← Marketecture: Get Smart. Fast.
Marketecture: Get Smart. Fast.
Episode 117: Chris Kane and Mike O'Sullivan debate supply and data
00:45:33
Loading summary
Transcript87 lines
- [00:00]
Ari Paparo
This episode is brought to you by Zeta Global. Do you know what it takes to transform marketing into a data driven profit center? Are you able to align the C suite around your AI vision and strategy? Zeta Global has the Playbook to help you get started Download Driving growth in the AI era today at ZetaCMO AI Book again, that's Zeta CMO.
- [00:32]
Eric Franchi
This podcast is brought to you by Audiohook, the leading independent audio DSP. Audio Hook has direct publisher integrations into all major podcast and streaming radio platforms, providing 40% more inventory than what could be accessed in omnichannel DSPs. What's more, audiobook has full transcripts on more than 90% of all podcast inventory, enabling advanced contextual targeting and brand suitability. Audio Hook is so confident that in addition to CPM buys, they offer the industry only pay for performance option where brands can scale audio and podcasting with peace of mind knowing they are only paying for outcomes. Visit audiohook.com to learn more. That's audiohook.com welcome to the market Tech Sure Podcast. This is Ari Papara. I'm joined by Eric Franchi. So today's episode we're going to play a great conversation we recorded in Architecture Live. It was one of the most popular sessions we had on the main stage where two really smart guys talk about really smart ad tech stuff. It's Mike O'Sullivan and Chris Kane from respectively the Trade Desk and Jounce, and they're talking about all kinds of stuff. They told me they were doing Frost v. Nixon for ad tech. So you can, if that makes sense to you. I don't know. Get out of the house. Eric, you saw, did you see this live when you were at the event?
- [02:01]
Chris Kane
Yeah. It was so good. It was architecture perfection. Like two of the smartest people on the deepest things in ad tech, just giving them the floor for an extended period of time. This is so good.
- [02:15]
Eric Franchi
Okay, well, we hope you enjoy that and then Eric and I will come back and do the refresh and talk about the news. Also, if you're listening to the podcast stream, you probably got a bonus episode from our new podcast from Architecture called the Brand Forum, where co founder Jeremy talks with Josh Palau and they bring in lots of interesting guests from the brand world. So one episode's out. Please go on Spotify or Apple TV and find the Brand Forum and hit that subscribe button to get updated. It's going to come out every other week and I hope you like it. All right, let's go right to the recording. Thanks.
- [02:53]
Mike O'Sullivan
I was going to Put that on loop on my iPad.
- [02:56]
Ari Paparo
Every time I listen to the podcast, I always think, who made that little beep boop, boop, boop, boop. And how much did Ari pay for? Because it sounds professional, but, like, I was like, where did he get that?
- [03:09]
Mike O'Sullivan
So let me try to set this thing up. Mike and I have been trading a zillion emails over the last couple of weeks about this discussion, because what we're going to talk about really is a debate that he and I have on an ongoing basis about signals in the bid stream. And so I'm going to take a swing at setting this up. And then, Mike, you can kind of add in or correct me if you see it differently, and then we'll get into the details. But the debate in general that we have and the debate that we're going to do our best to sort of have here is related to sort of what I think is some grayness in the way that publishers and their SSP partners communicate information about the available inventory in a bid request. And what Mike and I were talking about earlier was there's some ways of structuring signals in a bid request that make you a boy scout, and there's ways of structuring signals in a bid request that make you a fraudster. And that there's not a explicit binary separation of those two things. There's some spectrum, there's some gray area in between. And so what we kind of want to talk about is that gray area and sort of how to restore some trust in the integrity of signals that are in bid requests. Mike, anything you'd add to that?
- [04:17]
Ari Paparo
Yeah, I think that's right. I have a lot of sympathy for publishers. I think Chris is going to take more the sell side point of view, and I'm going to take more of the buy side point of view. But we both see both sides. I do have a lot of sympathy for publishers. And the one thing I would just underline is a lot of these decisions and topics we'll discuss, like id, bridging. I understand why publishers do it from a revenue perspective, right? Like that is the resource that they're focused on. But I do think there is another resource that they may not realize they're spending, which is this concept of trust. And so, like, how far you go into the gray maximizes revenue, but also spends a different currency, which is trust.
