is like...

Mark Zagorski (24:46)

Yeah.

Rio (24:56)

Any interesting things in a roadmap related to this? There's been a big buzz about agentic AI recently. Love to hear whatever you're able to share about that.

Mark Zagorski (25:05)

Yeah, look, AI is changing operations, it's changing our product road mapping, and it's changing the environment where our customers are sending ads. Like, you know, starting from the tail end there, we know, you know, there's so much AI slop in the open web right now. I mean, it's insane, you know. ⁓ We just published a paper on kind of AI-generated recipe sites, of which like,

Brett House (25:22)

Yeah, we talk about that all the time. Yeah.

Mark Zagorski (25:36)

vast majority of recipe sites now are all AI generated by creators and moms who are also, we find as they were telemarketing pictures and from a Russian website. So that the content environment has become really, really kind of nice.

Brett House (25:42)

Yeah. A wash with a lot of crap, right?

Mark Zagorski (25:57)

A lot of time, and we have, look, we've had customer engagements who've come to us and said, and like, this won't last, but they said, we don't want to be around any AI-generated content. And we're like,

Brett House (26:07)

Yeah.

Rio (26:08)

Good luck with that.

Brett House (26:08)

Yeah, see, good luck with that.

Mark Zagorski (26:09)

We're like, you're in finance, and by the way, 90 % of finance articles are AI generated. So that being said, but like there is something like we are detecting, for example, deep fakes in political content, right? That is something that advertisers don't want to be around. So it's created a more...

Brett House (26:15)

Yeah. Yeah, Donald Trump was supposedly dead like a couple days ago. Right? That was a deep fake. That was a big...

Mark Zagorski (26:29)

It's, I don't know, AI's, for all the good things AI is doing, it's doing a lot of bad stuff.

Brett House (26:37)

Yeah. Well, it's interesting, Mark. We did an analysis. It was like Krepsick and I, and he presented this in a webinar that we just did on sort of the content universe and how that landscape has changed when you talk about UGC and how in today's world, 50 % of all content that's produced is sort of short form UGC based content, which has much lower CPMs, different ad models, right? It's the creator economy. 30 % is sort of your Hollywood economy, like produced.

Mark Zagorski (26:47)

Yeah. Yep. Yeah.

Brett House (27:06)

Original content that we watch on Netflix and the like and then the 20 % that hasn't changed in like 20 years is basically live events in news Like live sports and live events and so that to my you know, the that UGC bucket Which it was a surprise to me 50 % and this was 10 % eight years ago

Mark Zagorski (27:14)

Okay.

Brett House (27:25)

nine years ago ⁓ is gonna be just a complete sea change in how we advertise against that content, right? Because it's so voluminous, there's so much of it. So how do you look at that? And this question, just sort of top of my head. ⁓ How do you think about ⁓ managing sort of ⁓ verification ⁓ efficiency spend against this just robust content source that's only gonna grow in volume?

Mark Zagorski (27:52)

So we've, mean, we've already encountered this. So you mentioned short form video and short form videos exploded. So we work with TikTok, we work with Reels, we work with shorts, we verify for suitability and safety and, you know, content, context and all of that stuff. And, you know, the shorter answer is we used AI to analyze AI. ⁓ So we do predictive modeling, for example, on short form video. So rather than look at every frame,

Brett House (27:58)

Yeah, yeah. Yeah, yeah, yeah.

Mark Zagorski (28:21)

of a video to determine what happens, we look at one frame per second and we determine that what's going to, you know, we predict what will happen in the next.

30 frames that happened in that second of HD video, right? So we're using AI to do predictive modeling. We're using AI to do labeling now. And we've always used large language models to analyze context, but a lot of those have been initially fed by what we call labelers, people, Either models. Now we're using AI to actually do the labeling as well. So, which has increased our scalability is increased our speed to contextualize content. So, I mean, we're just using this

Brett House (28:47)

Yeah.

Mark Zagorski (29:01)

tools that people are using to create content to analyze content as well.

