← Marketecture: Get Smart. Fast.
Marketecture: Get Smart. Fast.
Episode 137: The big debate around Transaction IDs with Chris Kane and Mike O’Sullivan
01:12:31
Loading summary
Transcript189 lines
- [00:01]
A
This podcast is brought to you by Incremental. You ever run a holdout test, wait between two and five weeks for the results, and then realize you just wasted a big chunk of budget? By the time the data comes in, the campaign's over, the money's gone, and the insight's basically a post mortem. That's why I like what Incrementl is doing. They give you always on incrementality measurement. No experiments, no holdouts, no delays. You get answers while the campaign is live and so you can actually make changes, not just analyze the wreckage. With Incremental, you don't just learn what worked, you do something about it. Check them out at the Future of Measurement. Hey, this is Ari with Markitecture, and unless you've been living under a rock, you've probably heard that Markitecture Live is coming up October 27th in New York City. The last Markitecture Live was sold out and this one will surely be as well. With speakers like Mark Grether of PayPal, Eric Seufert of Mobile Dev Memo, and Jenny Wall from videoamp. Plus, I'll be recording my podcast live with the one and only Antonio Garcia Martinez, author of Chaos Monkeys and now part of the team building at Coinbase. It's a stacked agenda and we hope to see you there. Go to market and grab your ticket while they're still available. That's marketexturelive.com welcome to the Market Exer podcast. This is Ari Paparo. I'm here with Eric Franchi and I was just reminding him we're on video. So we need to look good now, but Eric doesn't have a problem with that. I need some grooming.
- [01:41]
B
I comb my hair.
- [01:42]
A
Yeah, that's important. So if you're listening to us on audio, that's great, but if you want to see our smiling faces, you can also go to YouTube. So this is an interesting episode. I want to, like, set this up because we turned into part of the story here. So we're recording on Thursday, August 28th. And what happened here? Let's go to a timeline. So about two weeks ago, Michael o', Sullivan, who's been on the show multiple times, he's a founder of Sincera, he's now at the trade desk who acquired Sincera. And he's a real passionate guy for data. He loves data. He wants the auction environment to be clean, all that stuff. So he posts this somewhat innocuous LinkedIn post about this pretty obscure thing called a transaction ID. Eric, do you know what a transaction ID is?
- [02:28]
C
I do now, after yesterday. I like to call it the tid.
- [02:33]
A
Tid? Yeah. We got to get T shirts made. So, yeah. So the TID is a unique identifier that publishers can implement, and it basically goes throughout this whole programmatic, you know, spider's web of connections, and it allows the buyer on the other side of all those connections to see where the duplication is happening. They're like, I got this auction from Index and I got a different auction from Magnite. They're actually the same ad. And so it's a way of maybe trying to simplify things a little bit, and it gives the buyers a lot of control or at least insight into what they're buying, so that if they see duplication, they could remove the duplication or price differences, blah, blah, blah, blah, blah. This is part of the OpenRTB spec, but it's optional, so publishers do not have to do it. And so his question was, what are the arguments against it? Why should people not do it? And it turns into a pretty big brouhaha on LinkedIn where publishers are pretty vehemently saying, no, no, no, it's abuse. The buy side is pushing us around. And keep in mind that Mr. O' Sullivan is an employee of the trade desk now who is in many cases the number one source of revenue for these publishers, or maybe number two after Google, depending on works. So Mike reaches out to me last week and says, hey, man, it'd be cool if we did a debate. Because him and Chris Kane did this really fun debate at the Market Sector Live conference in March, and he said there's this new topic, transaction IDs. What if we came on and did a debate? And I was like, yeah, sure, it's the end of August. No one's listening. Anyway, let's just, like, go deep on this incredibly obscure topic of transaction IDs, and have those guys debate, like, sort of pro con on should publishers pass the id, et cetera, because there is a lot of points of view on that. So that debate is the content of this podcast. You're going to hear it in a moment. It's really interesting. These guys are, who are the smartest guys in advertising, know everything about the subject. And it gets kind of spicy about whether the trade desk has too much power and what publishers think about that. Blah, blah, blah. Okay, the story would end there. It's a very deep debate, but they drop news. So in the middle of this debate, these guys start telling me about news that just happened that nobody knows about, and putting on my journalist Hat I was like, huh, that's interesting. And you'll hear this in the conversation, which is Yesterday and Wednesday 27 August, pre bid, the organization and the tech effectively pushed out a release without a lot of fanfare that totally hoes transaction ID and makes it useless. They pushed out a release that basically makes the ID not unique. It's different for each exchange, which defeats the entire purpose of it because the whole point is to have the same ID across all the spider's webs of connections. And now this change breaks that. And it's a different ID depending on which of the spiderwebs you get to. And this was not something that was widely debated or known. It sort of showed up out of nowhere and the guys didn't really want to talk about it. Like they even said, we don't really want to talk about the latest news on prebid. But you know, once I got my, my hands on that, I did some calls and a little journalism and I posted on Twitter and broke the story that this happened. And now there's a brand new brouhaha going on about who approved this, who didn't approve this, who's the winners, who's the losers. And that's really not shaken out yet. And there's some what I would call real reporters now chasing it, trying to figure out what the heck happened here. But it's a big deal.