- [05:02]
Mike O'Sullivan
Trust is on a long timeline, though, and dollars on a very short timeline. There's, I think Andrew Casale gets credit for this thing. There's this line that I've heard, I think I've actually heard it mostly from you, that the bid request is a contract, that it is this sort of like commitment from the sell side to the buy side, that this is an accurate and honest expression of my available inventory. And so that's the thing that we're going to kind of push on a little bit. I think we'll start with one that hopefully is sort of illustrative for everybody, gives us something to agree on before we disagree on other things. One of the fields in the bid request for web inventory is the site dot domain field. It's supposed to declare the. The domain where the ad is going to run. If you lie about that, you get in trouble. I think we both agree that spoofing that piece of information is sort of clearly crossing that bright line into fraud territory. I guess the question that I have for you, Mike, is why is it so obvious to everyone in this room, to you, to me, that misrepresenting the site dot domain field is not in the gray area and is sort of like obviously fraudulent, obviously eligible for clawbacks, obviously outside of industry norms?
- [06:00]
Ari Paparo
Yeah. I think my view on that is actually tied to something that we've also discussed a lot, which is like publishers invest a lot in having quality experiences and quality media and the equity they've built up in a brand or what have you. For someone to basically usurp and misrepresent themselves as that, to me is a big reason why that is unacceptable, because this is not necessary. We use the term contract, but it's such misrepresentation because the entire opportunity is viewed through a different lens based on that one field. Right. So it's like I'm a buyer and I may like, you know, I use things like historically, what have been called whitelists, allow lists, block lists, and that is a big lever for choosing whether or not to buy and moving on and misdeclaring that domain. I think one triggers that publisher quality point, but then two is basically bypassing the buyer's filters.
- [07:04]
Mike O'Sullivan
I obviously agree with that. I'll build it out just a little bit more to maybe create some counterpoints for later. There's at least a decade, maybe two decades of the industry grappling with that issue and there's broad consensus that that's out of bounds. There's a bunch of specs that are meant to sort of protect specifically against this ads. Txt comes top of mind for me. There's a whole universe of verification companies that aren't perfect, of course, but try really hard to detect this sort of stuff, the spec is written in an extremely explicit way about. There's just no ambiguity of like, what site is this? All of those things to me sort of like lend themselves to a place where you can trust that signal. And I see very, very little examples of misrepresentation of the domain. I assume that checks out with what you see. Let's talk about an area where that I don't think lines up with any of the things I just said. Floor prices. If my actual price as a publisher is a 10 cent floor, I will sell the impression for 10, I won't sell it for 9. Really?
- [07:59]
Ari Paparo
Really.
- [07:59]
Mike O'Sullivan
I'll walk away at 9. Do you expect me as a publisher to express that 10 cent floor from your vantage point of the trade desk? Do you expect me to present a 10 cent floor in the bid request?
- [08:13]
Ari Paparo
I guess what I would love to see is if you're calling something a floor, make it be a floor. And so what floors are nowadays is essentially a bobbing lure in the ocean trying to get the DSP to bite. And so like how many publishers, I don't know how many publishers in the room. When was the last time you guys were like, we're gonna go set a bunch of floors across our SSPs, it doesn't happen. Okay, Correct. I stand corrected. But the exception proves the rule. What we would see when I was at an SSP is that floors would become fossilized and we're like, gee, this floo floor hasn't changed in two years. Oh my God. And we would have conversations like why even have floors? Why not let all of the demand move into the primary ad server and the top bid win or not win? Right? Because like the number of SSPs that publishers are using is skyrocketing. So you're doing 20, 30, 40 floor updates across all those SSPs and placements per day. We also see obviously the amount of dynamic floors that are present. So I guess the problem I have with this one is we would love if these were actually floors, but we see that a lot of times it's goal seeking and trying to goose out spend and it's an external expression of a margin optimization solution. And moreover, it just doesn't represent an actual floor.
- [09:39]
Mike O'Sullivan
Isn't that kind of the job of sellers to coax higher bids out of the buy side? Like if I'm in the real estate market, I'm not going to communicate to the market the absolute minimum price that I'll sell my home for. In fact, I'm going to hire an agent to Try to coax the highest possible bids out of buyers as soon as possible.
- [09:58]
Ari Paparo
Make a recommendation for OpenRTV 2.7 called recommended price or minimum bid for consideration or something like that. Don't hijack the floor object to specify that. I'm not saying sellers can't do those types of things. I'm just saying like, why is it in the floor object?
- [10:18]
Mike O'Sullivan
Is there a credible way for companies on the buy side to sort of change that incentive so that publishers are rationally incentivized to truly communicate the 10 cent value, not some inflated version of it?
- [10:33]
Ari Paparo
Yeah, I think there are ways. Like, I wonder, is there a future in like, I'm not saying don't do dynamic floors, but if I were a DSP or if I was a bidding engine not speaking on behalf of any large public companies here, but if I was a bidder and it's like, oh, this is floor type 1 publisher set versus floor type 2 systemic optimization. It's not like I won't bid when it's floor type 2, but I'll understand how to approach it differently. Versus floor type 1 is an actual expression from the publisher themselves. So I do think there are ways to sort of build our way out of this, I guess, on this particular object.