Brett House (29:05)

Yeah, and it's all about like applying metadata in a sense to that content to label it, know, taxonomize it, orient it, right? It's interesting because, you know, and you've been talking about the big move towards ⁓ CTV and social media and social video and digital video. We did some analysis that showed him of the fight like by twenty twenty eight, we're going to have about a five hundred billion dollar advertising ecosystem. ⁓ Two hundred and twenty billion of that is video driven. Right. And and fifty billion is CTV.

Mark Zagorski (29:10)

Yep. Yep.

Brett House (29:33)

about, you these are approximates. mean, you know, you can look at eMarketer. 100. Yeah, by 2020, 120 billion is going to be social video and digital video, which is the, you're, you're, guys, you're going college now.

Mark Zagorski (29:33)

Yep.

Rio (29:36)

By 2028, that's what they're saying, Brett.

Mark Zagorski (29:48)

Yeah, he just started, just started.

Brett House (29:49)

Yeah, but I did too because we used to talk about our boys and there's the skiing adventures they went on Where do you go? I just just off the topic but Nice. Well, they need to get miles my son's in Syracuse film school yeah, so well, we'll have to compare notes offline but but

Mark Zagorski (29:54)

Yes. NYU film school, as a matter

Rio (29:59)

choice.

Mark Zagorski (30:03)

nice. Yes, we'll compare our unemployed children's notes. They don't have jobs because film is dead and everything's AI generated.

Brett House (30:12)

I'm like, moved AI, moved to digital video. But yeah, so it sounds like you guys are taking that challenge of digital video and just the sheer growth of that, right? My point being with our boys, they're on that, I bet they're on that six hours a day watching short form video on TikTok. It's incredible the behavioral changes you see in the youth around us. The youths.

Mark Zagorski (30:35)

The youths, the youths.

Rio (30:37)

Yes.

Mark Zagorski (30:39)

Yeah, now it is. again, it, whoa, I lost my light here. Hold on, I have to stand up, Brett. Come on, give me something here. There we go. Yeah, now it is.

Rio (30:40)

Take it to sleep.

Brett House (30:46)

No

Rio (30:47)

Yeah, the motion sensor has to pick you up again, yeah.

Brett House (30:53)

There we go.

Rio (30:56)

And that's the Beastie Boys picture in the back. It's not Backstreet Boys, which Brett thought it was earlier, just so we're clear.

Mark Zagorski (30:59)

Yes.

Brett House (31:01)

He was still out of focus, guys. I know the Beastie Boys. Very intimately.

Mark Zagorski (31:04)

⁓ Yeah, no, it's a sea change for sure. know, engagement is all over the place now with regard to like, you know, how consumers are being engaged and, you know, advertisers. ⁓It's an interesting universe for an advertiser to be in. We work with a ton of CPG advertisers, right? And they've always been the site's sound and motion people because we need to a connection to a brand. CPG is all about branding. how do you do that? 30 seconds was hard enough. How do you do that now in five, in which someone's going like this?

I think the rise of influencers, the rise of trying to break through in these short form video universes is really interesting. But also the fact that you are now, at least when I used to run a commercial in daytime, I knew that if I was going be on during soap operas, I knew what the content was. I felt fairly confident that I knew what was going to happen. And if something bad happened, then someone in the network would hear about it. Today it's like I have to be in short form. I have to be in these social media environments and what I don't want to happen is have my ads show up in between two terrible posts about something that's really uncomfortable for me to be around, right? But there's no other place for me to be and that's the challenge, right? I have to be there to sell product because that's where my audience is.

Brett House (32:27)

Yeah. Yeah.

Rio (32:33)

Well, Mark, looking at like responsible media like UGC versus news organizations, like how would you describe your role in helping brands navigate that? Because I know it's challenging for them, right? They need to be in certain places, but you look at some of the content that gets uploaded, social is some of this pretty horrific, right? I mean, go to AI, a lot of ad tech is on X, I'm there all the time. And you have to curate your feed in order to just not go crazy sometimes, right? Like how would you describe your role in helping brands navigate this, pick where to go?⁓ Yet at the same time, support credible and responsible journalism and avoid disinfo. How would you describe that?

