Afternoon. I'm Laura Martin. I'm the senior media and internet analyst here at Needham, and I am here to welcome Tim and Chris Vanderhook, who are the CEO and CFO of Viant respectively. Tim is the co-founder and CEO, and Chris is the co-founder and COO of Viant Technology, ticker symbol DSP, where they have built the second largest independent DSP after Trade Desk. Their Household ID and recent acquisition of IRIS ID and TVision add unique data assets, which they are driving market share gains from competitors. Before Viant, they co-founded the streaming service Xumo, which was ultimately acquired by Comcast, and they made the high-profile acquisition of Myspace from Fox in 2011 for just $35 million after Fox had originally bought it six years before that at $580 million. Welcome you guys.
Thanks.
So nice to-
Thanks for having us
have you here.
We are shrewd businessmen.
You are a shrewd businessman. You always have been.
Somehow MySpace is still in our bio.
I know, but I really like that statistic.
We're going to change that. Yeah.
That you buy for $35 what Fox buys for $600.
We still lost money.
Yeah.
You didn't lose near as much as Fox.
That's true.
That's true.
Everything's relative in the end, right?
Yeah.
Anyway. Okay, let's start with recent earnings and go from there, and I'm going to try to find questions that pit you against each other, which I've never been successful doing. The first one I want to start with is leadership, because you guys are really different kind of leaders it seems to me. I want to start with generative AI is changing, it's transformational. My question is, what are the different things in leadership you have to do during a transformational technology change that maybe weren't relevant three or four years ago when things were more just execution excellence?
Yeah. I don't know if it's how we operate, which is I'll just talk how we operate. We get into the weeds. We're a part of the team that's developing it. We understand it from the ground up because we want to be able to speak to it at a very granular level because we're always out talking to customers, talking to investors.
Yeah.
We want to make sure we know what we're talking about. Even if I look at some of the newer things that happen, like ViantAI, which we released 2 years ago. That prototype was built by myself after the engineering team said it's not possible, the PM said it's not possible. I basically got pissed off and built it myself and then delivered the working prototype to them. I think that's the difference of a founder-led business versus that of the managerial class, some of the bigger companies. The founders have ultimate optimism and I don't really care about optimism. I just want to get the job done. For me, I like to dive in there and actually see if it's possible for myself. If I hit a wall, well, I can't blame anybody else.
Okay.
I.
Chris, what about you?
just to tag onto the other part of Tim's story there. When we went public, we had this vision, a product vision of autonomous advertising, which was powered by artificial intelligence. This is pre-OpenAI, ChatGPT, all that, right? We had a lab, and we had a guy who ran that lab, and he was an applied scientist. What we've came to figure out over three years was, I don't know what he applied any of the science to because we weren't getting any product out of it.
It was more research.
It was really that frustration. I think just as founders who run businesses, you typically find the good ones are, to Tim's point, getting into the weeds. We felt that we needed to be first. We wanted to be first. That's what precipitates change in an organization, and that's why Tim, over a week, basically spent his vacation doing that. On the other side of the coin, just as you run a company, what we do constantly, we're paranoid all the time, we look around corners, and we want to know where the boogeyman is. We want to know what could be lurking. We sit in the RTB infrastructure. We're in programmatic.
RTB, real-time bidding.
The real-time bidding infrastructure. We look at what could be possible threats. Oftentimes those threats become opportunities. When you look, AI, only recently in the last few months have we had this AI apocalypse, but it's something that we've been looking at, that we think it's an opportunity. At the same time, what we don't want to do, although we don't think it's going to be en masse in the next year or two, we do want to be prepared for that. We want to be able to enable our clients for that.
Just getting back to the leadership question. Three days ago, I just told my assistant, "Cancel all my standing one-on-one meetings with all the department heads. I don't care about that anymore. I want to meet with these five people next week, and next Friday when I meet with you, I'll tell you the next five meetings I want in the next week." I think as a founder, you're addressing the biggest problems in the company. You need to give the business context to the people working on the hardest problems. You need to see it through and make sure that things get done. Versus before, just the general delegation of tasks. I just don't see that working in the future. If I have an IR problem, I'm meeting with Nick, the head of my IR all week long.
