Loading summary
Transcript65 lines
- [00:00]
Alan Chappelle
Alan I'm Alan Chappelle, host of Market, where we talk about the big issues impacting digital media, from privacy to antitrust and the impact of big tech. We take the big ideas and break them down to help you see the future of the ad space. Join me wherever you listen to podcasts. Welcome to the Monopoly Report. The Monopoly Report is dedicated to chronicling and analyzing the impact of antitrust and other regulations on the global advertising economy. If you are new to the Monopoly Report, you can subscribe to our weekly newsletter at Monopoly Market tv and you can check out all of the Monopoly report podcasts@monopolyreportpod.com I'm Alan Chappell. This week my guests are Erez Levin and Gareth Glazer. Gareth is a programmatic veteran and entrepreneur beginning his career at AppNexus in 2011. He sold his most recent project RTK to Rubicon in 2019 and served as the chair of Prebid while heading all of Magnite's header bidding products. He has a new startup called Gamera and writes a blog called Gareth hates adtech.com. you should definitely check that out. Arez spent much of his ad tech career at Google where he wore many hats sitting in between the supply and demand sides, ultimately focusing on and becoming a strong advocate for the industry to move towards quality based buying and measurement. He remains a vocal advocate for industry reforms towards quality, effectiveness and truth, and supports buyers and ad tech companies through his advisory practice, Emmett Advisory. So these are two hardened veterans in the ad tech space. Arez comes at this mostly from the buy side and Gareth comes from more of a sell side perspective. So a few weeks ago they were riffing on LinkedIn about whether we continue to need a separate tech stack for each of the buy side and sell side. Now that may seem like an out of character discussion to have here on the Monopoly Report, but I wanted to bring Erez and Gareth on because these types of discussions happen all the time and most of the time they happen in what I call a regulatory vacuum where things like privacy, public public policy, competition, legal issues that tend to underpin the physics of these types of discussions. They tend to be a bit under considered. So let's get to it. Hey guys, thanks for coming on the pod. How are you both doing?
- [02:37]
Erez Levin
Great.
- [02:38]
Alan Chappelle
So this is sort of a unique discussion for the Monopoly Report. This all started from a separate discussion that took place on LinkedIn which started with the assertion that DSPs and SSPs are converging. We've heard that before. And then they morphed into A debate about whether we even needed separate buy side and sell side ad tech. And so I want to start with a little bit of background. So what was the original impetus way back in the day for having a separate tech stack for each of the buy side and the sell side?
- [03:11]
Erez Levin
I'm not a historian and I didn't start sort of as, as the industry was coming up, but I think like most markets there was probably, and again, I don't know the origins, but it came about like you have an exchange and you have a buy side and you have a sell side and it just made sense for folks to service those different sides of the market. So I can't give a lot more details, maybe Gareth, you know, a little bit more of the history there.
- [03:37]
Gareth Glazer
Oh, I got an opinion on this one. So I actually think you're right. I think you're largely right, Right where there emerges a desire for somebody to service both sides of the market. But I think what it primarily came from was originally, and this is before real time bidding, the market consisted entirely of networks that were basically isolated pods of liquidity and real time bidding in its first incarnation, which is debatable, it was either ad ECN or Right Media Ad ECN being started by Bill Urschel and Jeff Green and its own exchange that got bought by Microsoft for 60 million bucks back in the day. Or Right Media, right, which got bought by Yahoo, because gosh, those were the giants of the the world back then. So my read is that those first platforms, especially Right Media, was built to be a market maker for networks. It wasn't built specifically for buyers, it wasn't built specifically for publishers. In fact, it was built almost entirely to help networks bridge their liquidity problems. But that brought the concept of cross network, so cross pool of inventory targeting into existence, where now a single network could set up a campaign that runs across all of these different pools of inventory. And that led to X +1 where someone was like, hey, wait, there's a marketing and there's an expertise opportunity here where now if I'm building a network, I don't have to worry about going in public closing publishers anymore, which I would have had to do, right? Like value click for them to run their campaigns. Back in the day, they had to dial for dollars and talk to publishers. And now that we had this pooled connected source of liquidity, a company could say, ooh, I'm going to differentiate myself and I'm going to come out and I'm going to say, listen, all that I do is represent the Interests of buyers. That's all that I do. And I'm really good at it. And I use a platform like this or I build my own bidder back then. This is before OpenRTB, so they're all building their own things. I'm going to build my own, like infrastructure to just run campaigns. I think that that was the emergence where we had networks and it was probably like a portion technology, but really I think it was marketing. It was, hey, I can differentiate my business here where there wasn't really a need so much as there was an opportunity for a company to come in and say, oh, this is all I do. You should work with me.