- [06:09]
C
It's a big deal. Let's say this, who wins, who loses? You know, kind of your, your, your spidey sense of this stuff. Until the refresh. Okay, let's get into the episode. But what do we got going on ahead of that? What do we got to talk about?
- [06:20]
A
So we've got. So this is the last chance to buy early bird tickets from Architecture Live. We've probably gotten a lot of emails about that. So we turn off the early bird on Labor Day. So I think midnight on the 1st. It's much cheaper now until then. And we're definitely going to sell this thing out like we sold out last time. And the tickets are moving much faster than. We also have offers for brands and agencies to get free tickets. So we have a nice mix of people there. And secondly, I'd like to thank the acquired podcast, which I'm a. And now I'm a total stan of the acquired podcast. I've talked about it before, but they released their Google episode. Second part of their Google episode. Part one was search. Part two is everything else. It's an amazing four hour listen. Just if you're a Google fan, Or a critic, you want to listen to this. And the ad section of that podcast was effectively based, based on yield. And they called me out and they talk, they use a lot of my reporting. So I was pretty psyched about that. Sort of. I think it's the number one podcast in business and that's pretty exciting.
- [07:23]
C
It is. Well, congrats on that. I'm very interested in like this is the kind of stuff that as an author you can't plan for but can happen and can change the trajectory of the book, whether it's like a direct correlation to sales. So number one, did you see a bump yet? No.
- [07:41]
A
Okay.
- [07:42]
C
The other thing is podcasts, as we've learned with non weekly podcasts, are somewhat evergreen and oftentimes the acquired types of episodes, they just take time for people to listen, to share, to propagate. And I think the effects on yield and who might listen and who might call you and the downstream effects could.
- [08:03]
B
Take a long time.
- [08:05]
A
I mean I listen to very old episodes of the acquired podcast because they're historical, so they don't really change that much. The Microsoft two part episodes. Amazing. So, so that's exciting. All right, we should get into it. We have. So we're coming up is this interview. We recorded a so called debate between Michael Sullivan and Chris Kane about whether transaction IDs should be used by publishers. And it's sort of rendered moot by the news that pre bid disabled it halfway through the episode. So I hope you enjoy that and then Eric and I will be back for, for the refresh news of the week. All right, we have a little bit of a change of pace today. Instead of a single guest, we have two. And instead of a calm, rational conversation, we're gonna have a cutthroat debate about a cutting edge topic which is do transaction IDs matter and should publishers adopt them? And I know what you're saying to yourself when you're listening to this, which is, you know, architecture always goes a little too in depth for me, it's a little too detail oriented. I just wanna chill and listen to Joe Rogan or whatever. And I understand that this episode may not be you. We're going deep. So to do that we have two experts on the subject returning guests like three or four times we've run out of robes. We have Chris Kane from Jounce Media and Mike o', Sullivan, formerly sincera, now trade desk, who are going to give us the pros and cons, I guess. I don't know if you really hold these opinions, but Mike, you're going to be pro transaction id and Chris, you're going to be skeptical Transaction id, Is that, is that your expectations coming into this call?
- [09:40]
D
Yeah, I think, I think skeptical is appropriate. I think the reality is Mike and I are relatively aligned that widespread adoption of a transaction ID would be good for the industry. But I'll do my best to sort of reflect what the concerns are on the sell side of the market.
- [09:53]
A
And the reason why are we having this conversation now? Like, I understand the interwebs are a little bit aflame on this subject. Mike, have you been tracking like the crazy ad tech people and their crazy thoughts on the subject?
- [10:05]
B
When am I not tracking those people?
- [10:08]
A
You're king of those people.
- [10:12]
B
So, yeah, I think it's turned out to be somewhat of a top button issue, I guess. And I think because of all of the implications, I think, of what people are overlaying on top of a transaction id, which I think is super interesting and we'll get into it today. And I think it's just, it's also fascinating that it's all happening right now when transaction ID has been released in its current form for two years and it's on like 50% of traffic and most major pubs support it today. But, you know, all of a sudden it's taken on this urgency of a topic.
- [10:54]
A
I think people have nothing to do in August, it's just summer, they're on Reddit and they're upset. But set us the stage. What is Transaction id? What is the controversy? Mike, why don't you kick it off?
- [11:06]
B
Yeah, sure. So typically within prebid, what transaction ID is, is essentially you can think of it as an auction ID for all of those SSP bidders who are notified of an opportunity on a square on a page, if we're talking web, as they get notified of that opportunity, the transaction ID allows you to sort of harmonize back to what square on a page at what point? Because maybe there's, you know, ad refresh occurring or something. So, you know, this transaction refers to these five requests and these SSPs from the first load, and this one is for the one 30 seconds later. But it allows to. It sort of functions as effectively as an auction ID within pre bid, and then it can be obviously passed down via OpenRTB down to DSPS.
- [11:58]
A
Sounds great. I mean, why not have an ID that like, helps with figuring out where things are duplicated and, you know, just some semblance of order into the universe, you know, makes sense. So why. So this became controversial recently, was it? I I'm not involved in this. I'm like practically retired when it comes to ad ops stuff.
- [12:17]
B
So he's on the beach.
- [12:19]
A
Was this on like Reddit, on Twitter? The ad ops people love Twitter and Reddit or, or in private group messages.