Mark Zagorski (33:10)

Yeah, mean, look, the one thing that we are not in the business of is kind of directing ad spend anywhere. Like, you know, we are there to be the advertiser's partner and to help them navigate the places where they want to be. ⁓ You you mentioned news, and I think it's a place where, you know, we still just philosophically, as a company and personally feel like it's important for news to thrive. It's important for there to be quality news across the spectrum you know, ⁓ and for that content to be easily accessible and for it to be monetizable by advertisers. Advertisers want to be there. ⁓ So, you know, we launched the News Accelerator last year to help ⁓ publishers. We work with big publishers. mean, we work with folks like News Corp ⁓ and Fox and Reuters and ⁓ New York Times and The Post. You know, we work with a lot of big publishers and, you so the goal of that news accelerator was to help them better understand how they can package their content, to help communicate and educate advertisers on the quality and importance of news, and to help advance our tools, the technology that advertisers use to find their way to good content and make sure good being appropriate for them, and to stop using blunt tools like keyword blogging to ensure that they're being more sophisticated.

Brett House (34:42)

domain blocking, yeah.

Mark Zagorski (34:47)

And, you know, that's there's a lot of noise around how terrible keyword blocking is and how it destroys news. mean, but we.

Brett House (34:56)

Yeah, Luke Pascal is called a Roach Motel, right? It's like they go in, it's not, basically his whole point was, you know, these keywords and these domains that you've blocked, these lists go in and they don't come out, right? And they're not dynamic, they're not adaptable. They're not, they're not contextually predictive to say, hey, you might have blocked the Hillary Clinton keyword, that's kind of irrelevant in today's ecosystem. She's a grandma in Scarsdale. You know, how do you think about the nuances of that?

Rio (35:08)

Yeah, the lists are not curated for interrupt data frequently. Yeah.

Mark Zagorski (35:19)

Yeah, Yeah, I think there's a couple places we play there. The first is we don't recommend any advertiser use keywords as a solo way of blocking content. We've got categorization tools that are, you know, we have...

Brett House (35:32)

Yeah.

Mark Zagorski (35:38)

hundreds of suitability categories with multiple levels of sensitivity that are very sophisticated, that look at things like sentiment. That's what we promote, right? We should be using those tools. Keywords, you know, in conjunction with that, well, keywords right now represent 0.2 % of all the blocking that occurs. 0.2%, yeah. 0.2 % of all the blocks, the blocked impressions are blocked by keywords is to 0.2%. So...

Brett House (35:57)

really? 0.2 %? Wow.

Mark Zagorski (36:08)

That aside, even with keywords, what we've done is we've built new tools to allow keywords list to be, advertisers to be alerted when keywords are stale and not have been around too long, make it more seamless and easy for them to ⁓ update and upload keyword block lists if they wanna do that. ⁓

Brett House (36:31)

Can that be automated through some sort of AI decisioning engine that?

Mark Zagorski (36:34)

We have automated alerts. We're automating expiration periods. We're automating all of those things for advertisers. ⁓ This really comes down to advertiser preference. We can do everything from expire the keywords to ensure that those keyword lists are not stale and alerting them to that, to enthusiastically telling them you should not be using keywords on their own, you should be using sentiment analysis, you should be using our other tools. It really comes down to the advertiser and ensuring that they're the ones who are turning the knobs. We can't turn the knobs forward. We'll get as far as we can down that road. And look, there's a reason why we launched the news accelerator. And it was because we don't want like

Brett House (37:12)

Yeah, you can't control what they're doing. You can't make them more sophisticated. Yeah.

Mark Zagorski (37:23)

poor behavior to actually impair news growth. It's not a good thing. That's not why we're here. And look, the reality of it is that ⁓ without tools like ours, advertisers just won't buy news at all. That is the truth here. the reality is they don't need to buy it. There's an infinite amount of content. Advertisers don't need to buy content. Should they buy it? Yes.

Brett House (37:28)

Yeah, it impairs the open web, right?