- [06:06]
Alan Chappelle
I feel like you've articulated very well the solution. But in my view, and again, I'm saying this is the lawyer. So like, I don't know what the heck I'm talking about. But the problem that was being addressed were really twofold. Number one, advertisers and publishers fundamentally don't trust each other. And so nobody wanted to rely on each other's numbers. Therefore, each needed to have some ability or independent counting methodology. And number two, advertisers, because there were multiple networks and multiple publishers, advertisers wanted to have some ability to measure and count across multiple campaigns, across multiple networks. And so therefore the advertiser absolutely needed to have a separate accounting methodology.
- [06:54]
Gareth Glazer
Yeah. So I want to distinguish between two pieces of technology. Right. The counting is handled by the ad server. So this is Dart for advertisers, Dart for publishers. Those are the tools that were responsible for all accounts. And there are actually discrepancies between the DSPs, right, the demand side platforms and the buy side ad servers. Where like dfa, the, like the buy side ad server will say, hey, you served this many ads, here's how they performed. Meanwhile, the DSP X +1 or Media Math goes well, actually, we think we serve this many ads and this is what I'm billing you for. Now they have to argue about it. So I don't think the problem that the DSPs and exchanges solve so much was counting that's remained the purview of the ad server. This was really like, like Erez said, running the campaigns, that was their job.
- [07:42]
Alan Chappelle
That's an interesting one too though. And boy, I'm going to throw us down a rat hole. But what the hell, this is my podcast. The reason you have DSPS in many respects is because of what happened with DoubleClick, DoubleClick and Abacus, with their little privacy fiasco circa 2000, caused the ad server to be stripped down in functionality. I think my read was they were very worried about it becoming, doing anything other than being a dumb hype. They didn't want them really creating profiles. And so because of that vacuum, that's vacuum the DSP stepped into, and that's where you get into all the targeting and all the other stuff.
- [08:20]
Gareth Glazer
Yeah, I'm not familiar with the abacus debacle, but that's like, I would totally buy that. Even though Google blew it up because they like integrated DV360 and DFA and made them all like one thing.
- [08:34]
Alan Chappelle
Yeah, we're really weird here how, like, we pay attention to things. Like the most important thing in the year 2000 was, oh my God, you know, you cannot, cannot, cannot merge people's real world identities with their surfing habits. And then like, you're, we're now in a world where like, what was it, 2016, it might have been a little earlier, where Google just said, oh, no, we're, we're doing that. And deal with it. Like, then drop the mic. And you're like, okay, me and Adam.
- [09:01]
Gareth Glazer
Heimlich talk all about, all the time about how like in the early, like 2010s, right, every ad technic contract that I dealt with, and I dealt with lots of them, was really explicit. You cannot pass me any pii. Like, you do not do it. If you give me any of it, I am not responsible in any way, shape or form. It's a violation of these terms like, blah, blah, blah. And now we live in this world where we're shooting around pii left and right. It's like, oh, we're going to use email addresses, we're going to use all this stuff as these, these companies historically have not trusted themselves to handle these things.
- [09:40]
Alan Chappelle
Yeah, that's a whole, that's a whole other pod. I mean, every, every ad tech company in history, their entire privacy program was in one sentence, we don't touch pii.
- [09:50]
Erez Levin
Yeah.
- [09:50]
Alan Chappelle
And like, that was it. All right. Okay.
- [09:52]
Gareth Glazer
Don't give it to us.
- [09:54]
Alan Chappelle
So. So I want to get us back though, on track a little bit. So. So, Gareth, you'd made the argument that the ad space doesn't really need DSPs and SSPs anymore. And I'm just curious, is that just a nomenclature issue or there is a reason that we don't need separate buy side and sell side tools. And does this get into what you're saying, like you're separating DSP from ad server?