- [11:11]
Mike O'Sullivan
All I got one more for you on floors.
- [11:13]
Ari Paparo
Okay?
- [11:13]
Mike O'Sullivan
There's borrowing your language, floor type one, floor type two. There's publishers trying to sort of be advocates for driving up their own prices. And then there's SSPs sort of overriding what is communicated to them by publishers in order to further extract revenue from DSPs based on what you guys see through open path integrations, which I think has got to be pretty close to the source of truth for the type 1 situation. How much of the optimization of floors that you see in the bid stream is initiated by SSPs trying to modify the publisher's floor?
- [11:45]
Ari Paparo
I don't know if I have an exact number, but I would say it's very pervasive in terms of where the floor optimization is coming from. I would say for the most part it's software optimization. I also, again, I don't have a problem functionally with that type of feature. I'm just curious if it is their job, if I take what you just said and play it out, it's their job to extend and grow revenue. Great. How much of that additional floor revenue is going to the publisher?
- [12:16]
Mike O'Sullivan
Yeah, and I'm not sure there's a way to know, right? Like you know what you spent. The publisher knows what they've got, but it's really hard to reconcile that impression to impression. Okay, let's talk video signals. There's been a forever long debate in the industry about what constitutes in stream video. There was a redefinition of what in stream video means in 2022. I think this is going back to our domain swifting thing. To my mind this is a good example of where the number of words in the OpenRTB spec give you like a pretty good clue about how abused a signal has been. And the new definition of in stream is like so ridiculously explicit that I think speaks to sort of this stretching or misrepresentation over time. We're both in agreement of a variety of sort of best practices for how to transact video. So I'm not sure we've much of a debate on this one. But I do want to ask you about lessons learned. You were close to this thing. What did we learn as an industry from the redefinition of in stream signals or in general the redefinition of video signals? The way the whole migration happened. What are your takeaways from attempts to improve the trustworthiness of that signal and how should we be applying those sorts of lessons to other things that are currently unreliable signals in the midstream?
- [13:25]
Ari Paparo
Yeah, so I took two elements away. So one I had sent Chris actually we probably should have just kept this one slide but there is this CIA handbook on when your country has been invaded and how to be like a passive form of resistance against the regime. And it's like sign up for a bunch of committees and industry bodies in the interest of helping things along. Always ask questions, interject, constantly ask what does this mean to us? And blah blah, blah blah blah. And it was just like God feat. Lee feels like I'm like reading about like whether it's IAB or pre bid, like there are these situations and they're, you know, as Chris mentioned, like these institutions are very important but they can be gamed to the extent of like just outlast everyone else, extend it out and like keep the meetings going. What about this? Extend the language more ambiguous. So for me the first thing I learned is like there are weaknesses in our current approach to frameworks that can be exploited by motivated, you know, individuals and companies to sort of point one. But then point two, actually it made me sort of just like think about and honestly worry about, you know, the open Internet or open web as Ari calls it more broadly because, you know, there's this sort of gaming and back and forth about a specific field in video but it made me just realize like by default OpenRTB is open, right? Like it's default on. Like when you're buying it's like, oh, I added a publisher, added a new ad unit, I added something new here we've got a new property. And when you start to sort of like mix these things together, whether it's video placement declaration or another video one in banner video, right, that just sort of has been rising in popularity. And if we don't declare these things properly, like what's the terminus of this type of thing? We've got buyers who are default open, default on and we can't tell what's actually happening because of ambiguity in a field. The terminus is either one of two things. Buying systems become default closed and you got to basically ask to be added to an inclusion or worst case scenario. All the buyers are like, this is too complex. Let's just go to Walled Gardens.
- [15:51]
Mike O'Sullivan
From the perspective of a publisher, obviously both of those are quite alarming. I'm not sure what a publisher could reasonably do today though, other than perpetuate bad signals. Whether it's in banner video or whatever. It's one of these things where if I as a publisher am not stretching the truth on what I'm going to call in stream video or I'm not making my banner ads eligible for in banner video, no one's sort of like parading me up on stage here and sort of saying spend more money with these good guys. It's just that DSPs, Trade Desk and others send money to the publishers that are most aggressive. Yeah, how do I not do some of these things as a publisher?
- [16:31]
Ari Paparo
And this is where I think it's actually not just a sell side beat up session. Buyers have a significant responsibility here. This ultimately ends with buyers and they vote with their wallet and buyers need to be faster to react to these types of, you know, in the gray, I'll call them like exploits or you know, stretches of the imagination. Let's call it. Buyers need to be act faster and it's not about like publicly parading people up and making a big deal of it. It's like do we like this yes or no, let's move on and like bid accordingly.
