Loading summary
Transcript98 lines
- [00:00]
A
Look, big picture here, we've said this since day number one is that this concept of programmatic media execution as it relates to TV or CTV doesn't make a lot of sense. Programmatic CTV just leverages certain digital principles and technology which are ill fit for the market. I might be off by a billion dollars or something like that, but let's say the broader CTV market is $30 billion per year in the United States. About half of that today is these direct transactions. The other half is through programmatic biddable. This is the piece, by the way, that the press talks about every day. It's about $15 billion. And the hard truth of that $15 billion, probably half of it is fraud or low quality inventory. It's really only 7 or 8 billion dollars of premium inventory. TV media isn't necessarily the right media for the true SMBs that TV is no longer a big budget brand privilege as it was as little as five, 10 years ago. I think it's the realization that the biggest part of TV inventory and the most exciting part of TV inventory doesn't live in programmatic.
- [01:07]
B
This week on NEXT Media, I talk with Tatari CEO Philip Engelbrecht about the state of CTV advertising. Philip thinks the CTV ad market has been kind of mischaracterized as this highly dynamic, wide open market where millions of small advertisers are about to come on board. Well, in reality it's mostly about direct deals with a handful of small companies. But there's still a need to bring more brands into that fold and that's what Zatari is trying to lean in with deals with Fox, Disney and others. So we get into that as well as where data driven CTV is really headed over the next couple years. Let's get started. Hi everybody. Welcome to NEXT to Media. I'm Mike Shields. My guest this week is Tatari co founder and CEO Philip Engelbrook. How you doing Philip? Good to be here. Thanks. Thanks for, thanks for being here.
- [01:45]
A
Yeah, my pleasure.
- [01:48]
B
So many things to talk to you about. We were just kind of catching up. We had met a few months back and you were hinting at a lot of big changes but I want to kind of maybe can we talk big picture. We had one of your colleagues on a few months ago but I don't know, give people like the, the one on one journey of Tatari and I want to get your thoughts on like the state of CTV ad tech before we get into some of your news and some what's going on. But give us the origin story quickly.
- [02:14]
A
Yeah, yeah. And I think you're referring to the Andy or CRO meeting with Chime and being on the podcast. Yeah. Which is a good one. But so yeah, let me start with the big picture on Tatari. Tatari is actually structured as a holding company, infra short for infrastructure for TV advertising. The brand that everybody knows is Tatari itself, which is our demand side platform. Right. It's a place where both brands and agencies can come to manage their creatives, plan campaigns, execute media and then of course measure and optimize. And you know, within brands, we've got very small brands, we've got brands. We had four or five brands in the super bowl just, just a few weeks ago. So it's a pretty broad spectrum. Right.
- [02:59]
B
So not, I should be clear, not just a CTV platform.
- [03:02]
A
Correct, that's right. We, I mean one of the things in which we differentiate is that this concept of conver conversion tv, Right. Where we are truly fluent between linear and streaming TV and not don't have to make that distinction kind of giving the full spectrum. So that's the target as a demand side platform. Then we also have supply side tech. Yeah. This is the part that is renamed to upstream. And I'm not going to say more because I think this is where we're headed to in the conversation. And then there's a third piece real quickly, it's called Vault. It's a privacy and identity solution really focused on ctv. Think of it as a, as a special purpose data clean, clean room solution for ctv. I don't need to explain why we have it. The three of them combined makes for infra. And so that's the broader context.
- [03:47]
B
All right. That's a lot of stuff. And we are going to get to upstream and as you're hinting at. But let's just give me your take on. Because you were early, early in this space in some ways and then right on time maybe, because I think we even talked about this, there is a, you know, CTV has exploded. There was an argument that maybe we, the, the programmatic pipes were catching up or did, did the market even need to be programmatic right away? And what, what did you see early on? Because you, and, and because there's, there's a, there's definitely a desire to have more efficiency and more data driven television advertising across the board. But it's, it was a market in transition. Right. For a lot of this period.