Solving the questions that he's getting asked, understanding how to frame it, on what the opportunity is versus the risk.
All of these types of things. The AI story I gave you, it was pure product and engineering at that point with the data team. To me, I just view it as what's the biggest problem? What's the biggest need for the business?
We just attack it because the business can't move forward until you solve that problem, and then you're able to actually move the business forward.
One of the big controversies that's erupted on this stage today, because you guys are, I think, the 10th AdTech CEO team I've had on the stage is, do you think that generative AI agents will be built on new rails or on the programmatic rails?
The answer's both. It's on the programmatic rails in the near term. I think in the long term, it's on the agentic rails.
Okay.
The question is timing. In zero to three years, very low probability that it's on AdCP, the new agentic rail.
Yeah, MCP
MCP. That's on the new rails inside of There's a founder who called it AdCP. Now, I shouldn't have said that, yes. MCP as a way to communicate, absolutely. The question is 0-3 years, low probability, 3-5 years, mid-probability, 5 years plus, high probability.
It might take 20.
Correct. Yeah.
There's so much money, there's so much at stake if you this up.
Absolutely. Everyone talks about the hallucinations a lot. We've got hallucinations basically solved internally because I think most people think of AI like it can hallucinate. A lot of times when you make a statement into ChatGPT, it's unclear what you were saying.
When we view AI, we're building robots that do the same thing over and over and over and over again.
Right.
That's the way that you actually do it, then those robots become very good at solving the errors. I think we're deep in the weeds of it. ViantAI is an agentic platform internally, it's riding the RTB rails today. We fully plan that it will switch to a different rail at some point in the future, it's just a question of time on when that is, we'll be ready for it, no problem.
Okay.
I can give you a good example of where we've experienced this in the past. In 2009, 2010, Tim and I became obsessed with We were watching online video happen-
We wanted to look to the future, we were obsessed with a television will eventually be internet connected-
Content will be streamed, not delivered over copper into your house. At that time, we were going around to all these industry shows for broadcasters, this technology was out there called IPTV. That's what it was called.
We literally tried to hostile takeover-bid a private company called Move Networks-
that was early in the space.
Which has become Sling.
Yes.
Oh.
We weren't successful in that acquisition, but we immediately started building Xumo.
This was in late 2010.
Yep.
We believed that it was a sure thing it was going to happen, and it was going to happen in the next two to three years. We built Xumo. We burned so much capital, and we had installations with all the big television OEMs. We were getting our software installed on chipsets. We were so early. It didn't hit until 2020.
10 years later.
Correct.
10 years later, we burned an insane amount of capital.
Well, you probably got it back when you sold it to Comcast.
We did. You know what? We were first.
It should have never sold.
Yes.
Had we not done that, we would have maintained control. Had we maintained control, Xumo would be an independently, publicly traded company in the billions. It's worth $2 billion today, last valuation. That's what happens when you're a founder. If you miss on the timing, you've got to raise more capital.
Yeah.
You lose control, you have an idiot making the decision. That's it.
Even when you were right, you were wrong.
Yes.
Right.
You can still be right, but you're still wrong.
Yeah.
Well, it's like Wall Street on linear TV. We were right about linear TV.
Correct.
It's just going to take 30 years, not 10.
Correct.
It might take 50.
Right.
People might have to die.
Correct.
If you're too early, it's just as bad as being wrong.
Yep
if you're getting paid.
It's the same. Same outcome.
Same outcome. Let's talk about one of the things I wanted to ask, just on this last thing on leadership. One of the most interesting things that's being said is that maybe we need to get rid of functional organizations. No more marketing, no more product, no more engineering. That what we need are these SWAT teams, where you put teams of five people or 10 people from all different areas, and you say, "This is your deliverable.
In six months, I want you to do this because we're getting requested by clients.
Yeah.
If you don't, you all lose your jobs.
If you do, you get to get the next project." It's a big idea because it would destroy traditional architectures.
Yeah
you could implement that, or a startup would start like that. They would never have functional designations.
Yeah.
Big companies would be doomed because they can't get rid of the silos.
Yeah.
Do you have an opinion on that?
Yeah.