- [10:15]
Gareth Glazer
Yeah, so my impetus for this, and I'll try and keep it short, is Pre bid and I like have a whole spiel about this, but Pre bid and SSPs have tremendous feature overlap at this point, an undeniable amount of feature overlap. And I think what we've come to realize is that the concept of protecting the interests of publishers is actually a relatively fraught one because in a flat first price auction environment, I'm not even sure what it means. Like I'm not sure what interests are being protected nor even if they did exist, even if someone could explain to me really well, oh well, here's what we're doing to protect the interests of publishers. Do I trust a company that's being paid on a revenue share to do that? I'm not sure that I do. So the trajectory that we're now seeing is because of the functionality duplication that exists between exchanges and prebit, we are migrating away from SSPs being these guardians of publisher interest and pre bid just being this nice open source mechanism that allows flat auctions where publishers are kind of coming to the realization that's like, oh, you know what, as long as I keep everyone bidding in the same auction, that's all I really need for Programmatic. That's the extent that my interests need to be protected. And that's why I think that the distinction is going away. Because SSPs are seeing this and panicking and going, oh my goodness, like we have Open path bidding directly into pre bid. That's not good for me. I don't like that they're going around me. So now I need to basically build DSP functionality because Open Path is bidding right into pre bid and we're entering this world where there's just pre bid bidders and that's like kind of the future.
- [12:03]
Erez Levin
And again, I think this does come down to nomenclature. And so I think of SSP by its sort of literal, you know, the word supply side platform and DSP is demand side platform. So so two sort of technology layers and entities that are really representing one side of the market or the other. And so in that sense I still think that there is that need. And so I don't know if pre bid falls on one or the other or it's neither one of my sort of views, and this is a bigger thesis that I'm still sort of working on is the concept of markets, generally our market, but all markets, two sided, multi sided, being cooperative or being adversarial and there's sort of a range. And as I've done little research, I honestly came into this, this industry, you know, programmatic was this cat and mouse game. So it's, it's inherently this adversarial thing. And I almost assumed that all markets were adversarial to some degree. And a, what I've learned is that's not, not necessarily true. A lot of markets, and maybe even most markets are more cooperative than they are adversarial. And even advertising was more cooperative before it became digitized, before we got into this sort of automated real time system which almost inherently made it much more adversarial. And so my view is, yes, there's going to be a part of our market that returns, right, deals, premium inventory, really scarce, that is much more cooperative, that doesn't need to live in this adversarial world where you don't need a separate buy and sell side, you just need sort of these two. You know, you can have the publisher and advertiser talk and have one technology that's helping facilitate that. But I still think the majority of the market, this sort of really liquid market, requires different entities representing each of those different sides because they're not just adversarial with each other, but with their peers. Right? Buyers are adversarial with their competitors and they're trying their, you know, their advertising competitors and trying to get an edge over them on access and price and then same thing on the publisher side.
- [14:11]
Alan Chappelle
So does that make sense to you, Gareth?
- [14:13]
Gareth Glazer
Yeah, I think that the world that's scary in the programmatically traded world is one where demand is really consolidated. Because if demand is really consolidated in a programmatic environment, they can set prices and they can set prices that are divorced from the value generated by the inventory. And don't get me wrong, there's a level of that that happens everywhere. But that's where you see Google taking a 50 or 60% margin, right? Because they can, they know, oh my post click reporting for this advertiser on GDN is driving a net effective CPM of $10. But I know I can get this inventory from this publisher at $4 and there's not really another game in town. So I'm going to pay them the 4. Where things become more collaborative, where there are economies of scale is when there's competition on the buy side. Because where that 104 delta gets squeezed is when there are three other companies capable of running those campaigns for those advertisers and are saying, well, we know we're making 10, so we're going to bid 6, we're going to bid 7, we're going to bid 8. And pre bid is a fantastic facilitator of that. So my fear and what I think, where I think this, this becomes really honestly cooperative for independent ad tech is we all need websites to survive if we're going to have an independent web websites survive by making higher prices from their ads. The way to make higher prices from their ads is competition. The more that I think about any other solution to make more money from your ads, anything that's not competition is just scary to me it seems fragile.
- [15:52]
Alan Chappelle
But to your point, that's sort of the trouble here is that we might very well have some competition on the sell side if the DOJ has its way. But the sad reality about what's going on with the DOJ remedies in the ad tech case is that they almost entirely keep off of the buy side. And so you may find yourself to have additional competition on the sell side, but you're still going to have Google owning all of this demand. And I'm not sure where we go with that, but Eric?
- [16:23]
Erez Levin
Yeah, no, just on the last point. So of course if all, if all the demand is concentrated in one player, that's really bad. I don't see that happening. Right. There's going to be enough sort of different buy side competitors and demand coming together. Maybe it will consolidate, but it's not going to be one and probably not two or three. I think the factor that we've not really thought about as much and something I think about a lot is the scarcity of supply right now. The buy side does not treat the supply side as a scarce commodity. It is right, especially higher quality inventory time attention. Those are all really scarce, but we don't treat them as scarce commodity because you can always create a replica, another cookie, another, you know, real or not, and more impressions and stuff them on a page. And so that to me is actually one of the bigger issues when I don't think the buy side can set prices if there's scarce supply. And I think this is actually going to move the power is going to move more into the sell side that has some of that scarce, higher value, higher quality supply.