- [04:28]
A
Yeah. Look, big picture here, I mean we've said this since day number one at Tatari is that this concept of programmatic Media execution as it relates to TV or CTV doesn't make a lot of sense. And actually a lot of people kind of like maybe would have criticized us on that in the past, but I feel like the tides are changing right where they actually see why, why, why
- [04:54]
B
is this any different? What are you talking about?
- [04:55]
A
Right. Yeah. Right. And so, so ultimately comes down to how the market is structured. Unlike other media, say display, the concentration of supply and CTV is, is phenomenal. As a matter of fact, looking at Atari, 90% of all of our streaming impressions placed for brands or agencies come from the same top 10 publishers.
- [05:16]
B
There is a long talent TV but it's very different in terms of what we think there is.
- [05:20]
A
Yeah, it's, it's a big head, even a very small torso. And so if, if that is the case, and this is a fact, these are numbers that we have, why do we organize the market through this whole SSP system which is really good for digital, right. Where you have millions of inventory, you got to organize.
- [05:38]
B
You need all this aggregation and you need this.
- [05:41]
A
Don't get me wrong, there's a place and time for Programmatic but just not for ctv. And so we have always held that belief and we also kind of put our money where our mouth is. And so years ago we bought supply side technology company named the Viewpoint. We quickly learned that competing in the SSP market is not easy. It's hard to innovate, it's highly commoditized, not that much money. But instead we have now increasingly used this technology to build those direct relationships with the publishers.
- [06:19]
B
Interesting. So, so because I was going to ask you, is there enough, you know, there, there's a lot of, there's a, there's a lot of people trying to you know, reproduce the, these display model or business model in television.
- [06:31]
A
Y.
- [06:31]
B
There's only, there's only so many companies that are big sellers and there's only so much inventory that they put out there. Right. So that's just like a different dynamic. Did you find that was just harder to have the reason to. For being of this full stack product or what was the. What did you learn?
- [06:46]
A
Yeah, multiple layers here. So.
- [06:49]
B
Yeah.
- [06:49]
A
So let's first go back to kind of like what technology is in place in Programmatic. Right. And this is where I always say it's kind of like all of the Programmatic CDV just leverages certain digital principles and technology which are ill fit for the market. And so what are. I mean Programmatic is not bad, it's automation, but it does come with a Bunch of well known issues. Fraud, brand safety fees, DSPS fees, SSPS fees, data fees. I mean, you can just name it up, right. Sometimes half of a media budget goes to fees. And so if you can remove all of this with different tech, then we're already better off. We bring it back to the automation. And so ultimately this is where upstream was born. I think the second layer of your question is like, where is it relevant? And so let me maybe kind of paint the CTV market in broad strokes for you.
- [07:49]
B
Sure,
- [07:52]
A
I might be off by a billion dollars or something like that, but let's say the broader CTV market is $30 billion per year in the United States. About half of that today is kind of these direct transactions. The other half is through programmatic biddable. This is the piece, by the way, that the press talks about every day. It's about $15 billion. And the hard truth of that, of that $15 billion, probably half of it is fraud or low quality inventory. It's really only seven or eight billion dollars of premium inventory. This is kind of again, what we talk about in the press every day. This is where we see a lot of other companies truly hanging their hat on. Right. But they're ignoring the other $15 billion. And that's where we started.
- [08:39]
B
You're talking about stuff that gets sold in the upfront or through direct, a lot of direct deals or, and also private marketplaces or like how do you categorize, where do you put all these pieces?
- [08:49]
A
A lot of that inventory is simply not accessible through programmatic bidding. Right, right. I'll give you an extreme example. I can give you many more. The Super Bowl. Nobody can buy the super bowl programmatically.
- [08:59]
B
Right, right.
- [09:00]
A
You, and you know, today you start the conversations with Disney about 2027 and you buy that direct. But it's just not the Super Bowl, Mike. I mean it's, you know, especially in sports, anything that's, that looks and smells like a playoff or a bigger game,
- [09:16]
B
there's no reason for those companies to.
- [09:18]
A
You're not going to find that on the trade desk as an example.
- [09:22]
B
Right.