Why don't you start?
It's what I think we try to put in practice almost every day.
Okay.
Not formally per se, where we say there's no more marketing, there's no more engineering. They certainly are there. Every time, we have our top priorities, and that's basically what we do. Let's say that you're going to build a new product.
Yeah.
Okay? I'll use our Outcomes product that we launched at CES-
Yeah
as an example. To successfully launch that, you can't just have a product manager.
Why don't you tell them what it is first, so they know what you're working with?
Yeah. Outcomes is our first fully autonomous advertising product, but it's for performance-based marketing clients.
Right.
It basically is fully automated. They give us four pieces of info, the advertiser name, the budget, what your goal is, you want sales, and what's the flight dates of the campaign? From there, the AI runs.
They give you the creative, presumably.
They give us the creative.
Okay.
From there, the AI runs the entire campaign. It picks the media plan, auto-generates it. It decides how much budget it's going to allocate. It is in charge of all the optimization. It constantly regenerates new sites, new apps. It runs the measurement. It's incredible. However, to get this product, traditionally, a product manager would scope it out, they'd write a PRD for it, they'd give it to the engineer, they'd build it, and then they just expect that these things come out perfect. They don't. You have to get people from the customer. You've got to get customer feedback on what kind of guardrails does this need? It needs brand safety.
They need transparency. Okay, great. You have to bring in the product marketing team as well, because the minute we launch this, we need salespeople ready to go. You have to bring in sales enablement and training to learn how the product works so they can help train the salespeople. That is kind of what you're talking about.
Yeah
these SWAT teams. Tim and I also just started this kind of along the lines, we don't want these one-on-ones, but what we do is a standup call. It's the most expensive call in our company because there's probably about 20, 25 people on this call.
Isn't it just on internet protocol Zoom? Which is free.
It is.
It's still an hour of 30 people's time.
Yes.
I see. It's expensive from.
From a.
personnel-
Yeah.
Okay.
Every week, those topics can vary, but they're our top priorities. Not everybody talks, and they know.
Is it presentation or Q&A?
No, it's just us.
It's we run it.
We run it and say, "These are the-
We drill people with questions. What's the problem? I ask the next person.
I see.
We play judge and jury as well. "Nope, this is the right answer. Move forward.
That's what we're trying to do, is move at a pace that's very fast.
Break down silos. Something's stuck again the second week in a row. What's the limiting factor? What's the issue? We have all the I shouldn't say all, but we have a lot of context across the organization, and that person, who might be an engineer or a product manager or a salesperson, maybe doesn't have all the context. We can provide that. That has proved very valuable. It moves things quickly through the org. It goes back to the original point that Tim said. We don't believe in what we call the managerial class of the CTO.
You're worthless.
Yeah. You just preside over these meetings.
If you went to Harvard Business School, Stanford Business School, you don't have the domain expertise, you are worthless. That was a bill of goods that got sold to everybody, it turned out not to be true. You just have an expensive college bill you've got to pay back.
I think I would probably disagree with you on that.
By the way.
Being one of those Harvard people. Okay.
Let me just also bring up, just to answer the question, the absolute answer is yes. SWAT teams with heavy domain expertise are the future. Just look at Stanford Engineering School. Big Tech used to suck all the talent that came out of all these schools, just looking at Stanford, number one. Okay. They used to hire 50% of kids coming out of that school.
Right.
It's down to 7%.
Yeah.
Why? Because they don't have domain expertise.
You're not valuable in business today because AI has the coding expertise. It can help you get these things done.
Right.
To me, I think what you're going to see by big tech is many people are going to lose their jobs, unless you are absolutely critical with domain expertise. That's it.
You have good human judgment, or you can interact with people, because you're still going to have to manage people.
Sure. Yep. Relationships are still important.
Yeah. Okay. Very interesting. Let's go to the quarter, which was excellent. Just to recap, but I should let you do it because 18% revenue growth, guiding to 23% revenue growth, compared to The Trade Desk at 12% guiding to 8%, and almost every other today, Nexxen came out, and I think theirs was highest under yours, which was 13% revenue growth and guiding to higher. The point is, you guys are double, okay, 50% higher than the rest of industry and accelerating as you get, you said every quarter will accelerate. Talk about what's driving that really formidable growth that the rest of the industry, because we cover 10 of these, nobody's close to you guys.