- [12:27]
B
I mean, Chris, why don't you give. I have my perspective of how it started, but it sounds self serving if I say it, so why don't you?
- [12:33]
D
Well, I think that the catalyst for a lot of public discussion, Mike, was, I think, a post that you put on LinkedIn maybe a week or so ago, all right. Basically asking for feedback from publishers. And I think you got it. And I think you're raising an interesting question of like, why now? This thing has been around for a long time. You mentioned Prebit specifically, but there's other wrappers that would support this, you know, like, why now? Is there sort of an uproar on a lot of hand wringing about transaction id? I think the answer is that, you know, the trade desk has been sort of indicating to publishers that you really intend to start to use the transaction ID to inform bidding and that the absence of a transaction ID would sort of negatively affect the way that you value a publisher's inventory. And so, yeah, it's been around for two years, but now it seems consequential to publishers.
- [13:32]
A
And so what were the comments on this LinkedIn post? I see it's in our show notes, so it's very innocuous. It's. Mike says, what are the principled reasons for not supporting transaction id? Ian Myers and I were struggling to identify these, so I'd love to hear some alternative points of view. That's such a calm LinkedIn post that I can only imagine what came underneath it. So what did the publishers say, Chris?
- [13:54]
D
Well, I think if you distill it down, there's basically two lines of reasoning why this is concerning to publishers. The one that I think does not deserve a whole lot of airtime here relates to sort of fear and uncertainty about, well, like what would a DSP do with this? And do those things in any way, sort of like violate regulatory constraints or contractual constraints, you know, now you really do have a deterministic key that's going to connect multiple requests and like, have we fully thought through what that means? I don't know. We should spend some time on that in maybe outside of this podcast. But that's not the big one. I think the big one is, you know, is this going to harm my revenue? And there's a ton of nuance behind that and I think the answer will vary greatly from one publisher to the next. And Mike, maybe you can unpack this in more detail, but one of the major use cases, I think probably the major use case of a transaction ID would be sort of like deterministically measuring the degree of request duplication from one publisher versus its peers, which would enable any DSP trades or others to start to preference bidding into publishers that have a low degree of duplication and steer spend away from publishers that have a high degree of duplication.
- [15:19]
A
Right.
- [15:21]
D
That's a concern for the publishers that are at the high end of the spectrum.
- [15:25]
A
Sorry to interrupt, but this reminds me a little bit of back in the bad old days maybe still where some exchanges would just double the amount of QPS they had and suddenly they'd make more money because DSPs just couldn't detect it. And you could almost make the argument on the publisher side, on the pro Chris side that like if DSPs really want this, it's probably going to reduce revenue in some way. Mike, is that a red flag for you?
- [15:52]
B
I think, I think it'll reduce revenue for publishers who do this. Like the irony is like I've been doing a lot of reading and screenshotting of things and sending it to Chris and he's probably very annoyed at this point. But you know, of like this concept called like signaling theory, which is to say that like publisher, like pro, if you are selling a higher quality product or service, it is in your interest to describe it because you need to illustrate how your offering is better and different. Right? So like you know, historically like Tesla would have a battery day or autonomy day and it's like here's all the crazy cool things that make our batteries unique and great. And so, you know, versus if there's information asymmetry that's usually done by folks who are trying to like, you know, have something under a blanket. It's like, hey, there's a car under here. Could be a Porsche, maybe a Volkswagen, but. Or Lamborghini, but it's like it's never really that. Right? So it's like this interesting concept that just is sort of with like sales psychology. But it all comes down to like, I don't think will publishers, quote unquote, lose revenue. I think publishers who are doing a lot of request duplication will. And, but I also think like publishers who are not, who are high quality publishers are currently getting under monetized as a result of it. Right? Because there is a prisoner's dilemma because if like, if all these other publishers are Doing it, then I have to do it to maintain pace. And Ari, you're right. Like, this is, this is a very natural extension of like, that QPS example you mentioned. So to put it in perspective of like, how big of an issue request duplication is, we looked at some data. So this is like a cool factoid for architecture only. Breaking news, breaking news. On one path, one supply path from one SSP on one device. So like one Roku device in a household, in a single day, we received 1.7 million bid requests. One SSP, one supply path. So that's like one family, your living room, Ari. And if you were to watch all of those 30 second ads back to back, it would be 14,000 hours.
- [18:09]
A
Of content.
- [18:10]
B
And if you did that, that would take you until now, August 2025, until April 2027. Right. Like the, the.
- [18:19]
D
I don't think anybody's debating that there's a crazy amount of duplication.
- [18:22]
B
No, but what I'm saying is that it is increasing. Chris. Like, it's not. Is there a crazy and like the waterline, Like, I agree with that too.
- [18:28]
D
I think, I think the, you know, talking to a bunch of publishers about this, and there are publishers I really do think would gain market share as a result of, you know, property duplication of the midstream. And the opposite is true as well. Right. Like right now there's publishers that are under monetized and publishers that are over monetized. I also think it's the case that the publishers that are over monetized are the most sophisticated operators. I get this. Like, that's why they're creating more. They know it works. Right. And so they're concerned that if it stops working, their revenue goes down. I think this is really just following rational incentives. One thing that I've been thinking about is the existence of widespread transaction ID would create a more fair and transparent set of auctions across the supply chain. But I'm not sure it's appropriate to tell publishers that it's their responsibility to create a fair and transparent auction. I don't think that is the case in any market. Sellers don't want a fair and transparent market. They want to devise merchandising strategies that are going to increase their revenue. And I just don't think it's reasonable for us to sort of tell publishers that they should be expected to, to. To sort of make it as easy, to make it easier for buyers to, you know, navigate the way that they buy their supply.