- [17:32]
Alan Chappelle
Sounds like you disagree, Gareth.
- [17:35]
Gareth Glazer
Well, I happen to be in a company so shamelessly plugging myself, my business is the evaluation and the quantification of the quality of publisher inventory. That's what I do. So I do it without judgment though. And I think that that's really important because the, when I say without judgment, I mean we should be working to quantify the components of inventory just as accurately as possible and then making them optimizable. Nodes in campaigns. The reason that I draw a distinction between this and attention and the concept of quality is because I think that we have this really nasty habit in ad tech of replacing old gods with new ones and not thinking about why we are doing the new thing and what the actual components of the new thing are. And so I trade in components. But what I see happening, and I worry about this, this is my. Because I actually like really appreciate attention for what it is. I appreciate the attention like economy. I appreciate them moving towards it. But I worry that we are just making a new viewability and we're going to have buyers buying cookie IDs with attention scores. And that's what we're going to. And that's where we end up. And that's like the new thing. And that doesn't solve the issue because I do think if we fix the buy side, right, and this is the big if, if we get the buy side actually optimizing to incremental business outcomes, whatever those may be, however accurately they're measured, media mix, marketing, whatever it is, competition increases properly for that inventory you're talking about. Because I definitely think then I. And I get a lot of arguments about this. I just got into one button in the beer slack. I don't think humans should be setting prices and sellers in terms of pricing. When they do it by hand, they tend to not do it very sophisticatedly or in a very sophisticated way. They tend to pull a number out of a hat. And that's how advertising worked, right? Like a long time ago, they would pull a number out of a hat, take it to market, see how much they had, and then, well, if lots of people wanted it, it's like, all right, great. Lots of people are calling my phone up, numbers going higher, like. But in the digital world, when you look at the average CPM on Facebook, it's crazy high. Like 20, $30. Things that are like that don't make sense in Banner World. And that tells me that if competition is there properly, everyone could be making way more money. And our problem is that the buy side is not assessing things competitively. Therefore they're not willing to pay the high numbers. But if we can fix it and if we can make that happen, that will get us the high numbers. But I'm not a big power dynamic guy. I'm a big competition guy.
- [20:21]
Alan Chappelle
But how do we fix it? I've got two really smart people here, like, how does this get fixed?
- [20:27]
Erez Levin
I. So I have a couple of thoughts on what Gareth. Gareth was saying and maybe it's related. One, I just don't like trying to replicate what Facebook and Google are doing. I don't think it's possible. One, I think they have way more intense data, not just audience data, but intent data that will be very hard to replicate at scale across the web. And two, they have figured out the best way to take credit for outcomes. Some of them which, you know, are actually incremental and some of them were not. And so. And I think that is also going to be very hard to replicate. So I generally say we should move in a different direction. Going back to your other point, you know, as one of the people who was sort of championing attention for a very long time, I pivoted a few years ago to talk about quality. Even when I was talking about attention, I actually am still against a standard for tension because then it become another viewability, it would eventually get gamed. What my sort of big focus as a, as a marketer, as somebody who sort of thinks about the effectiveness of marketing short and long term value. Most media should be thought of in terms of its relative value, not in terms of its absolute value. And so yeah, humans shouldn't be setting prices. They can set a maximum sort of price for an impression or types of media based on what they think, you know, conversion rates and brand lift and things like that and how that's going to affect their, you know, return on ad spend. But generally speaking, they should be thinking about the relative value of that media and going back to your competition guy, so am I, I'm a supply and demand guy. And so based on the available supply, based on the demand for that supply, the systems should be deciding how to price that, right? How to find the best price available to fulfill whatever the objectives of that campaign are.
- [22:20]
Gareth Glazer
We're comically aligned on the second one, so I'm just going to leave that one. But the first one. So I've done a lot of media buying on Facebook for various things and a lot of my buddies are in D2C. I am not so sure that their data is so good. And don't get me wrong, my reason for this is a simple one and it is one, their internal consistency in reporting, their engineering generally across the board for their ad system is like a albatross. Like they really, it's, it's not a reliable system. But two is if you set up the same campaign five different times, same creative, same everything, performance will vary wildly. And what that tells me is that the chance component means that there is not so much like proactive intelligence when you're setting up a campaign and they have like this data set is going to work for this campaign, they have data sets that are underneath the hood, they throw the can paint out there and they, you, if you get a good one, you get hot and you go, you don't, you don't. So I think we can take them.