Yeah.
What's driving that, and why does that continue?
It's just because we have the best product. It's not like in the end yet, but there's certain portions that are execution. I think advertisers are very smart. There's lots of data. They measure everything. Ultimately, it comes down to who has the best product in connected television. The most thing I'm most proud about, those are great stats, and we're really proud of it. We didn't mention that EBITDA was up 81%. We should. Cash flow was up quite a bit as well. From a management perspective, we're doing a great job. We flat out have the best DSP, independent, buy side only, and we think that's the right way. Others will disagree. Ultimately, our growth rate in CTV was up 40% for the third year in a row. It's like, how many years can you put up before investors figure it out?
Our stock is down today, apparently another year. We'll wait till next year. I think when it comes down to it is we just flat out have the best product in CTV. We think that is where the decision is made. We have Iris, which gives you content-level intelligence. It's exclusive data. We just did the acquisition of TVision, which customers are about to get the benefit of that. Now a customer can target at the show level. We get the lowest price because we have the best bidding algorithm that's out there. AI Bidding has 90% adoption. It was the fastest-growing adopted product that we ever released. Now with TVision, we now have the data from a U.S.-based panel feeding us the data before the ad break hits.
About attention.
Around not just attention, but reach. What's the co-viewing, the number of people in the room watching this content? What's the demographics of those that are in there? Are there eyes on the screen, or are they looking at the phone? That's attention. I think when you think about the data that we have before the ad break hits, now we're able to buy the ad way more effectively than any competitor, then single-handedly, the product that's needed. Today, when an advertiser buys, they buy from Google Direct for YouTube, Prime Video direct from Amazon. They buy linear TV from all the television networks, and it's measured by Nielsen. YouTube measures themselves. Prime Video measures themselves. The open internet provides an actual statistic. You have four different measurement systems that are measuring.
For the first time, TVision puts all four of those, we measure all of them as an independent provider. Now you can understand linear. NBC provided value number 1. Peacock might've been low. YouTube, we see over and over, it's not as high as everyone thinks. Prime Video, depending on when you bought. Having unified measurement, which is there is no alternative product to it, that is what is the number 1 need for marketers today. I think as we go forward, our revenue in CTV will continue to accelerate. The overall metrics of the business will continue to accelerate. But it's really due to that it's just the best product. Over time, advertisers figure it out, I think investors will figure it out, too.
Why wouldn't you sell, if that TVision really does such a critical need in the measurement space to put walled gardens versus the open internet, why wouldn't you sell that, productize it to sell it separately also? As a measurement product.
Sure
Maybe other DSPs can integrate it, but.
We do. We do sell it. Advertisers buy the measurement, the content owners buy the measurement. When it comes to media activation, that's exclusive to our DSP.
That makes sense.
Yeah. We are garnering revenue across the board. All of our competitors would like to have it.
They could've bought it as of last month.
They could've, and some of them licensed the data. We chose to acquire it, and there's contracts out that we're going to let expire, and we will pull that data back in to be exclusive to our DSP. Now, why do we want to do that? Nielsen is the currency in linear TV. We're not interested in being a currency from measurement. The old way was measuring potential reach. We'll just say that's what Nielsen does a good job at. When it comes down to being the currency today, this is the RTB transaction. I'm about to buy an ad from Peacock. What's the price of this ad? We think by putting it into a DSP, that's the best way to understand the value of one ad impression versus another.
Right.
You have this thing called CPM adjusted.
Attention-Adjusted CPM.
That's what it is.
Yeah.
Attention-Adjusted. If there's three people sitting in the room, you can pay a little more and win the bid because you're reaching three people, not one person.
Exactly.
If somebody's not paying attention, you're bidding half as much because they're not paying attention anyway.
Exactly.
It's a low impact CPM.
Just to add to that, the only people that have that information are the advertisers that work with Viant. They're the only ones who have the information ahead of time. We represent the advertiser, we want them to have the strategic advantage.
When you sell it to third parties, it can't be ahead of the activation.