- [17:07]
Mike O'Sullivan
I mean in a way that that framing makes it sound like default off might actually be what's necessary to create the right incentives that only verify there's no shortage of supply in this industry. Only verifiably high integrity supply chains get demand or get demand at sort of like full throttle. Is that where we're going I think.
- [17:27]
Ari Paparo
That is where we're going. I think that's different than default off slightly because the bid value can vary. It's not binary, necessarily. There is the decision to bid and then there's the value of the bid. What I. So I do sympathize and I am empathetic to publishers because you need to be in the gray, you know, to. To earn like that, you know, your revenue and especially if you have your competitors are in the gray.
- [17:55]
Mike O'Sullivan
Right.
- [17:55]
Ari Paparo
And so, like, ID bridging is a great example. Like, was I rewarded for not doing ID bridging? Oh, actually, the folks who were doing ID bridging, they were rewarded. So I think this is where I think buyers have a bigger role to play. And I don't necessarily mean, like, being on stage and things like that. It's more just on executing relentlessly, on finding the things that you want to value and the things you don't want to value. So there is no incentive for publishers to do it. And the last thing I would just say on the video side is if we can't tell with the video opportunities, like, hey, this property does in banner video, and they have a video player, it's like, well, there's a chance that this could be having two videos running simultaneously. Let's just not bid on any videos from this property, right? So because of the limitation, because it's been sort of hidden in this video declaration, we don't know if there's concurrent videos playing back in this dystopian experience, which I'm sure we've all had, when you load a webpage and you're like, God, with three videos playing simultaneously and your plane's plane, your laptop's about to take off because it's like a 4K ad in the bottom right or something running. So, like, the downside there is just like, if we can't. If you are, there's a downside to sort of saying, hey, we've blended this in banner video such that you can't tell versus the other video. And it's like, well, then what do you think we're going to do if we don't want in Banner video?
- [19:26]
Mike O'Sullivan
So keep playing that out. And you're welcome to switch this to bridging, which I think could also be a good example of it. It feels to me that there is this history, I was actually talking with somebody in the hall about this earlier, that there's this history of, okay, Boy Scouts are over here and the fraudsters are over here, and there's this gray thing in the middle. And I don't want to be a boy Scout. I don't want to be a fraudster. I got to find the right spot in the middle. I think there's even an argument, like, actually, it'd be way better to be close to the fraud side than the gray area side, because the worst thing that has ever happened with any of this stuff, multiple videos on a page, injection of user IDs in the bid stream, whatever, is a slap on the wrist. No one's ever been banished from this industry. And so I'll just wait for the slap on the wrist, and I'm going to take the revenue benefits as long as this game lasts, and then I'll take a half a step away from that.
- [20:12]
Ari Paparo
I think, like, if you ever. If you know Will Dougherty and you follow him at the next industry conference, I think you may change your mind on, like, whether anyone's been, you know, permanently punished. Because people are like, will, Will, can I have a minute? Like, I want to discuss, like, how do I get out of the penalty box or what's going on? So, like, I made slightly challenge on specifics, but I take your overall point around where the incentives are, and I think that just comes back to how quickly we need to execute on discovering these and just not make it an incentive for publishers. Let's invert it and take it out of all this negative stuff. I talked a little bit about this, if I can. Ari. Sorry. At the Trade Desk Forward event, where it's just like, we're investing so heavily and historically as an industry in detecting bad and identifying bad. And, like, this is, you know, deep gray versus light gray versus ash or whatever. And it's like, what is quality? What is super good bid requests look like? What do amazing publishers look like? Because let's give. If publishers are saying, does it even exist anymore, though?
- [21:18]
Mike O'Sullivan
Or is everybody sort of beaten into having to stretch the truth?
- [21:22]
Ari Paparo
I think, well, you know, knowing that it's easy to be cynical about our space, like, nobody ever got sort of, like, booed off a stage for being cynical about ad tech. But, you know, you guys affected monstrous change with mfa. So change is possible around quality. And I would argue, you know, if companies start saying, we think this is what great quality looks like and there's guidance to publishers, I do think that publishers are like, oh, I'll earn more if I have transaction ID appended. I'll earn more if my GPIDs aren't warping all over the field. Oh, and I can see it when I Turn it on the same way. I can see with ID bridging, I think there is a future there. But in both cases, the upside as well as downside, risk mitigation, we just have to cycle faster about making those changes visible to pubs.
- [22:17]
Mike O'Sullivan
We're almost at time. Quick last question. And obviously only share what you can, but you know, Sincera is one of the very few companies that have ever been brought into the trade desk externally. Presumably there's some interesting ideas you and the team have. Can you share anything about sort of the future of trade desk providing more of that feedback to publishers and the direction that you want to take so that publishers understand more specifically when their signals are sort of perceived to be useful versus, you know, questionable.