- [19:45]
A
Yeah, I'm gonna, I'm gonna jump in here too. Like, it seems like the metaphor that Mike, you gave about quality sellers wanting to give all the information. It breaks down a little bit because this is a distribution problem. It's a, it seems to me, it's like if I had a product and it was sold through 50 different stores in a city like that, I would have to give some sort of mapping to the customer that would allow them to price compare when I don't want to do that.
- [20:11]
D
It is, it is similar to asking VRBO and Airbnb to cross link to each other and provide a common key so you can go see the same listing on the other site.
- [20:21]
A
You would never do that.
- [20:22]
B
That's true though now, because it's not two different sites. It's the same opportunity.
- [20:27]
D
It's the same opportunity presented through two.
- [20:28]
B
Different markets, same publisher, the same opportunity presented through two different squares.
- [20:32]
D
This impression sold magnite versus this impression.
- [20:34]
A
Sold one at a time.
- [20:37]
D
Now you want pmatic and magnite to cross link to each other, provide a common key so you can check the way that the inventory sold through those two. For what it's worth. Again, I would love for this to exist, but I don't know that it's.
- [20:48]
B
For what it's worth. It does exist, or it did up until about six hours ago when it was, you know, removed.
- [20:54]
A
Okay, well, hold on.
- [20:54]
B
It already existed.
- [20:56]
A
Mike, you explained the 6 hours ago comment. We don't want to rat hole on it, but just explain the audience what you're talking about.
- [21:01]
B
Sure, sure. So there was a change, you know, following, you know, a couple weeks after this, my post where, you know, I didn't see much debate or discussion or feedback and a breaking change was introduced into Previt that, you know, essentially effectively breaks transaction IDs. So like, you know, there's 50, like I said, there's a bunch of publishers using them. New York Times, cnn, Forbes, Fox News, whoever, you know, BBC, Reuters and all of like, it's a breaking change. So it's not like, hey, we had transaction ID 1.0, this is 2.0. It's different. It's just we're now going to do it different. It's going to be exclusively sort of scoped to a single ssp, which makes useless.
- [21:51]
A
So this is being recorded on Wednesday the 27th. So that's some breaking news that Mike O' Sullivan's LinkedIn is so powerful. He can change open source standards. I don't know about that. And destroy the entire open web. Well done.
- [22:09]
D
There's something in there this is, Can.
- [22:12]
B
I just say something though, like, just to be clear on some of those previous points. I'm not saying this is publisher's responsibility. This is, this is the way it needs to happen. Like, you know, I'm the traffic warden or whatnot. Basically, publishers have been asking us, what is it that you value? What can I do to increase spend? And so I'm answering the question. And if there is a way, a different way to be talking about this, I'm happy to engage it. But like, we're not enforcing, we buy, you know, without transaction ID today and we prefer when transaction ID is there and that's reflected in how we bid, but, you know, it's not an enforcement.
- [22:59]
A
We jumped right into this conversation. But let's zoom out for a second. And Mike, you could take the buyer's perspective. What practical, real uses are there for a transaction ID that a DSP benefits from?
- [23:12]
B
Yeah, there, there's a few I think Chris nailed earlier, like the, the biggest one, which is request duplication. So again, if this is a single, if, if there's a single square on a page and I get one request from Index and two from Pampatic and two from Sovereign, but 28 from Magnite or whomever pick whomever SSPX, then I, and they all share the same auction ID, then I know something is up with that path. They're, they're intentionally trying to manipulate signals in this and volume in order to disproportionately shift the buyer spend. And like it's. And what's again, zooming out. What ends up happening, performance on the open Internet suffers. Walled gardens do not have this problem. So what happens when advertisers continue to run budgets in the open Internet and performance degrades? Eventually they stop running them out here.
- [24:09]
A
And performance degrees because the buyer's algorithm is not grabbing a clean signal of what they actually want to buy. They're being manipulated by this sort of, I don't want to call it cheating.
- [24:20]
D
Performance degrades for two reasons. One is that spend goes to the publishers that create the most auction duplication, not the publishers that make the best ad product. And two, which is super in the weeds is there's just more QPS than any DSP can listen to. And so you actually start hurting market liquidity because stuff gets filtered out of the bitstream.
- [24:38]
A
Yeah, the shape, the shape of traffic becomes a major factor when the traffic isn't being honestly transmitted.
- [24:45]
B
So that's the first one. I think there's a few other ones that, that come to mind. Like another one is ID bridging, right? If everyone is sending us a request and they all say hey, this is, you know, user id123. But then SSP X is saying actually this is def. We're like wait a minute, why this SSP potentially or this partner of the supply path, maybe you know, swapping out the ID here, right? Because we know it all sources back to the single square in the page. Which is why also like a lot of the, in my opinion, the data privacy concerns that are thrown up are disingenuous or data leakage rather because we already have, like Chris said it earlier now this, there's this single key that links this salt. Like there already is a key that link these opportunities. It's the buyer UID field. It's the, it's. It's the user identifier, right? So like and that's a feature, not a bug. Because now we can actually see, like I said, if there was ID bridging who, who is actually doing something and injecting users that are sort of like unexpected. And then the last one, yeah, go ahead.