- [23:29]
Alan Chappelle
So here's, here's two things I would just challenge. Number one, if you compare Meta and Google in terms of data quality and depth to any publisher anywhere, they're a hundred times better. So you might be right that meta's data is kind of crappy, but it's all relative because as compared to like the New York Times, not to pick on them, but like meta's data is significantly better. But that's not really my number one point. My point is that you're pointing to campaigns and you're saying the same campaign conducted four times and has widely variant outcomes. Isn't that less an evidence of bad data and the fact that the algorithm is really weird and it may not be working in your best interests?
- [24:16]
Gareth Glazer
Well, on your first point, I'm not a believer in publisher data and this is, well, I am as it pertains to the quality of the inventory. But I think publishers have been sold fool's gold for a while here with people telling them your savior is your first party data. I just don't think, I think that your point is super salient like the, the, even the New York Times who probably has spectacular data, it doesn't compare to intent data from the buy side now. But I do think we have independent players in independent ad tech doing things to get intent data right, harvesting, search, harvesting credit card logs, harvesting all sorts of stuff. We do all sorts of stuff. And I do think there's some great data out there, there. I think it exists. I wish I had a company to shill for this, but I don't. But I know that there are some really intelligent like user data, intent data companies out there in independent ad tech. I, I, I believe I am.
- [25:12]
Alan Chappelle
I think that they all bow at the feet of the masters and the two masters right now are Google and Meta.
- [25:17]
Gareth Glazer
Yeah, I mean Google, I don't think, I don't think anyone's ever gonna, if they're pulling search data across.
- [25:22]
Erez Levin
I don't, I mean, but just keep.
- [25:25]
Alan Chappelle
In mind, but, but I, I actually know, I want to harp on the second point though, Nera, not to cut you off but like, but like you, you're, you're holding that out as a data quality problem and I think it could be evidence of a whole bunch of other problems but, but really underlying a trust problem. Because if you're holding four campaigns and having four wildly different outcomes with the same creative something is amiss and data would not be the first thing I would look at.
- [25:48]
Gareth Glazer
No, I tend not to blame malice where incompetence is explanation. I think it's a confidence. I really think that, I really think there is an optimal way for that campaign to run and if they could get it fast, they would.
- [26:02]
Erez Levin
So just on that point, you know, I've long believed that we've overinvested in Dr. Right. And money needs to move back towards sort of mid and upper funnel. And that also means that intent data has been overvalued. It's not been that incremental. And that includes on the walled gardens and off the walled gardens. And as we sort of realize it's not been that incremental, it's something that, you know, the walled gardens do. And unfortunately the industry is sort of so hypnotized by averages, right. They look at average click through rates and average conversions and things like that, but there's a long tail of stuff that doesn't actually work. But on the whole the campaign did so they could actually cut off a lot of that fat. But I think as things move towards that mid and upper funnel, that's where I would challenge you, Gareth, that I think that publisher data, it's not very rich in intent signals, but it's still more valuable than potentially other, first of all, other inventory that maybe doesn't have even basic demographic data or some other sort of insights that you might know about a user. And even more so than the walled gardens, which will continue to try to just sort of take credit for, for outcomes that were likely going to happen anyway.
- [27:12]
Alan Chappelle
So how much is. And you just touched on this, Erez. So how much is attribution theft or just the ability to sit. If you have a big enough footprint out there within the interwebs, what's your ability to claim credit for things that you might not deserve credit for? Like how much is attribution theft a thing and how much is that? How much is that? Like if we could solve for that today, how much of that would address some of the other concerns that we're sort of quibbling over?
- [27:40]
Erez Levin
I'm biased, but I think it's a very big thing. How much? I don't know. But in the, you know, double digit percentages easily. And I think it is, it sort of corrupts everything Right. It just sort of steals money from other opportunities, other valuable media that could not be attributed to an outcome. Right. As easily or something else was able to take credit for it more easily. And so I think there's going to be sort of a big reset of sorts when, if and when we move folks off of sort of that obsession with attribution. This is a bit of an aside, but I think it's still relevant. I ran a poll last week, maybe you saw it. And I asked people, of all of the purchases they make in a given day, what percentage do they think were influenced by an ad that they saw in the previous seven days? And it was an overwhelming, like 80 plus percent said less than 5% of the things that they bought on a given day, in a given week were influenced by an ad that they saw in that seven days before. Most of the things that we buy is not just based on that last click. It's not that thing that you saw on Meta. You're spending, you know, 100, 100% of whatever you're, you're spending is for that year. Most of those things are ads that you've been seeing for months and years. And so those are things that you cannot do attribution for. Those are the things that we need to understand, not just tv, but billboards and also what car your grandfather drove. And those are the things that are getting undervalued today in our marketing system.