If I was selling it to Peacock, they're going to raise the price. Do you see what I mean?
No. I don't see what you mean.
If Peacock was aware there are three females
I see
are 18 to 54, they would raise the price of the CPM for that one ad impression.
I see. How does the product that you're selling to third parties differ from what you're using for activation?
It's delayed.
It's delayed.
It's just not in real time.
It'll tell you.
Okay
Let's say for the month.
Oh
Peacock had.
Average co-viewing
60% attention. It had 1.4 people on average watching.
Per screen, all that involved.
Right. Really, marketers.
Okay.
That's good to know when you're.
Yeah
media planning.
Yes.
You might commit a certain amount of budget, let's say, to NBCUniversal Peacock. But really what you want to do is, Peacock has 5,000 different titles.
in CTV underneath their umbrella.
Yeah.
Every show behaves differently.
That's true.
Right?
Second by second, it's different.
Right. The marketer wants to have the advantage of the information to know this impression right here, before this ad pod, before the javelin throw in the Olympics.
Right
What is the co-viewership? How many people are in the room? How many of them are actually paying attention second by second? If it has really high attention signals, you might bid up.
You might bid a little more. If it has low, you might bid less or not at all.
The point of advertising is to grab the attention of the consumer so that they're watching, so they watch for as long of your 30 seconds as you have them.
You have to gain the attention of the actual consumer.
Persuade them.
Yes.
Do you guys feel like this pushes you into content? Because content is a really big driver to whether people sit in the room and watch the full ad.
We license the data to the content owners and it's Netflix, it's Prime Video.
I was thinking the ad.
the ad.
The ad content, not the programming.
Absolutely
content.
That's what's on the product pipeline. we've showed off ViantAI, which does autonomous media planning. It implements it executes the campaign, it optimizes after reading the reports. Making the ad is the next critical component. I'd listen to John Wren of Omnicom, and he mentioned it's all about media, and I think where the profits currently are is media, but I think you're going to see creative and media come back together. the best example I can give you is if I know I'm going to show a 30-second spot on "The Kardashians" on Friday night, why wouldn't I make a custom 30-second ad for that? now with GenAI, you have the ability to make that ad for $1,000, whereas before, TV ad production cost $1 million for 10 spots. they're all basically the same, slightly different.
if I know where the ad is going to be placed ahead of time, I can actually make a custom creative. when you put the creative matching the content, the performance, the attention, the persuadability, and ultimately the buy rate, the conversion rate into sales skyrockets, and that's where advertising is going to have a renaissance and where GenAI plays a big role.
What I was thinking on product roadmap is since you can tell ahead of time, you can A/B test five ads that you create. Let's say they deliver you one piece of creative that they've approved. You can do five things.
You can modify.
Turning the watch over, you can turn it into blues, then do A/B testing over the first five ad blocks then say, "Okay, the one that works best is the red one that doesn't move.
Yeah.
Serve that one the rest of the time. It feels like that's a natural place for you guys to go, given that you can do these very fine distinctions about what's holding customer attention versus what's getting people to look at their phone.
Yeah. What GenAI gives you the ability right now, a marketer like Molson Coors, one of our customers, they only have 10 spots for Coors Light. Why? Because the ad production shoot is $1 million plus dollars. They can't go do another shoot. The longevity, it takes forever.
Yeah
The timeframe. Now you can start with a baseline ad. Let's say it has 60% attention, and the goal is improvement.
Yes.
60%-65%, 65%-70%. That way, the working media dollars are actually working harder.
Yeah.
You're leveraging creative as another lever to do that.
If you look at a trend over in the social space, if you look at Meta, they do all of their Advantage+ system when those e-commerce marketers come in. They control all the targeting and the media optimization. Meta controls all of that.
Okay.
What is the advertiser responsible for? Creative. The trend in that for the last two years is more and more and more creative. They have to pump out so much creative. What Meta needs is to be able to make, whatever you're watching on Instagram Reels, if you're looking at mine
I can't chip worth a damn right now in golf, I'm just watching short game chip stuff, next thing you know, what am I getting? I'm getting an ad for a golf aid for chipping. They have ads matched to micro-categories of content. That shows you. Some of these advertisers are pumping out hundreds of pieces of creative every month. It's cheaper than television creative.