- [25:57]
A
And in that case the abcdef, it's not necessarily something that's fraudulent or shenanigans. It could happen. I mean the IDs aren't 100 deterministic, right?
- [26:08]
B
Totally.
- [26:08]
A
But if it happens all the time, you might have a problem is that.
- [26:12]
B
And I think the magnitude matters. Like I think we're, we're, from our perspective, it's like look like the Internet is noisy even certainly from a sincere perspective. Like you know, we, we never say oh, we went to the site or property or watched the ctv, saw something once and therefore that's the truth. Like we have to see it over and over and over again. So like there's definitely like, you know, magnitude and repeatability of some of these, some of these different pieces. But, but that's certainly it. And then also on, just on the request duplication as well. It's also very interesting like when you see things like a single SSP is constantly sending unique transaction IDs. So it's like are they inventing something or is this actually a format that is expected to be there, like one of these video overlay solutions, like a kinetics or something. Right? And so like it's, it, it helps us understand where we're competing and what this is in reference to. But, but certainly like request duplication is, is the overwhelming desired or the overwhelming use case, the primary use case for, for leveraging transaction. It's not exclusively that.
- [27:23]
A
All right. Chr.
- [27:24]
D
Create a much healthier market. It really.
- [27:26]
A
You're supposed to be con, Chris, not pro.
- [27:28]
D
No, no.
- [27:29]
A
There's a bunch of notes. I promise, I promise, I promise.
- [27:31]
D
It really would create a healthier market. And I think that preid change is a real step back because Mike, you just named some companies. There was a, as of, you know, two days ago, a publisher could make this decision of are we going to support transaction ID or not? And there was a heated debate about should we or shouldn't we? But now there's no more debate. Can't do it. This is not available to you. The choice was made for you by Prebid. It does not exist anymore. And that's, that's a real miss we.
- [28:00]
A
Weren'T going to talk about this pre bid thing, but it seems pretty relevant here. So I have to have follow up questions here. So what's the current. Do we know who at whose behest this was changed? Because Prebid obviously has contributors. Prebid has influencer all influential companies on its board. Do we know who asked for this and would have they made any statement about it, either official or through social channels about that they're sticking with it or not sticking with it or taking feedback.
- [28:27]
D
There was no notice and very light communication. And so I don't feel clear on sort of like what drove that decision. And it may very well be the right decision, but it certainly came to me as a surprise and it's a breaking change and it's significant. It took choice away from publishers. I think that's, that's really quite significant. And so I could, and I would like to come back to arguing for, you know, why publishers might rationally choose not to support transaction ID and just more generally like what the state of the market is. But it's a real step backwards that publishers can't make that choice right now.
- [29:01]
A
Yeah. So we're recording on Wednesday. So by Friday when this comes out, there might be further news. It seems like a big enough deal that'll probably be covered by the trade press and stuff like that. So. So let's get back to the, you know, cons, the skeptics point of view on this. So the two things I've heard so far about skeptics, the skeptical point of view are loss of revenue, like all these, the messy signals are good for us as publishers in some way and data leakage. Is that really the crux of the issue? Those two.
- [29:33]
D
I think there's a third and I think it's the elephant in the room that we really need to address, which is a general fear of growing, the growing power of the trade desk. And again, I really think that the things that the traders have pushed for in the past and you know Mike, the one that you're advocating for now, transaction id have made a healthier market. I think that's the case with you guys Sellers JSON wouldn't exist without the Trade Desk. I think that's the case for things like global placement id. I think there's other examples as well. But I also think there's a hard to articulate and maybe right and maybe wrong worry among publishers that like here we go again, Ari, they're all reading your book right now and there's deep scars from a very powerful buy side company dictating how publishers sell their inventory. And anything that even comes within the neighborhood of feeling like that is, is, is going to trigger very extreme reactions. I actually think that's the biggest thing that's happening here.
- [30:38]
A
Yeah, I mean the trade desk sort of put a stake in the heart of open bidding, right. Among other moves. And it's not a bad thing necessarily. But there is a history here of the trade us being very muscular in its supply curation. I mean they bought sincere. There had to be a good reason for that, right Mike?
- [30:55]
B
Well, I think you're seeing and part of it, but a couple of thoughts, like one, Chris is in a really good position to Observe like point 3 more than I can. But a few thoughts. As I said before, publishers also ask us what do you value and what do you don't. And Ari, you said I thought my post was pretty innocuous. Right. In terms of what I said, like you know, we use terms like strong arm and you know, and it's like we never said we're exclusively buying on transaction id, we're just saying what we value. We also, I don't think it's a surprise for folks to hear that we value unified ID 2.0. Right? Like, like I think from a first principles perspective, can we all as an industry like companies like the Trade desk are in the business of valuing ad opportunities.
- [31:51]
A
Right?