- [22:44]
Ari Paparo
Yeah. So you know, what I would say is the same thing I told Alice and Chef who's somewhere here is for those folks who know Sincera, like, we are very focused on publisher quality and giving feedback to publishers. And when we started we worked at the trade desk to do the first global 100s and P500 and publishers would come to us and be like, what do I need to do to get better? It's like, yes, I know I'm. I suck at something, but like, how do I get better? And we were focused on that problem. And what I would say is we are still extremely focused on that problem. The only thing that changed is the W2.
- [23:23]
Mike O'Sullivan
All right, thank you, Mike. Thanks for having us.
- [23:29]
Eric Franchi
This podcast is brought to you by JW P Kinetics. JWP provides the technology that manages and delivers video, maximizes visitor engagement and monetizes it through positive ad experiences. Find out more about how JWP kinetics can help your company achieve results@jwplayer.com that's jwplayer.com.
- [23:51]
Chris Kane
All right. And we are back. Welcome to the refresh everybody. A bunch of stuff to talk about today. So we will talk about the analytics report that dropped last Friday. We will talk about Applovin and then a little trend that we're picking up on fee wars amongst DSPs and those competing for ad budgets. So, bunch of stuff to get into here. All right, let's start with analytics. Christoph strikes again. So the firm alleged that verification firms are not catching ads that are being served to bots or otherwise invalid or non human traffic. Back up, 40% of the Internet is bots or non human. And the article claims that DVIS and Human, those were the three companies that were called out serve ads to these bots and or miss non human traffic. And the question is this is where a lot of the debate has been are brands paying for this and is this technology like not doing its job?
- [24:55]
Eric Franchi
Yeah. So there are distributed well known lists of bots. So you hear bots, you think bad stuff. It's not true. You know Google, Google search bot collects information so you could be in the search engine. There are hundreds, thousands of bots out there and there are lists that are provided to people of those bots so that you can know it's a bot. I think the allegation here is that these companies were not blocking bot traffic and bots were seeing ads even though it was in one sense very easy to detect them. But we don't know if anyone was paying for them. That's kind of a big open question. We also know that the providers, especially Human, has in the past said that they don't want to block bots because it would give a signal to the bot owners that they would use and alter their behavior and you're better off letting it through and then giving reports later on. So that's the two sides of the coin. And as with many Atalytics reports, and I think one of the criticisms of Atalytics, which is totally valid, is that it's very hard to see if this is a minor problem or a big problem. The volume numbers are essentially impossible to know unless you are in kind of an ad server situation. So it'd be interesting to hear if an ad server like a mediaocean innovate could come back to us and say here's how much of spend was wasted on these non blocked bots.
- [26:27]
Chris Kane
Yeah, that's right. AdExchange had a very deep article on this. We'll put it in the newsletter, we'll point people to it. And the verification companies, DV and I s both came out with responses relatively quickly and DV was very definitive and clear in their response, which is basically this. DV identifies bots nearly 100% of the time and brands don't pay for ads served to bots. Is response was different. It was a walkthrough of the technology. And on my read, and I only read it once, you know, they, they didn't make the same statement that brands are not paying for bot traffic. So that was the difference. And I don't think Human gave a statement at all.
- [27:08]
Eric Franchi
We've had Chris at the Market Extra Live. He does PR and mod up for a double verify and he's whipped them into shape and they are responsive, they're on top of things. I do question this assertion that brands don't pay for it because effectively a verification company like Double Verify isn't really in the flow of money. They can, they can tell the brand, don't pay for this, but whether the brand actually pays or not is a matter of negotiation and conversation. So in some cases, brands might actually be paying for it.
- [27:39]
Chris Kane
Yeah, I guess that's the right question. Okay, well, I think this is not the last we're going to hear about this issue. This is the last that we're going to hear from analytics.
- [27:49]
Eric Franchi
His whole ecosystem. He publishes a report, the Wall Street Journal gets a scoop, you and I get something to talk about, Double Verify PR gets something to talk about, and the world moves on.
- [27:59]
Chris Kane
And it always happens on a Friday, the day after we record. So the one feature request for Christoph is if you're going to do this, just push it up and have it drop on a Wednesday.
- [28:10]
Eric Franchi
That's, that's just my Wednesday, not Thursday. Don't deliver news on Thursday if you want to cover it in the market sector, but Wednesday, absolutely.