That is where the future is going. A Molson Coors needs to be able to create hundreds or thousands of variations of creative to match the content. Is it NFL football? Cool. Make a football-themed ad. Is it Kardashians? Make it women-themed. Whatever it is, you want to theme the creative so that it matches the content. To Tim's point, it grabs attention. Every one percentage point that the marketer gets in attention, every increase-
it's a 1% increase in awareness for their product, a 1% increase in ad recall, and what we need to do now is tie Outcomes, the sales.
What does that translate into sales?
You would have a different business model because wouldn't you take a percentage of the lift in the outcome and the percentage sales lift?
A commission on selling the product. We'll see where the business model goes to, but ultimately, if you have an advertising engine that moves product off the shelves, that's very valuable. Whatever the right business model is, it'll reveal itself to us.
Okay, you think you're moving. You think that this new TVision with the attention actually helps you drive Outcomes which then maybe implies a different business model for you?
Over the long term, it implies a different business model. Today, as a percentage of the spend is the way that it operates today. That's a frictionless way to drive revenue. Obviously, we're public, and we're worried about quarterly results-
Yeah
throwing a brand-new business model-
Yeah
as a public company is harder.
Risky business, yeah.
Certainly, if you're making the ads, you're automating the creation of the ad campaign, the optimization of the campaign.
In a self-reinforcing loop. That is very valuable. I think with the acquisition of TVision.
Our focus on strategic proprietary data that's valuable, that we've described, we're transforming from what was a once media execution platform, like The Trade Desk or DV360, into an advertising intelligence company. That's where I see our transition moving, one that offers proprietary data that doesn't exist anywhere else, independent measurement, and then media execution, all in a self-reinforcing loop. To me, that's an advertising intelligence company, and that's our future.
One of the things that a lot of DSPs have done is layer in data, like 300 data providers when you're making audience segments.
Yeah.
Are you now in channel conflict with those because you have these two proprietary data assets, so you lose all the revenue you were making from those.
I-
300 third party?
No. Take a look at IRIS.TV. It's a one-of-one solution. They're the only ones that went out and did the dirty work.
Show level
heavy lifting to go out to the content owners, do all the integration with 1,900 different content management systems, gain the trust of the content owners to be able to send all their video files of every show and episode that they have, let them index it, run all the computer vision, what's this video about, and then pass it through the bid stream. There is no competitor to it, and we acquired it.
No channel conflict?
No channel conflict. There's no other content identifier in the bid stream. It's only IRIS.TV.
There's no other TVision.
Yeah
U.S.-based panel. There's Nielsen, but it's a people meter, 1989 technology. We have a camera that's passive that does facial recognition, and no one does anything, and it's just automated. These are one-for-one assets that we look for. That's what we're looking for in M&A. To understand that Tim Vanderhook's email is blank blank blank @gmail.com. I don't want to do it on video. I'm going to get spam.
We get enough spam.
Just to know someone's email address, that's commodity data. There's lots of third-party commodity data. Most of it is stale. People know what car I drive, that I drive a Tesla. You can get that data from 100 different providers. In the end, does that data have value? It has some value versus not knowing it. When there's 10 providers with the data, ultimately, it gets commoditized, and that's most of the data that's out there today. There is very valuable data assets out there.
Very few.
As part of our strategy, when we go and pitch a marketer, and they're like, "Hey, I'm using another DSP already. The only reason why I would switch is if you have unique data that's exclusive to you, or you represent content that I can't get anywhere else. Which one of those two are you? Tell me.
For years, it was okay to compete in the space because you could win two out of 10 marketers.
If you want someone to switch, you have to have something exclusive to you that they want.
That's very valuable.
One of the things that's come up on this stage a lot over the last couple of days is there's sort of a new entrance on DSPs using agentic buying platforms.
Yeah.
There's this One of the reasons the stocks are under pressure, I think, on the DSP side is we're not seeing this on the SSP side because you got to get a guy to adopt your agents. On the DSP side, it's cheaper to enter the market now and then just be a buyer. Sort of talk to that competitive threat about how agentic systems make the entry into your competitive set of DSPs more competitive.