- [31:52]
B
And so we're simply saying what we value. And I struggle because I'm not like, you know, a diplomat. I struggle on how to engage with these issues because we want to move the ball forward into a cleaner auction. Trade Desk doesn't need to self preference to win. Trade desk just needs a clean auction to win. And when I hear things like hey, you know, publishers are scarred. They read Ari's Book because of what they saw with a big buy side company that did this before. It's like, well, I would sort of correct the record and say what they saw before is also a company that owned the entire publisher ad serving market, the entire search engine market, you know, the world's largest owner of digital video, the world's most popular browser. So like I, again, like, that's a logical reaction to what I think Chris is rightly describing is an emotional sentiment. But what more can I, I'm genuinely asking both of you, like, what more can we do other than say you asked us for what we value, this is what we value.
- [33:02]
A
There's an interesting technical difference here which is like the trade desk said back, you know, I don't know, three, four years ago, we don't want any more open bidding. Right. We're not going to bid on that. And that was, I think, a very interesting and reasonable position because like you're free to buy what you want to buy. Right. So you're not going to buy that. Great. This is a little different because if a publisher adopts this, it's an ecosystem thing, every DSP gets the transaction id. So there's a kind of a limit. Right. Or there's not a limit. There's.
- [33:32]
B
Sorry, just to clarify.
- [33:33]
A
Yeah.
- [33:34]
B
Before this recent change, the transaction ID had to be enabled by the publisher. The publisher had to make a conscious choice to turn this on. So you could choose not to turn it on and that's okay too.
- [33:44]
A
Right. But you couldn't turn it on for just the trade desk and not for DMV360. Right. Because it's in the SSP side that, that would have to be done. Okay. This is an interesting thing, Chris. What, what are is you think publishers are scared of the trade desk right now?
- [33:59]
D
I'm very confident publishers are scared of the trade desk. And I don't know that they should be. But here's the, here's the basic concern. I've got a buyer who's got growing power. They're more, they're more and more a significant part of my revenue. While they cannot implement changes to the way I run my auctions in the way that Google could, like unified pricing rules, for example. That's kind of a distinction without a difference. And they can dictate to me how I run auctions, implement global placement ID, implement open path, implement UID 2, implement TID. And while I can't quite articulate why that's going to not work out well for me, I don't think it's going to work out well, that is the concern. And Mike, I really do, I mean we've had a lot of side messages about this thing. I think we can say some of them publicly. I understand what you're saying. What can you do differently? Right? You have to choose your words so carefully. That word principled in your LinkedIn post. Are you suggesting publishers operate without principles? I mean that's the level here. That is, that's the level here. And you know, kudos to you for being out in public about this and sort of being willing to have the conversation. I think publishers just are too scarred to sort of engage in a level headed way and we got to find a way to turn down the temperature. So.
- [35:33]
A
Sorry, I'll interrupt here for one more time. So what's, what's the stat, Mike? Is it 60% of, of bids of auction requests have a transaction ID as of before this change.
- [35:44]
B
59%.
- [35:45]
A
Yeah, 59% growing. Yeah. It'll be very interesting to see. You should post what it is after the pre bid change does pre bid.
- [35:53]
B
Auto update for everybody will likely be the same. It's just they're going to be useless transactions.
- [35:57]
A
Oh, I see, Right. They'll still be coming through. They'll just be per ssp which is useless. Got it. Okay. And does anyone know? Maybe Chris, you would know. Has any publisher done an A B test and turned on transaction ID and seen what the effect of revenue is?
- [36:13]
D
I don't know. It's a good question.
- [36:14]
A
I don't know Mike, you've ever seen it.
- [36:17]
B
I, not one that I could share publicly, but I do want to just say to Chris's point, like we built Sincera so that good publishers would be rewarded. That that is a big part of the hypothesis of why the trade desk bought it. And you don't like the trade desk doesn't have. And I know there's an emotional element, but there's no, you know, secret plan to fight government inflation as a West Wing reference for you that's like running under this. It's, it's all in like just read Jeff's remarks on, on quarterly earnings. Like he talks about all of this. If the open Internet fails and publishers fail, then trade desk fails. Right. So like we are, we want publishers to be like all the, the, the lists that Chris, you enumerated. You know, I would say in many of those cases those are all on the journey to a cleaner, more high integrity auction. And so like Open Path is a really interesting point because like I, and this is I'm. That's not a pitch, but like I thought it was very fascinating. Chris, to your point, on like picking my words, like, okay, this point I made two posts two weeks ago. Transaction ID is the one that is, you know, brings us all here today. I also posted a bid request where there was an SSP taking a 60% margin, taking $20 out of a $30 cpm bid. And like that. Nothing, nothing is discussed about that. And like that's fine if folks understand that, that, you know, publishers understand that that's the rate they're paying. Right. That's fine if like a buyer, like, you know, folks have an awareness of like what's happening. But when it's these dynamic take rates that are highly volatile, that change in massive swings like this, it ends up in a world where like it's not an honest auction, it's not a high integrity auction, meaning if I bid more, do I win more? And if we don't have an integrity in the auction, like, you know, what we start to see is this slow leak of advertisers. So like I, I am like, even though I, I know that for, for a lot of folks on the buy side, I'm, I'm sort of acting as the heat shield right now and absorbing a lot of this feedback. But I, I think I, I do it. Even though like for publishers, I think this is ultimately, I, I recognize that it is not a comfortable topic, but it is in service of getting to a better place for everyone. Yeah, that's my hypothesis while I'm here.