- [09:22]
A
That's just impossible. And so this is inventory that the publishers very much prefer and will only sell directly. Right. And, and, and so the unfortunate piece about selling direct, of course is often that comes with manual work. Right. When you say manual, it's small.
- [09:41]
B
It feels like you're giving up the things you like about programmatic when you do that.
- [09:44]
A
Exactly. And so we use our supply side tech and ask ourselves a question. Well, can we hold on to all the Benefits of being direct. Right. And we can talk about in a second what that actually is. But can we also give the automation the beauty of programmatic and that is upstream? And so we painstakingly worked for almost two years on these one by one, custom tech integrations with the five leading publishers, Disney, Warner Brothers, Tubi, nbcu. And I'm forgetting somebody, Paramount, of course.
- [10:16]
B
Yes, yes. Well, they're only going to be three or four by the end of this podcast with all the deal making going on.
- [10:24]
A
Exactly. But, and so now we have these integrations in play that when we have to do a direct buy, whether it's for $100 or $3 million in one fell swoop, that is fully automated. Right. And when we get to such level of automation, it's something that really plays well for the publishers. Right. Because less manual work, they're much faster, they can transact any amount. Right. It doesn't matter. It's a few hundred dollars or a few million dollars less errors. Right, Right. And so obviously it took us a little while to demonstrate, convince them that we can do this, but this is in place.
- [11:08]
B
So how do you make sure they feel good about it? Control and just transparency has always been a big thing.
- [11:13]
A
Right. Ultimately they see, they see one change in their business with Atari, right. Is it's, it's more business.
- [11:19]
B
Okay. That's, that's what might be people happy.
- [11:21]
A
You have hundreds of say, brands in tow that want to buy premium inventory, but sometimes, you know, they don't. Maybe the transaction amount is too small or maybe they don't, you know, the speed to market can't be there. And so now we kind of like change all of it. And so with one of our early publisher partners, we pretty much doubled volume in the six months test period. Wow. Right. Because we could do things that we otherwise couldn't. And so it takes a little bit of time and you know, like change is scary for everybody. But as we've built this out and the publishers see how it works and they realize that they don't lose control, they drive more business, all is good, off we go.
- [12:05]
B
So you mentioned, you know, like it hasn't been there. Isn't TV is not wide open. Like maybe some would like it to be where it's all in one exchange. Everybody just buys whatever they want. And the, you know, this hope has been that all these new advertisers would come on. You guys have seen newer brands come to the market, but how, how will this accelerate that? And do you think this is going to be the like really small pizza place type advertiser or smaller brands. Mid, mid. What do you think it's gonna look like?
- [12:33]
A
Yeah, before I answer that question, I mean, look, there's a fundamental belief that we have at Atari is that, you know, TV media isn't necessarily the right media for the true SMBs. The 20, 30 million or so SMBs that we have in the United States, the corner store, pizza shops, whatever it is. We do think that for a brand to find success with TV they have to be of a certain caliber. Think of something like a million dollars per year in general marketing, maybe $10 million in gross merchandise value. Okay, okay, so it's, it's not for that long, long, long, long tail of businesses. But I can also tell you that TV is no longer a big budget brand privilege as it was as little as five, ten years ago, maybe 2,000 companies only now. And so we definitely moved the needle to 10, 20, 50,000 companies that are now eligible for TV. For them, for example, to buy into a premium placement, things that were sold directly. They even couldn't talk to the networks today back then. Sorry.
- [13:32]
B
Right, right. Yeah. Where were they? Where would they even get started? Like that's where you even call.
- [13:36]
A
Right, like, and by the way, to the, to the publisher's defense, I mean like all of this is manual. They weren't ready to handle all that. The demand, so highly fragmented. This is exactly what's changing.