Mark Zagorski (37:51)

because they think it's important for them to be there. And if they have tools that allow them to navigate around things that is more appropriate for them, they should leverage those tools. And that's what we want to promote is the idea of you can use a scalpel or you can use a meat cleaver. We'd rather use a scalpel and this is the way you use it.

Rio (38:09)

Yeah, we've talked a lot about news publishers in this show, Brett, right, which I think it's been good. Like Lou was great last episode and yeah, we have a great one with Jessica Ho coming up from Hearst. I it's interesting how like, and I get it, like complaints about keyword blocking and how it's impacted display. But I do think that like there's a broader problem about the business of news publishing and whether you go back to Craigslist or ⁓

Brett House (38:13)

Yeah, with Stephanie Lazer and the Lou, yeah.

Rio (38:38)

how that eviscerated classified ads, right? And how, you know, look at Google Meta's impact on siphoning traffic away. It's been a broad, just, I think, change of the business that they haven't responded very quickly to with not a lot of easy answers. So think blaming keyword blocking and, you in ad verification, brand safety is probably just, I don't I think it's missing the farce for the trees to certain extent.

Brett House (38:59)

Yeah, that's a good point. mean, it's a little, ⁓ yeah, it is a little, maybe a red herring, right? ⁓ Like, innovate or die, these publishers do have to innovate how they engage and attract brand dollars, whether it's their content models, their advertising models, et cetera.

Rio (39:07)

It's a factor.

Mark Zagorski (39:17)

No, it's a tough time to be a publisher.

You know, we want to help advertisers better direct spend there. I I think the existential crisis for large publishers right now obviously is AI chat bot summaries. I mean, you we already know it's impacting publishers. So there needs to be a universe in which they can figure out how to monetize that as well. Otherwise, you know, or create content that's unique. ⁓that isn't just fact-based snippets, it has to have a POV. And again, I'm not a news person, there are much better prognosticators here. But I think the reason why we talked about recipe sites, the reason why recipe sites are AI-generated is because they're just a list of So even the recipe sites are going to be summarized in a

Brett House (40:10)

Yeah, there is no perspective. There is no nuance or opinion or empathy.

Mark Zagorski (40:19)

and a summary and I won't even need to go to the AI generated recipe site. know, content needs a POV and I think that's what will thrive and survive in the open.

Rio (40:27)

That's a good point. So you look at like free press, they're getting acquired by CBS for what, 200 million, right? It's just basically like, yeah, it sounds like it is as of today. I mean, it may not, but like I was reading that it's going to Barry, that's Barry Weiss's organization, right? I mean, it's just basically like a bunch of blog. It's a podcast combined with, you know, some great content that they've been publishing on Substack, right? And that's being acquired, right? For, she's, I mean, said that good quality content.

Mark Zagorski (40:32)

Is that actually happening? Is that confirmed?

Brett House (40:38)

Barry Weiss

Mark Zagorski (40:40)

Yeah.

Rio (40:55)

people value it, people read it. mean, like, I think that proves there is a market for it, but it's probably, I the format's just changing, right? And publishers need to embrace that. And I mean, I don't think that anyone knows what the future is gonna look like to your point, Mark, but I think that things are changing and there is an appetite for good quality content that's made by humans.

Mark Zagorski (40:57)

Yeah? for sure. Yeah. Yeah, for sure.

Brett House (41:17)

So ⁓ maybe a little bit of a pivot is I would love your perspective and you may not be able to say a ton on this, but ⁓ Scope 3, another topic, Brian O'Kelly ⁓ launching this product initially to sort of, yeah, yeah, to decarbonize the advertising ecosystem, which we all thought, how long is that gonna last before he pivots back into sort of classical ad tech? It seems like his latest pivots are moving into the...

Rio (41:30)

Speaking of pivots, right?

Brett House (41:44)

the ad verification, IAS DV space. What do you see of that? you see them? Yeah.

Rio (41:49)

know, agentic ad verification, right? I mean,

Mark Zagorski (41:51)

Yeah,

Rio (41:51)

that's

Mark Zagorski (41:51)

well, there may be a third pivot here because I don't know if you saw today there were layoffs at Scope 3, published it. Yeah, and they want to go into agentic AI and I'm sorry, agentic media buying. So I think there's a, you know,

Rio (41:57)

Saw that, yeah. The sales team, right?