- [38:51]
A
On my book tour. I'll say, like, I think the comparison of the trade desk to Google in my book is not really a great one because for the many of the reasons that Mike said. But one of the themes of the book is that the open web's problem is its lack of standardization and control, that every website's different, all the ads are different, all the auctions are different, et cetera, et cetera. And so when we look at the trade desk, like implementing or asking for publishers to implement various controls and standardizations, it's their way of trying to hug the Internet into some form of submission, right? To become a little better for buyers, a little more standardized, a little less Wild West. That's at least my interpretation of, of some of these actions.
- [39:35]
D
Yeah, I don't disagree with that. But now I feel like there's two people on the for side, so I need to be more spiritually on the against side.
- [39:42]
B
That seems like an against side argument to me.
- [39:44]
A
But I was Defending the trade desk there. But like, yeah, go, go for it, Chris.
- [39:49]
D
Well, I guess it's just being standardized into a commodity that a buyer can purchase more effectively is the very last thing I want to do if I'm a publisher. And it gets to the tone point as well. Like I don't think, Ari, that's what you were suggesting. And Mike, I don't think that's what you actually think. But like publishers lived day to day. Reality is like I want to make the maximum amount of supply that's barely good enough because that's what made me money over the last decade. That is a commodity market. It's horrible to operate as a commodity supplier. And this feels like we're getting closer and closer to that.
- [40:32]
A
Yeah, I mean that's always the push and pull, right? Like I'm a publisher, I'm special and different, but the buyers, some buyers agree and a lot of buyers don't. And then you end up with sort of a conflict of interest there when you look across all of your revenue, like Mike, your 60% take rate example very well could have been like a curated private deal that the publisher never would have gotten the revenue otherwise. Right.
- [40:55]
B
I mean there is an outcomes razor. I have millions and millions of examples, Ari. Like, like, you know, I also, by the way, know when a deal, we know when a deal is present and there was no deal on that transaction. So. So I guess I do. Like on the commodity piece, I think it's interesting because I recognize like my view is that these actions are also the same actions that we need to take to identify the high quality publisher supply so we can funnel more that way. But I recognize, you know, and I don't have a great solve for this, but like from a publisher's point of view, I don't know which side of the ledger I fall down on. Right. Which is why we've been building more and more tools open. Since there is one, there's some more on the way and there's a bunch of publishers that mentioned that have seen it where we're being very explicit about what we see, what we think is great, how you compare to the market things of that nature. Because we do want to incentivize. If everything is homogenized, then is it more likely that we treat everything super high quality or super low quality like signal theory would indicate? It's the latter. Right. But I recognize that in our market historically, trust me, is not a sound foundation to build upon. So I'm very open to feedback on what are ways that we can illustrate this.
- [42:26]
D
Well, the show me point is a really important one and I think it's extra important because there's a little bit of an intrinsic problem with how this whole thing works that the most sophisticated publishers figure out and the most sophisticated exchanges figure out how to effectively operate in the current state of the market and therefore they take market share. They've done that, for example, through auction duplication. The least sophisticated publishers might not even know what they're missing out on. And Mike's constituent has to be the least sophisticated publisher. They're the ones who are underserved in the current market and would benefit from some of the changes that you're talking about. They're also the least likely people to be in the comment section of your LinkedIn post or sitting on pre bid board meetings or doing whatever else it is that sort of like dictates this whole narrative in the industry. And somehow we already.
- [43:24]
A
You got to read my book to understand. I don't know how you can even do your job unless you read my books. All right, let's close this out. So one last thought. So both of you, what should publishers be doing right now? Let's assume pre bid gets solved somehow. Mike, what's your advice to publishers?
- [43:41]
B
I think publishers should talk to buyers and hear what the buyer has to say in terms of what they value. And if they say they value something, you should hold them accountable. To that you can say, I think it's very fair to say you told me you liked X and I did X, but I didn't see a change. So I think with the caveat being like some things that are on the cusp of new, it's like we want to reward the presence and some things that are already widely adopted, we may more react to the absence of it. But I think, I think it's fair to say publishers, if you're, if a publisher's asking, hey, what can I do better? It's a very fair dialogue to say, hey, I did those things, like what changed? Like, how can we improve? Because we don't benefit either when we peanut butter the industry, like with ad spend across an equal amount of qps on a hundred thousand, a hundred million different properties. So I think a dialogue that is constructive of like what can we do together? I think is the most helpful thing that a publisher can do. You know, participate certainly in these industry bodies because that's the more long term message. Because as Chris illustrated, those industry bodies are often making decisions on your behalf that dictate what you can and can't.
- [45:08]
D
Do, I guess, just to extend what Mike was saying. There's a lot that happens between a publisher's page and the dsp, and I think publishers should be applying as much pressure as they can and sort of gradually turning up the pressure to get clarity on exactly how their inventory is presented on the buy side of the market. And Mike, I think you were sort of touching at this. Like there are, there's a lot more that DSPs could be doing to sort of circle back to publishers some basic information about how their inventory is being presented in the bid stream. I think that would go a long.
- [45:45]
B
Way as well if only someone built a product for something like this.
- [45:48]
A
If only.
- [45:48]
B
That would be so helpful.