Yeah. If you go to buy, let's say, 100 ads-
99.9% of the time, you would buy the wrong ads-
Okay
if you used agentic. I think, again, it comes back to which users within Peacock or within ESPN-
are you going to buy out of the 100 that are presented to you, and answering the question why. The agentic protocol exists today. There's very little adoption.
Right.
The technical capability is there.
Yeah.
In our business, if that were to come and come to fruition, we would lose no customers because our customers choose us because we have IRIS.TV, because we now have TVision, and because we have Household ID. We can resolve back to we know this is Chris Vanderhook's household. We know he's watching this show, and we know he's likely in market for Coors Light. He's in market for Coors Light all the time.
All the time, because now you have to drink Molson.
Post earnings.
Yes.
Post earnings, you're always in market for Coors Light. That's what I mean is, again, it all comes back to your value proposition is about exclusive data. It's not that we can bid and buy an impression. Because agentic or RTB, everyone can already bid and buy an impression. Let's say it gets cheaper, great. Our operating costs drop and our EBITDA skyrockets.
Okay.
That's the real reason why we want agentic to happen, and I hope the adoption happens faster. No one's choosing us because we have the RTB protocol. There's 1,000 DSPs that are involved in the RTB protocol today. There could might be 10,000 in the agentic protocol of tomorrow.
Right.
What I mean by that is there's tons of competition for bidding and buying for ads. The advantage is which users should I buy, what price should I pay, and what's the estimated ROI to the advertiser?
Let's go back. One of the things you said up front is we really believe that single-sided DSPs are the way to play.
Yeah.
Well, you just said, Wall Street agrees, that unique supply, either content or ad units that are unique-
Sure
are actually preferenced, like they have pricing power. Just like unique data, which you guys now have bought.
Mm-hmm. Yep.
Why wouldn't an acquisition of a proprietary content base becoming two-sided be actually on your roadmap of M&A?
Well, let's see, like by YouTube? If you think about it, if I owned content, let's say that we turned MySpace ad-supported, which it's not ad-supported at all today because we were tired of nasty articles being written about us. We said, "That's it. Take all the ads off. It's not worth it.
Okay.
It has to be large enough
Yeah
for it to draw a large marketer.
Okay.
If you bought some news publication
We call it, is your content of high enough quality or reach that you're a must-buy.
Right.
There's very few must-buys. Let's say there's 10 or less that are out there.
Yeah. That's not in our thinking. It's also not in our thinking because it would just change the thesis of the company.
If I become a content owner, my business model completely switches. If you're the marketer, it's not about doing what's best for you. I need to fill that inventory.
You're maximizing margin.
That's my number one priority. Xumo, our previous business.
We were on the sell side. That's what that business was.
You get 100 ads that you can sell. You must fill all 100, otherwise your margin is less. You fill it whether it's good for the advertiser or not.
We-
This is why, as a buyer of ads, you can't trust a seller to actually deliver the ads for you.
We give.
They're going to give you the worst ads.
Yeah. We gave a talk at Xumo once, and we had one slide, and it said, "CPM times fill rate equals your business.
Your business.
We were telling the team, like, "Get religion on this. We have to fill more ads." If you're a content owner, that's your business model. It's not bad. We're not saying it's evil. It's just that you can't actually-
It's a conflict
service the marketer because my incentives are conflicted. We don't want to do that. I think that the proprietary data piece is all angled towards when we go and talk to a marketer, I want to figure out a way to grow your business. I would say we're getting RFP'd a lot right now. The ones where we are most-
RFP. Define.
Oh, RFP, request for proposal. marketers-
New customers, sales pipeline, baby
new customers. Yeah. It's-
Like backlog.
The lifeblood of all business, new customers. The ones that we feel that we have the best chance, where we feel we're in the driver's seat, and it's many of them. We look at the corporate earnings of that company, and if they're lagging the top 10 companies in the S&P 500, we know how they spend their money in marketing. They're predominantly giving it to the sell side, Google, Meta, Amazon.
They're spending 60+% of their budget with these guys.
Yeah.