- [28:20]
Chris Kane
All right, cool. Let's move on. So Applovin, we've talked about Applovin a bunch over the course of the past year. Stock was up or is up still 700% despite it going down. And they've been under attack by short sellers. There was a report last week or the week prior from, I think it was called Muddy Waters with a bunch of allegations. Ari, you did an awesome breakdown, as usual of, you know, both sides of the Applovin argument, the bull and the. And the bear case. In the newsletter this week. I thought what was interesting is Applovin. So the CEO Adam Faroji has, you know, generally been kind of like not press shy. I don't think he's, you know, does press much, if at all. We haven't heard his voice all that much in the, in the market, despite the business just being on this incredible trajectory. He started making blog posts, writing blog posts explaining Applovitz business, explaining how things work and really being on the proactive here and outlining effectively, your. Your bookcase on it. I'll pull a quote from the. His blog post when we talk about it. So, you know, he's talking about how they got tens of billions of dollars of spend from, you know, sophisticated marketers. I'll start the quote. If we weren't delivering, advertisers would go bankrupt or stop paying. Yet our collections are rock solid. That's a testament to the sharp minds driving these decisions and the real value we provide.
- [29:43]
Eric Franchi
Yeah, I think the proof's in the pudding. If the customers are smart and they're not being defrauded, then the customer should be expected to move their budget where it works. We do know that mobile marketers are pretty smart and the mobile marketers are seeing these reports and they're looking for themselves. And there was sort of a little bit of a mob on Twitter who was pressure testing some of the Applovin results and they came back and said, eh, it's pretty good. So I think those facts are more important than many of these forensic analyses of exactly the way the data flows around. That said, at some point there could be some revelation that they are doing something that is wasting marketers money in some way. I don't want to say illegal or fraudulent. I want to say wasting marketers money because if a marketer finds out they're not getting the roas they think they're getting, they will dial back. And to my knowledge they haven't found that out yet. It's interesting. I was at a dinner last night that had a lot of media and finance people and I would say this was the most discussed topic that I heard, which I wouldn't expect if you hang out with a bunch of media Wall street people. But really Applovin and the short sellers. The general consensus from these business hedge fund analyst types of people was effectively exactly what we said. No smoking gun if the money keeps flowing. It sounds good to me. Let me know if Apple's going to do anything.
- [31:10]
Chris Kane
The other piece here that I think is newsworthy is outside of Adam started to go on the offensive. The company is going on the offensive. It has lawyered up and hired Quinn Emanuel to investigate the claims from the short sellers. So with this Muddy Waters report that came out last week basically said Applovin was misusing data and breaking rules set by some of the partners caused the stock to drop 20% and then Applovin announced that they were lowering up and the stock came back. So I think that the market likes what they're seeing out of the company. Keep an eye on it.
- [31:46]
Eric Franchi
Yeah, I never like the suing the short sellers approach, but it seems to be part of the playbook. Short sellers tend to be small guys and then they have to lawyer up and nothing good ever comes of this.
- [31:57]
Chris Kane
Whole thing unless you're a lawyer. And if you're a lawyer, you're getting paid.
- [32:02]
Eric Franchi
There was a little side drama on this whole thing, which was for some unknown reason Adam's blog post said written by Grok, which is like a couple of questions immediately come to mind, which is? First, why can't you write your own blog post? Secondly, even if you did use an AI to help write your blog post, why would you tell everybody? Third, why grok, of all AI systems is not known as the one people are using to cheat on their homework just yet. And the immediate speculation is this was to draw the attention of a certain ketamine influenced billionaire. And it did. That person who will remain nameless immediately was like, oh look, he wrote it using grok. And it's just a sign of our times that you have to wave a certain kind of flag in the air to get attention.
- [32:53]
Chris Kane
Not to give Chris too much more airtime, but he's the one, Chris Harhar, he's the one that pointed out that blog post, which was excellent, by the way. Very well written, very well written. There was the footnote and I don't know the exact language, but effectively something to the effect of kind of created by Grok and then Adam edited. Is that how it was written?
- [33:16]
Eric Franchi
Yeah. I don't remember exactly the wording, but I know if I was a major public company CEO, I would just be shitposting using GROK all day. It seems to help your stock price. There's a little correlation here.
- [33:27]
Chris Kane
Chris's point was this is an awesome post and this line right here is all that we're going to talk about. And it seems like that is. That is the case, we should talk about grok.
- [33:38]
Eric Franchi
Do we have to? Yes, I want your take on.
- [33:42]
Chris Kane
Yeah, I want your take on this. So X AI, which is the AI company that Elon Musk owns, acquired X last week. So basically it's a $45 billion price tag, including $12 billion in debt. And Xai is valued at $80 billion. Sure. What's your take on this whole thing?