Mark Zagorski (42:09)

I think there's a path there where they're gonna go into actually purchasing media based on the signal that they have. I'm not sure. I I wish Brian all the best. We had worked with Scope 3 when they initially launched for carbon, for decarbonization ⁓ and had them as a partner, as a signal coming into our system and as a reseller. And they obviously pivoted into other areas which were quasi-competitive. ⁓ But, you know, I think it's, we will see where that ends up, right?

Brett House (42:13)

interesting.

Rio (42:14)

common agency. Yeah, I think it is current administration. Anything carbon is going to be tough. You know, that's for sure. Carbon intelligence and you know, it was interesting. I this is I guess this is another pivot and maybe they become an agency buying media. I don't know. But but I do think like the one point they made, though, was interesting markets that like instead of like maybe there wasn't that much of a market for carbon intelligence or carbon accounting. Right. But but like there is some there is a mark. Like if you take a lot of the crap out, you mentioned this before.

Mark Zagorski (42:39)

is a lot. in.

Rio (43:07)

It does lead to better outcomes and every impression is almost a gram of carbon generated. So if you're actually cutting out lot of AI generated crap, the garbage can focus in real outcomes and real performance. It should be better for the environment anyway.

Mark Zagorski (43:21)

Yeah, no, for sure. I like, I think the one thing everyone can agree on is that efficiency is better, right? More efficiency, more effectiveness, less spend, higher ROI. I think fewer middlemen, you know, and I think, you know, that obviously is our thesis too, which is if you can find quality cheaper that performs better and we can show you that it performs like that's.

Brett House (43:32)

fewer middlemen.

Mark Zagorski (43:46)

hitting the three legs of the stool, is quality, efficiency, and performance. I think that's what every advertiser wants, and that's good. And the output of that is economically positive and is environmentally positive. Those are great things that you can point to.

Brett House (44:08)

Yeah, and I love how you've just simplified it down to it's like we had three, it's like three tent poles, three pillars, and your entire business approach over the last five years has been moving down in those three buckets really, right? So looking ahead, and I'm sure you probably have to run in a little bit, ⁓ looking ahead, although I thought you said you had until 5 p.m. Eastern.

Mark Zagorski (44:18)

Yep. I did put a suit on just for you guys, right? That's what I am. I have nothing else.

Rio (44:27)

We appreciate it

Brett House (44:29)

You came out of a client call I think our client meeting somebody told me that ⁓ so looking at three years What's your what's your sort of bold prediction for? verification optimization You know a unified sort of quality to outcome OS which seems like you're really leaning into that theme Which is it seems critically important? ⁓ You know kind of a more federated ecosystem What's what is what's your edge at diverse at DV and what do you see the future looking like?

Mark Zagorski (44:55)

Yeah, I think there won't be an engagement without all three of those components as part of it. Like verification and optimizing the cost of that verification are just gonna be interlinked. You won't even have a discussion with a customer because, know. the idea of finding quality cheaper just makes sense. Like they're going to come together as one. ⁓ And that third leg of the stool, which is actually proving the effectiveness of that performance, is just a no-brainer. So when I look at the business three years out, I see a vast majority of our engagements ⁓ being across all three of those areas through an integrated solution, which we're building out right now. We launched our first product, Authentic Advantage, which is our first optimization in verification solution together. I see all those coming together into one platform. You know, the one thing that we didn't talk about and we could talk day is about the increasing rise of black box solutions. Without transparency and without any assessment of quality, only assessment of performance. And I think what we want to be is the counter to that, which is we're going to open the black box through transparency and verification. We're going to show where you actually went and we're going to help you understand what worked and what didn't work. And we're going to do that in a very independent way. Because I think advertisers love the simplicity of the black box.

Brett House (45:56)

Yeah.Yeah.

Mark Zagorski (46:21)

But everyone I talk to fears of what that actually means.