We know that we can come in with the information and insights and show them how they can actually spend more efficiently, and it'd be like getting. When you work with us, it's like getting another 20% or 30% bump in ad spend from your CMO or your CFO, because not only are we more efficient, but we're going to put those dollars to work much harder. We want to grow their business.
Okay.
Whether that was Molson Coors or Whoop or any of the other ones that we won that we've announced, that is our pitch when we walk in the door.
I would just put it differently as well. The total addressable market for the buy side is the entire advertising industry.
Yes.
The total addressable market if you're on the sell side, let's say I'm an SSP or a content owner, is a percentage of those budgets.
That's true.
Okay. That's the way I look at it. I can help Molson Coors spend their entire ad budget, but if I was Prime Video, let's just say, and a great streaming app, you're still going to get a percentage of the share of wallet. YouTube will get a percentage of the share of wallet. I view the TAM when you're on the buy side as the total ad industry versus if you're on the sell side, it's a percentage of the ad industry, depending on your size and scale.
One other point. If I'm Amazon and I have the Amazon DSP, what do I think that their future is? If you want to buy Prime Video
Yeah
You can only buy it through the Amazon DSP, great. That's what's going to get it.
Yeah.
YouTube, the only way to buy it is through DV360.
Yeah.
Great. They're a walled garden. That's fine. They will get their sliver. For us, we want to be able to sit on top of much of the spend as possible to help the marketer actually grow their business, not just in a small sector.
All right. I get it. Okay, questions? We have about five minutes left. Any questions for this? Chris?
Yeah. How big is the
Yeah. The panel in the U.S. is 5,000. It's in the U.S. only. 5,000 households.
Yeah.
It's about 14,000 people within those households. To compare that to Nielsen's people meter is between 10,000 and 12,000 households. We are a nationally representative sample. Nielsen is both national and local, so they're in 210 DMAs, but they have announced they're retreating from 130 DMAs they're not going to support. You're going to see their panel size shrink. It's too expensive. What's great about the TVision panel, it's a much lower cost to operate the panel. Nielsen sends out a whole team. It takes them three days to interview you and set you up. They have lab coats. It's ridiculous. They literally wear lab coats.
They have high churn.
Yes, very high churn. This is a passive panel. We send a box into the household. They plug it into a power source. It does the ACR of what's on the screen, and then there's a camera that you mount on top of your TV, plug it into a power source. You register on the app, everyone in the household, they take a picture of everyone.
It takes three minutes to set up. I have it. I'm a panelist now. You have a box the size of a Mac mini. The camera plugs into the box. You give the QR code because you pre-registered for the panel, you're up and running. You don't have to push buttons. You don't do anything. Our churn rate is very low of people leaving the panel. That's as it stands today. The excitement that we have of the R&D we're doing, what we can provide advertisers with that panel and market is going to be exceptional.
If it's so cheap, why wouldn't you make it 50,000? Why wouldn't you ship these things out to anyone who agrees-
Okay. Yeah, sorry
Make it a lot bigger?
Really good point. We have 5,000 households today, and we are scaling that panel across every single DMA. You really just have to get to statistical significance, and then there's no-
In each market.
Yeah.
We're statistically significant for national buying. That's the opportunity.
In 35 DMAs.
Yes.
Out of the 220.
Yeah.
My follow-up question was going to be, is that a large enough panel size to inform pricing strategy? That Attention-Adjusted CPM, right?
Yeah.
Are you extrapolating?
From a national basis, absolutely.
Okay. Let's talk about how we do it. TVision, had we not acquired IRIS.TV, we would never be able to bid and buy in real time. TVision is measuring streaming apps, linear television networks and shows. IRIS.TV is supplying that content ID, which is the show, and that's how we can model onto real time who's watching, how many people are watching, what's the attention, and bid and buy in the RTB stream. If we had not acquired IRIS.TV, we wouldn't be able to do this. If we had not acquired TVision, we would not be able to do this. It was a very focused strategy of IRIS.TV, scale the IRIS.TV system, and now buy TVision and scale the TVision system.
Although we're not interested in being in the currency business, we do in effect believe that TVision is going to grow the IRIS.TV carriage.
Penetration.
Penetration and market because we're going to be.