- [29:19]
Alan Chappelle
Gareth. Agree, disagree.
- [29:20]
Gareth Glazer
Yeah, so I think that there are, there are two separate parts of the market to consider here. One part is what I refer to as boardroom budgets. So boardroom budgets are large brands that have significant margin marketing budgets deployed across lots and lots of channels for them, I completely agree. If they're running on Facebook as one of their channels or Google is one of their channels and they're looking at post click conversions or they're like just looking at platform numbers in Facebook or in, in Google. Yes, it is a problem. I am not sure, however, the extent to which the big brands actually do this because my understanding is people who are spending, you know, Once you're over $50 million a year in marketing spend, get tend to get smart and know that, hey, if I count up all my conversion numbers in Facebook and all my conversion numbers in Google, it's bigger than our sales as a company for the year. So, so something I can confirm, I.
- [30:25]
Erez Levin
Could my Instagram, I can confirm it on interrupt, but I can confirm that that is not accurate. There are certainly some, I call them enterprise buyers that are buying the smart way. But there are a lot that are wasting their money. Right. You know, my passion of, of fake in stream. There's a lot of very big buyers that are buying this stuff and overpaying for it in a very large magnitude and valuing things according to view through conversions, viewability and all that kind of stuff. And so they are not immune.
- [30:51]
Alan Chappelle
I have one metric for you. And this sort of, I think, drives home the whole, the whole shebang. The people who are writing all of the checks have almost zero influence in how the marketplace is set up.
- [31:03]
Erez Levin
So.
- [31:04]
Alan Chappelle
So I don't know if you want to call that smart.
- [31:06]
Erez Levin
Well, they're not writing all the. Right. The SMBs. And I want to hear the rest of your thesis, Gareth, because I think you're going to Contrast that to SMBs, which I think is a very fair sort of split gun.
- [31:15]
Gareth Glazer
That's exactly right. And when I talk about going after Facebook and Google money, I'm talking about the SMBs. I'm talking about single channel advertisers who only run their marketing because their budgets are not huge. They spend a couple million dollars a year tops. Those are the big ones. The small ones are spending $50,000 a month and they just run on Facebook and it does increase their sales. I do not think that the incrementality problem is as pronounced for people like that. I think it's a lot smaller and I think that's the market that I get really excited about because it's a big market and they don't run programmatic because they log into a DSP and the DSP will say it drove no conversions. And that's a whole separate discussion about quality. But like I, I think it. That's so I think that that's my thing.
- [32:05]
Erez Levin
I totally agree, Gareth. I always try to separate those two things. I focused more on the enterprise because I think about branding and things like that. But even within the SMB, and even if you're talking about buyers that don't do programmatic, how many of them, and I don't know the answer, how many of them are truly single channel? How many of them are really just Facebook or just Google or how many of them are Facebook and Google and maybe, you know, one of the other self service platforms or doing direct mail because as soon as they're doing that still Facebook is taking credit for a lot of the direct mail that's also doing a lot of the work. This is where I'm concerned about sort of the performance, CTV for SMBs. If it's the only channel they're doing. Maybe they can show that there was a lift, but it's going to be really, really hard for them to if they're doing all. Any of these other channels as well.
- [32:52]
Gareth Glazer
I totally agree. So my experience in general is with dtc and most DTC companies that I know have a channel that they ride. And the reason that they do that is they're people who invest in what works and their appetite for trying new things that don't work out of the gate if they have something that's working is very, very small. So I think that normally they'll hit a certain amount of critical mass where they're like, I can't scale anymore. My CAC is going up now I'm looking for more channels. And now I look into like the multiple attribution problem. But yes. Oh my gosh. Do I share my concerns for performance? Ctv, like, wow.
- [33:28]
Erez Levin
But, but again, so, so, yeah. So going back. So I, I think there's segments of the market, like you're saying dtc, which absolutely, like, makes sense. I just imagine that is a much smaller part of the market than my local plumber or pizza shop or somewhere else where they're still doing a lot of traditional. They're also doing a little bit of, you know, display on Facebook or whatever that is. And the factors get a little bit confusing there, especially for an SMB that doesn't know how to, to properly look at all that.