- [13:49]
B
Digital advertising is facing growing signal loss across browsers, devices and connected tv. That makes accurate identity more important than ever. Intent IQ is a privacy first identity resolution leader, helping advertisers, publishers and platforms recognize and reach real audiences across both cookie based and idealist environments. Powered by a patented identity graph and advanced signal enrichment, Intent IQ delivers the scale and accuracy needed to drive measurable performance and better monetization for both advertisers and publishers. To learn how Intent IQ helps turn fragmented signals into real results, visit intent iq.com okay, so this goes well. You're. These guys are going to see the Paramounts, the NBC's are going to see new demand. It's going to, they're going to have the control that they like. The pricing is not going to go crazy. That's all should be good. So what does this, where do you think upstream takes the market? Like is it, you know, are we going to see a shaking out of DSPs? Are we going to see more inventory available? Where do you envision things going?
- [14:52]
A
Yeah, I mean we're starting with a big chunk of the market. I would argue it's the most valid piece of the market that direct sales. Right. And it's something that publishers will hang on to forever. Where do we go from here? Look, many possible ways, but I think one thing we're definitely looking at and considering is like okay, well when a publisher does these types of direct sales with automation, obviously right now it all plays to the benefit of the Tatari demand and brands.
- [15:29]
B
Sure.
- [15:30]
A
In the next iteration, could we possibly work with publishers and allow them to benefit from that for other demand which is not Tatari. Right. We're building tech to be used. Yeah, yeah. So that's one possible next evolution. And this is again super early. Haven't done anything in, in that direction. Next evolution is a little bit more complex in nature. But can we help publishers potentially drive higher yield on their inventory? Right. Because if you look for example at Tatari Mike. And then again it is the demand side. We have so much first party data and what we call performance media data. Right. We know across billions of dollars and so many years what type of results a certain piece of media can do. Can we deploy that data so that the publishers and their direct sales can sell that inventory? Possibly. I wouldn't say with higher intelligence, but maybe stronger smart so that they can discriminate on pricing and maybe drive a higher yield. That's something we're definitely looking into as well.
- [16:34]
B
Now is that going to be you kind of led me to my next question which was going to be you're talking a lot, we're talking a lot about new demand and newer, broader base of advertisers. Presumably there's going to be more opportunities to be more data driven. You're built on that. But I think people are impatient that maybe that's not the right word. But they want TV to become super data driven overnight. It's not easy. Do you think this accelerates that and that'll help some of your partners unlock more value because they'll be able to target better?
- [17:04]
A
I mean, look, it's just, it's another step in the right direction. No doubt. Right. Once we have automation, we can, we can apply data, we can, we can apply machine learning to that. And, and, and, and you know, whatever it is that matters to the publisher in most cases, drive more volume. So I mean, vague answer, high, big level answer. Sure. Yes.
- [17:27]
B
What, what do you, what do you think's holding that back?
- [17:29]
A
Maybe just add one, one small piece. Like I'm not saying that program that, that upstream is in odds with programmatic.
- [17:39]
B
Right.
- [17:40]
A
Right Again, there's a place in time for Programmatic. We do Programmatic CTV at Atari as well. It's just very small potatoes. It's like probably less than $70 million a year. Yeah, right. Because there are instances where we have to do or we are seeking a level of, let me call it hyper targeting.
- [17:57]
B
Yeah.
- [17:58]
A
And programmatic, like. Like retargeting. Right. So the highest form of targeting, so to speak. And so like programmatic is really valuable then. Right. It's not that one will cancel out the other. It's just we're highlighting, we're demonstrating that maybe there's too much talk about Programmatic.
- [18:15]
B
That's not where the majority of the demand is. Right. I mean, that's simple as that.
- [18:19]
A
Yeah. And just pay attention. Right. So I think there was just something published where I think it was Denso and WPP retracting from Open Path.
- [18:33]
B
Yeah, yeah.
- [18:34]
A
Right. And the reason side. And this again, this, I'm not, I wasn't privy to any of these conversations. This is what I read in the press. The reason sided were all the issues that come with Programmatic. Right. And just, just highlights that Open Pad or some of these other supply pad optimizations. Ultimately they're still programmatic in nature. They're just header bidding solutions where sometimes it's just kind of like a curation package. Right.