- [34:00]
Eric Franchi
It's just financial engineering. Clearly we all know that Elon overpaid for x $45 billion price tag immediately wasn't worth that amount, maybe worth 20, and it had a lot of debt. It was looking like dead money for a long time. Then about a month and a half ago, I think they refinanced the debt at 45 billion. That was bolstered by improved vibes around the company. Probably related directly to the political situation, but who knows why. But somehow they got that 45 billion, which probably let some of those debt holders get them off the balance sheet, get out of the deal that they probably regretted. And then the follow up is to swap the equity and some of the debt for the significantly overpriced equity in the AI company because AI has got eye watering valuations right now. So you're able to do that, take the overpriced valuation and use it to bail out the underwater equity in such a way that kind of solves two problems at once.
- [35:05]
Chris Kane
Yeah, I mean, sounds like just good, smart deal making.
- [35:08]
Eric Franchi
I mean if anyone got screwed here, it would be the equity investors in Xai who invested at I guess an $85 billion valuation and then ended up owning a no growth social network. But it's not a clear cut case where it's a bad investment. It's just not exactly what they invested in. I want to talk about a different subject related to this, which is do we think that the data in X is strategically valuable to an AI company? Because I've actually heard the opposite from well informed sources saying there's corpuses of data that are vastly more valuable. There is data you can buy, be it Reddit or New York Times, whatever, if you want to train your model. And the X data is not that valuable. But I actually not sure I agree with that. But that's a point of view I've heard. What do you think, Eric?
- [36:01]
Chris Kane
Sounds like we might be aligned on this one. If there's value in what, I don't know what you would call this group, but maybe the smartest, maybe like, you know, the most Internet addicted, maybe the, you know, just the percentage of the people that are creating the majority of the content on X, if there's value in that data, if there's value in what people are talking about, if there's value in the sentiment, if there's value in the now to train an AI, like on the most immediate, you know, real time data, there's real value in X and then it implies that there's real value in X AI. I don't want to put more words in your mouth, but I think we're thinking about this similarly.
- [36:40]
Eric Franchi
I think that, yeah, I think that's right and I would say that whereas say a cursor AI has got the best data about coding and I assume they've licensed enormous amount of data about coding and as a result they have a good coding product and X does not have access to that necessarily unless they pay for it. What is the X data valuable for? Breaking news, et cetera. I wonder if you put aside costs and GPUs and stuff like that, which obviously costs will come down over time, could you use the X data to produce fully AI produced nightly newscast?
- [37:15]
Chris Kane
Of course.
- [37:16]
Eric Franchi
Right. Or in the morning, many People listen to NPR for their short version, 5, 10 minute program every morning. You could actually create a version of that entirely using xData.
- [37:28]
Chris Kane
You could probably have some very unique data that can inform financial decisions, stock trading, investing, like short term stuff, and let alone like what these models over time can point to in terms of trends. And then the other thing about XAI is they don't have to pay a dime for licensing outside of, you know, the $45 billion, you know, implicitly valuing the company. So there's, there's value there.
- [37:55]
Eric Franchi
On an inside baseball note, do you think Linda Yakarino still has a CEO title?
- [38:00]
Chris Kane
I haven't heard otherwise.
- [38:01]
Eric Franchi
Yeah. All right.
- [38:02]
Chris Kane
I mean, I would assume X continues as X as an operating entity and there wouldn't be consolidation on that front. And maybe this should segue into the other news of the week related to X. There's commercial deals to still be done. So this article made its way around. Basically, Omnicom is giving clients incentives to spend on X. So I'll pull a quote from the article. We can talk about it. Clients who did not spend on X in 2024 are offered 50% in added value, up to 200,000 for their investments on the platform during Q1 of 2025, according to a document called OMG X 2025-2026 Global Strategic Partnership. Give me your take on this.
- [38:46]
Eric Franchi
I don't really have a take unless I find out the reason this is happening. Is there some guarantee payment that they're trying to work through? Is there an incentive coming from X to them? Are there political motivations? I don't really know. I think it's pretty well established that advertising on X is a bad idea, at least on the main timeline. And I think that most people in the industry kind of agree with that. And you can tell if you're an X user when you look at the ads. And the ads are all the worst possible ads you could imagine, other than the special deals they have with like the NBA where you watch a clip and there's a pre roll. The standard ads on X, I think unambiguously are bad. So, you know, you need incentives to get back into it, I guess.
- [39:32]
Chris Kane
I read this and this just sounds like sales incentives. Again, not not knowing if there's anything behind it. Right. What was the character's name in the Godfather? Mo Green. So I'm the Mo Green of ad networks. I made my bones during the days where you would start the quarter at zero and you would try to figure out how to make your quarter. And oftentimes you would come up with incentive plans. Right. Added value. Free this, free that, discount on this, discount on that. When I read some of these quotes that are supposedly directly from the sales materials, that's what I took away.