- [33:55]
Gareth Glazer
There's probably a huge opportunity here to go in and educate them and like, win SMB dollars. I really want it. I want, I want independent ad tech to be the thing that does it because I just think that it's become the easy button because the first thing to do was always to advertise on Google and because, hey, Google search, right? Someone's searching for a plumber. If I can rank locally for plumber, Google's sending me leads. They are doing it. And I fully believe that that worked. And Google could just use that wedge to be like, all right, now I'm going to sell you all my other things. But yeah, I think we're in agreement on the SMBs.
- [34:31]
Erez Levin
And I try, I try to look.
- [34:32]
Gareth Glazer
Maybe I have too much faith.
- [34:33]
Erez Levin
I try not to use searches in the example too much because it is just a totally different paradigm. Advertising as a pull versus a push advertising. It's not even akin to marketing. So it's worth sort of thinking about very separately too.
- [34:47]
Alan Chappelle
So this has been a fascinating discussion. I got one more question for you guys. I'm going to ask for Your reaction to something. So in a recent Monopoly Report newsletter, I had openly questioned why little ad tech, independent ad tech, I think Garrett, the term you used. Why does little ad tech need to move to an ID less targeting world when big tech effectively gets to continue to build a footprint that in many instances leads to real world identities? I mean if we're, we're trying to hold out, the walled gardens is offering better outcomes for advertisers. Why not just try to emulate some of the practices? Why, why the push to be cleaner than clean when you know, the direct competition has, has, you know, has got a howitzer to mix metaphors. Anybody want to jump into that?
- [35:31]
Gareth Glazer
Oh, I'll gladly do it. I'm very opinionated about this one. I think that the privacy like campaign, this is conspiracy theory, tinfoil hat, but privacy. The entire push for privacy has benefited Google and Facebook. Don't get me wrong. The fines have been unpleasant for them, but the additional regulation hurts little ad tech.
- [35:52]
Alan Chappelle
Cost of doing business, my friend. That's just a, that's a rounding error.
- [35:56]
Gareth Glazer
I totally agree. And it just destroys a little ad tech because if we're not allowed to have IDs, meanwhile they have Gmail, right? They have logging Facebook makes users log in. They have tremendous ability to get consent. Like really coercive ability to get consent. Like they, they can like really get users to give them everything. So the entire privacy initiative where like users need to have all this say and we do all this TCS stuff, like my gosh, I'm, I'm applying for a GVL ID right now. It's wild. And it just, the, the whole thing is, it's, it's, it shores up their moats, it makes them more powerful. Privacy, that is a sword for big tech. It's not a, something that puts them on the defensive by any stretch of the imagination. So I have been very vocal in. We need to stop being scared of IDs. Like, there's nothing wrong with, with cookies. Like, in fact they're, they're remarkably anonymous compared to the alternatives. Like, and so.
- [36:54]
Alan Chappelle
Okay, Eris, I would love your reaction.
- [36:56]
Erez Levin
So yeah, I agree with that piece. My perspective is that data is not as valuable for the walled gardens as we pretend it is. Right Outside of the DTCs, the single channel advertisers that can really be confident in whatever attribution is truly incremental. The vast majority of advertisers, from big enterprises to even larger, you know, medium sized businesses are advertising on multiple channels. And so having an ID in one place that you cannot connect to all the other touch points like you could with a sort of, kind of could with the cookie before. It doesn't make it useless. But let's not ascribe too much sort of value and power to it, when in reality buyers are going to have to find all kinds of different ways to do measurement and understand the value of different media, which a lot of the walled gardens, a lot of their inventory is lower quality, it's longer tail just because as an id, it's still a small little ad, it's still something, you know, it's still music videos playing on the background of my, my TV that I'm not even in the room for. And so I think there's other ways that we can assign value beyond the IDs that I'll actually show that some of the, the open Internet is more valuable than what's in these walled gardens, even with their IDs.
- [38:09]
Alan Chappelle
Okay, so it sounds like we're actually in more agreement on this than I thought we would be. The one bit of daylight I would say, Erez, is I agree with your sentiment, but if you take it to its logical conclusion, which is that a idless environment is at best, what, 75% as effective as an ID environment. I mean, those are all the sandbox tests and those are the ones that were more beneficial to the sandbox. Like that's a huge delta. And so I guess my, my I would still go back to my question is like, why in the world is little ad tech talking about moving away from IDs to protect privacy when the, our biggest competitors are, are, you know, have IDs and actually have IDs that are linked to identity. If it was a 95 to 100% Delta, I feel like you can make a moral argument that you go ID free, but if it's a 60% to 100% argument in terms of effectiveness, really hard to make that argument.