- [18:59]
B
They're in the name of simplicity. But you're still getting a lot of
- [19:02]
A
the same issues, but it's still all the same. And this is how fundamentally different Upstream is. Right. Because the connections are done directly inside the ad server.
- [19:12]
B
Right. Now the obvious question is the big. There's a couple of big eyes that are not part of this yet. These walled gardens that like to do their own thing. Do you think they are ever going to be interested in something like this? Are they always going to be sort of. We're not going to play ball like this. Who knows?
- [19:29]
A
I'm an eternal optimist by nature, Mike. So the answer is yes. I mean, the question would be why not anything that drives that level of automation. Why would they not?
- [19:40]
B
Right. Yeah. I mean, if you could talk about new demand, you would think that's something that they're all interested in. It just. It depends on the terms, I suppose. What? I'm going to just jump back from a brand perspective, like for your existing advertisers, what does anything change for them?
- [19:58]
A
Yes and no. No, because they've only always had the benefits of buying direct publisher direct when working with the tire. Right. Where they get the benefits of sometimes lower pricing. But above all, right Brand safety, no fraud, no intermediary fees and things like that. What is changing is that right, these smaller brands, I mean like we sign more than one brand every day new to tv. Yeah, right. These smaller brands now, unlike, you know, as a few weeks or months ago, they now get to see, they get access, they have the ability to buy some truly high quality premium inventory which till recently was only available via direct sales channel.
- [20:46]
B
Right.
- [20:46]
A
With direct big money at the table.
- [20:48]
B
Right. They're not sitting, they're on. They're not only getting the leftovers like, or the, the feeling that, that they can only get, you know, a certain level.
- [20:54]
A
That's right, yep.
- [20:56]
B
So yeah.
- [20:56]
A
So it matters for them as well.
- [20:58]
B
Yeah, sure.
- [20:58]
A
But up Upstream was always built for publishers ultimately as we can show and bring to life technology that leans into that inventory. It does behoove the demand side as well, right. The brands.
- [21:13]
B
I gotta ask you this, you're, it's interesting. You're, you're making an announcement in 2026 that is not all about AI, you know, or that's not thrown in a thousand different ways here. Can there's, I guess in television there's the AI. Can AI really change creative? Is it going to accelerate the, the, you know, everybody making a new spot? So I'll ask you that, I guess. But also can we have, do you think, do you, do you envision some kind of AI optimization layer coming to TV advertising, Allah, Meta and Google, or is that really going to be difficult given the fragmentation?
- [21:47]
A
Absolutely. And this comes back to, I think the question you had like what's next? Right.
- [21:52]
B
And tell me the future.
- [21:53]
A
Yes, the next is, well, hey, hey publishers, would you like to use upstream for your other demand as well so that you can truly automate every single piece in your business? By the way, we've never asked them, right? Why not? The second piece is exactly what you say, right. Mike is all right, well publishers now you sell all of that inventory, but can we help you through data that we have, all the first party data and measurement data we have over the years and machine learning or AI, can we help you drive a higher yield out of your direct sales? Right. This will take time, but there is a definite window of opportunity there. And it's interesting, right, because if you look at the traditional SSSP business for CTV and I alluded to it, it's very commoditized but it's like I feel like they don't. How can they drive better results volume or pricing for the publishers today with the tech stack they have the short answer is they don't short of bid application.
- [22:54]
B
Right. Which is not. Not the idea.
- [22:56]
A
Yeah.
- [22:57]
B
What about are you bullish? You know, you, you, you've kind of been vocal about. You don't, you don't see necessarily the millions of SMBs coming to TV, at least not anytime soon. Is AI going to revolutionize TV creative? Is it already happening? Is that going to be years away because it's so difficult to change processes?
- [23:17]
A
It will. I will say this though. I mean we have many brands that have built their own AI creators Fallicious as an example. And when we go to them and we, and we ask the question how did you do it? Tell us about the experience. Turns out that they will often use multiple tools from SORA to eotree. We'll spend five days on editing on it and all that kind of stuff. And don't take me wrong, the creators are great and it's fantastic. But five days goes again of the spirit of what we like.