- [40:06]
Eric Franchi
Yeah. All right. I guess I'm the Robert Duvall of ad tech. I'm not really sure. I'm definitely not tough enough to be the James Kahn and I'm not the Fredo. So it's a little tough to come around to where I fit in.
- [40:21]
Chris Kane
I'm not Mo Green, but I just, I remember the last. All right, I retract. Let's talk about incentives. One more topic here before we call it. So I'm seeing something out there and I want to get your take if this is a thing. So there are some incentives being pushed by big companies around demand side fees. So two articles this week. The first was universal ads which is the new platform from Comcast basically to, you know, for. For marketers and mid market customers to buy ctv. Led by James Burrow and the crew. They're charging zero demand side fees. Obviously there's fees for Freewheel and all that, but you know, example one, zero demand side fees for. For buying ctv. Example two, well written article on Digiday. We'll point to it in the newsletter about the DSP market, but most mostly about Amazon. Amazon lowering DSP rates to win new business. This seems to be a thing, right? Like two in a week kind of. What's your take?
- [41:21]
Eric Franchi
I don't know. The universal ads thing is funny for two reasons. The first is they're a seller. I mean it's NBC, right? So why would they charge DSP fees? They're a seller and.
- [41:31]
Chris Kane
Well, it's a platform.
- [41:32]
Eric Franchi
Yeah, they could sell, they have, but they set the rates. So like what's the point of a fee? If you set the CPM rates, you could just make the CPM higher and not charge a fee. It's sort of meaningless to say you don't have a fee just because they've carved that out as a new brand name. Also the second thing that's weird about it is the type of customers are going for those sort of performance Y E commerce ones don't care about fees at all. They care about ROAS. You can charge them 100%, not 100% fee. You can charge 99% fee. And if the ROAS is good, they keep advertising. The people who care about fees are the ones who don't have really good roas. Calculations meaning brands and brands response type campaigns. So I don't really think that's relevant. I think the Amazon sort of attack on the open web where they have said publicly they want to be the dominant DSP in open web that should send chills down a lot of people's spines because for so long we could say the Amazon dsp. Oh yeah, that's just for, you know, Amazon retail, retail media and then maybe do a little audience extension. They got Neil Richter, they have Brian Thomasette, they have a murderer's row of talent over there and they're going after it with very large budgets. And the Amazon, your margin is my opportunity. And that's maybe the biggest story I've heard in adtech in a while for sure.
- [42:54]
Chris Kane
So maybe the universal thing is a nothing burger. But if there's more and more of these instances of companies just saying we're not charging fees, we're going to be aggressive on fees, perhaps there's a meta developing here that's going to put pressure on fees and market and you know, it could spell, you know, difficulty for other DSPs.
- [43:16]
Eric Franchi
Yeah, I think the CTV market in particular has had fee pressure for a while because it just seems obvious you shouldn't be paying 15, 20% and so much of CTV is on programmatic, guaranteed and non biddable ways of transacting and as a result the percentage of media fee doesn't make as much sense. This is all pressure on like the trade desk who is 40% of the revenue I've heard is CTV and you know, fees start compressing in that area, it's no bueno.
- [43:45]
Chris Kane
Yeah, agreed. Hey, before we go, just one other funny quick thing that they hit the feed this week. So yesterday there was an article posted. Nielsen's patent lawsuit against Videoamp was dismissed. And Ross McCrae, the former CEO, founder, board member of Video amp, he posted on LinkedIn, I'll quote it verbatim. Haha. Boomers at Nielsen can't even win. The only thing they're good at anymore, crazy face emojis, smiling emojis.
- [44:13]
Eric Franchi
Love it.
- [44:14]
Chris Kane
I just thought it was funny.
- [44:15]
Eric Franchi
Love it. Boomers. Yeah. I have one patent to my name and it's from my time at Nielsen. Nielsen loves their patents. There are so many of them.
- [44:23]
Chris Kane
Yeah, so I don't know. It was funny. Wolfpoints are in the newsletter. Everybody get a laugh. And with that I think we got the news of the week.
- [44:30]
Eric Franchi
All right, this was awesome. Yeah, the refresh. We renamed it the Refresh. All right, this was awesome. We'll see you next week.
- [44:36]
Chris Kane
Take care everybody. Thank you for subscribing to marketecture. New interviews are added every week at.
- [44:42]
Ari Paparo
Marketecture TV and your favorite podcasting app.
- [44:54]
Eric Franchi
Thank you for listening to the Market Podcast. New episodes come out every Friday and an insightful vendor interview is published published each Monday. You can subscribe to our library of hundreds of executive interviews at marketecture tv. You can also sign up for free for our weekly newsletter with my original strategic insights on the week's news at News Market tv. And if you're feeling social, we operate a vibrant Slack community that you can apply to join@adtechgod.com.