- [39:10]
Erez Levin
No, I, I, I think the delta is not as big as sort of, you know, various tests that had flaws in them tell. And again, I think there's other variables. It's not just the id, but ultimately I do think that the premium Internet, the premium open Internet, will move into a much more sort of authenticated, logged in sort of world. And so they'll be able to compete in theory on just sort of having an id, which will have some benefits, but not in terms of attribution.
- [39:41]
Alan Chappelle
Okay, well, that makes sense. Gareth, what's your response?
- [39:44]
Gareth Glazer
I mean, there's wishing the world was a better place. And there's dealing with it as it is now, if advertisers were successfully convinced to be smarter about how they attribute things like, yes, we could move away from cookies and we could move away from IDs, but until that happens, the destruction that will be reaped upon little ad tech by abandoning IDs will be so profound. Like, like, we will be in such trouble that I just. We, we need to do everything we can to remind the public, I think that there's like a public sentiment here. There's also a lobbying component like, that this initiative is not. It says it's for users, but it's not like what this is. This is for monopolists to increase their moat. And we, we have to keep banging that drum.
- [40:46]
Alan Chappelle
We need more people saying that that's the one. If there's one thing, there are so many separate instances where privacy has been shown to be a canard of big tech. And yet I feel like I'm screaming into the ether when I, when I, when I say this. But like, more people need to recognize that because there's this weird, like, oh, no, we can't. We got to stay away from any type of UID or, you know, anything that doesn't use cookies is fingerprinting. And then like, I'm like, well, how do you define fingerprinting? And somebody points me to a, to a definition created by several browser companies and it's like, yeah, of course they think that's fingerprinting. Like, are you kidding me? So I, if I could wave a magic wand, two things would happen. I'm just going to throw it. Maybe we're going to end with this because what the hell, it's my pod. Number one, I would like to see the net net result of a Chrome divestment be a requirement to include a idfa, like pseudonymous identifier that can be used by multiple entities. That's number one. Number two, somebody has got to give Rick Bruner, like four or five million dollars to do some research on the attribution and the measurement side of this. Not using IDs, because I do think that, that there is some value that, that, hell, we're 25 years into this and it's still underdeveloped. We need, we need more research into, you know, research and measurement using either zip codes or some flavor of. Of without using it.
- [42:11]
Erez Levin
I think there's advertisers doing, I think there are advertisers doing it. They're just doing it quietly. So it's happening the transition is going to. We just need them to accelerate it and actually talk about it.
- [42:22]
Alan Chappelle
Amen brother. Hey guys, this has been a great discussion. I really appreciate you coming on. I was really hoping to get you guys arguing a little bit more, but but I still think it was an interesting debate, so thanks so much. I really enjoyed that discussion. We each come at the digital media world from a different perspective and I thought going in that there might be more daylight as between our respective positions, as it did feel like we were agreeing more often than not in this discussion. I do have a few observations. I was struck by Gareth's claim that there is so much variance in terms of campaign metrics in walledgard and ad campaigns. Now, assuming that's true, I am pretty skeptical that the root cause is simply a data quality issue. I guess you can judge for yourself. And I was happy to find at least some common ground on the idea that moving away from pseudonymous UIDs when the walled gardens are not playing by those same rules could have a catastrophic impact on little adtech and even some publishers. And Arez made a great point in that there are at least some brand advertisers who have figured out compelling measurement models, but they just aren't talking about them yet. So here's hoping that we see more advertisers, particularly the larger ones, take a more vocal approach in touting what's working as as that will help shape the marketplace towards quality. We've got a bunch of other fantastic guests coming up on the Monopoly Report podcast over the next few weeks, including a discussion with Arie Papero talking about his new book Yield. I can promise you that my interview with Arie will be unique. I've got some great questions in mind. Please subscribe to the show@monopolyreportpod.com or on Spotify, Apple, YouTube, or wherever you listen to your podcasts. And thanks for listening. Thank you for listening to the Market podcast. New episodes come out every Friday and an insightful vendor interview is published each Monday. You can subscribe to our library of hundreds of executive interviews at Markitecture 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. And if you're feeling social, we operate a vibrant Slack community that you can apply to join@adtechgod.com.