- [23:49]
B
It's not type of prompt and we got, you know, a thousand ads.
- [23:52]
A
We'll get there, but I don't think we're as far as people like to say. But the other things, however that we can do today. Example, when you go into premium placements, right. Unlocked by Upstream, there is a great scrutiny on the quality of the creative. Right. The claims makes and then kind of the approval process which can sometimes take days, if not weeks with the publishers. This is a piece, we have this going and beta already at Atari where we look at all the creative reviews over all these years. Tens of thousands of creators. And so when a client comes to us now and says like will of be approved by the network? Well, yeah, in, in a matter of a minute. We run it through our machine learning to AI systems for creative and it gives us an indication of you're good or go tweak this so that you can pass. I don't know NBCU's requirements.
- [24:47]
B
Right. Okay.
- [24:48]
A
So that, that we can do.
- [24:50]
B
That's. That's a timesaver. That's a, that's an accelerant.
- [24:52]
A
That's a great use of AI. Yeah.
- [24:54]
B
All right. That makes a lot of sense. A couple more for me. What big picture, I keep coming back to this like where what is holding thing back either brands from using more first party data, intelligent advertising or having a better way of, you know, doing the kind of retargeting they're looking for that that's also. Should television even be approached that way? What is kind of holding things back there on the Performance front of in your mind.
- [25:19]
A
Yeah. Well, and ironically it's kind of like even just capturing the data, that first party data Mike, is very difficult. I mean we've done this for so many years, right. Through combination of Pixel instrumentation, server to server integrations and it's, it's. Yeah. And just to give you some flavor of like quantified as well, I don't know, we'll probably ingest something like 150 million events daily. Right. Sites visit, conversion across sites and mobile apps. Sometimes just that infrastructure is very difficult. When we launched Vault. Right. Which is kind of the second thing we have to solve for is kind of the privacy concerns. Right. But when we launched Vault, we then also learned that for example, we would capture way more events than you can do via standard pixelman instrumentation. And so to answer your question, it's possible. It's not as easy as people make believe it and we had to build special purpose tech for it. Right.
- [26:26]
B
I think there's a perception, plug it into a cloud thing and you're done.
- [26:29]
A
It's not, that's not how it works. But once it is there, I will say this. Once it is there and it's humming and the kinks are ironed out, it does like, I mean look, we can, when a client runs a campaign with us, we can quickly do the measurements, see what works, make an assessment of which audience really worked well and then use that to build lookalike audiences. By the way, that's the use of machine learning again or AI to go into kind of the next level and scaled up campaign. So we're there. It just, it's just harder than people like to suggest it is. It's also why I think we see certain companies, especially those that confine themselves to kind of programmatic tv not running much of their in house measurement because it's got damn difficult.
- [27:15]
B
Yeah, that, yeah, that's very, that's revealing when you see that. I guess last thing, if there's one thing you want everybody to kind of take away from your news here, the publishers or brands, like what would that be?
- [27:27]
A
I think it's the realization that the biggest part of TV inventory and the most exciting part of TV inventory doesn't live in programmatic.
- [27:39]
B
It's a little bit of a reality check here for television.
- [27:43]
A
And if you weren't in there five weeks ago, it's understandable because it wasn't always possible. You had to be of a certain caliber, size, whatever prestige to get it. And if there's one thing I want people to take away is that that paradigm has shifted. That paradigm has shifted. And I think this is also why we see these top five publishers joining in with both feet. They get it. They're smart.
- [28:06]
B
Yeah. Yeah. They're revealing where they think this is going by being part of this together. That makes a lot of sense. Yep. All right. Fascinating stuff. Great overview. Thanks so much for your time. Let's, let's chat again down the road.
- [28:18]
A
You bet, man. Good to talk to you as always. Cheers, Mike.
- [28:21]
B
A big thanks to my guest this week, Tatari CEO Philip Engelbrecht, my partners at Inten iq. If you like this week's episode, please take a moment to rate and leave a review. We have lots more to bring you, so please hit that subscribe button and we'll see you next time for more on what's next in media. Thanks for listening.