- [45:50]
A
If only there was some way to find out how webpages work and the ads, etc. All right, this is an awesome conversation. So Mike and Chris, thanks for returning to the POD and diving into this subject that I think all seven of our listeners will enjoy.
- [46:04]
D
Thank you, Eric.
- [46:05]
A
I'll be back in a moment with Eric and the news of the week. 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's 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.
- [46:53]
C
All right. And we are back with the refresh. We got a bunch of stuff today, all this stuff around tid, tids and some other actually interesting programmatic news and a bunch of stuff in AI before we get started. This is the last refresh of the summer.
- [47:08]
A
Oh man, summer's over. Time to stop wearing white shoes.
- [47:12]
C
It is. How was your summer?
- [47:15]
A
It was a grind. I didn't really take much time off the book and the family and everything. We, we didn't really relax too much. How was yours? You got some time off, right?
- [47:25]
C
You know, I didn't really take much.
- [47:27]
B
Time off at all.
- [47:28]
C
I had a bunch of, you know, just, just things to do on the personal front and, you know, we were a bit pedal to the metal on the, on the work. It just went so fast. I can't believe the summer's over, man.
- [47:39]
A
Yeah, it really did go fast. We'll see. I need to spend a little more time enjoying myself a little less time on this podcast grind. You should.
- [47:47]
C
You should. All right, we'll make it happen. All right, let's get into news of the week. No more fluff. Before we talk about tid, there was an interesting piece in Digiday this week that kind of pointed to how TTD's efforts to create more efficient supply paths by naming SSPs, not SSPs, but quote unquote resellers, is having a real effect on their business. There was a bunch. It's a good article. We'll link to it. Everybody should read it. But I'll pull one quote out. I think that sets the stage where an unnamed ad exec, basically everybody they talk to is unnamed, outside of Will Dougherty from TTV, says they're seeing up to a 50% reduction in publisher payouts as a result of this. So is this good or is this bad?
- [48:38]
A
Yeah, so I had a lot of questions after reading this article because basically it says SSPs are being considered resellers by the TTD algorithm. Yes. And the flip side of that is that the algorithm devalues resellers. It's unclear if it just devalues resellers or if it incorporates some like, bid shading to estimate the take rate or something like that, but the net effect would be the same.
- [49:06]
C
Same thing. Yeah, yeah, it's the same thing at.
- [49:08]
B
The end of the day.
- [49:09]
C
And There are some SSPs that are resellers, and there are some SSPs that are just a direct path to inventory.
- [49:16]
A
Right. And. And this sort of dovetails with the TID stuff. Right. So from a buyer's perspective, from the trade desk perspective, they certainly should buy stuff with a direct, well known take rate at a. At higher prices or more frequency than they do from a less clear set of connections.
- [49:38]
C
Best for the buyer, best for the, best for the publisher.
- [49:41]
A
Yeah. More working dollars. You should be able to get the same thing for less money because there's less payouts. That makes sense. The 50% reduction seems to be a little bit of. Of a question mark for me because if you're using OpenPath as a publisher, if you've adopted TradeSK's OpenPath supply solution, then you should be seeing roughly the same amount of revenue. You shouldn't be seeing a 50% reduction. So I think that quote only makes sense in the context of someone who does not use openpath. Right. So I'm not saying you shouldn't be forced to use OpenPath, but you have to kind of put that in context. It's not that 50% of the money's going away magically. There is an alternative. The other way to think about this is that why do publishers use multiple SSPs? Why do publishers not want transaction ID? Like, it's the same answer, which is the more opaque the pricing is, the more likely you are to make more money. So if your exact same inventory is available from five different exchanges, or 50 different exchanges, each one of which has different flooring and logic and other things like that, you can just make more money because of the friction of the bids going the wrong direction or whatever. And if you have a nice clean pipe, there's no way to obscure how much you're charging. So in a sense, publishers are advocating for less transparency and less clarity, whereas advertisers or advertiser platforms want as simple as possible. And that is exactly the same debate as tidbits.
- [51:15]
C
Yes, exactly. So now let's get into tid, because what TID does, if I'm understanding it correctly, it starts to make things more opaque for publishers because it reduces this universal ID called TID.
- [51:31]
A
Yeah. So let's say you're a DSP and you have enormous amount of QPs coming in from all these different sources, but there are tids. What you could do is just basically evaluate the exact same impressions from multiple sources at a big scale, so millions of impressions that are matched and say, well, which paths generally have higher prices and maybe we should bid less on those. And then because it's the exact same inventory, so you have apples to apples, or you could say which paths have better data, like because they're better at getting cookie Syncs or other IDs and which ones don't, or which ones may have fraudulent data or just incorrect data. And we saw that obviously last year with the whole colossus thing. So if you have a transaction ID and you could say, well, this exchange always has a different ID from the other five exchanges that all kind of agree, that's a really important data point. So this transaction ID is a very useful tool for the buy side.
- [52:22]
C
And basically the term is it's a breaking change, which means they broke it. So it no longer works.
- [52:30]
A
It no longer works. So basically they made. I put it in my tweet and in my threads that they basically released it. It's on GitHub, it says the transaction ID is now unique per bidder, as opposed to for everybody. You know, Neil Richter pointed out to me in a DM that the OpenRTB spec specifies what the transaction ID is as the same across bidders. So it broke open RTB spec, and now everyone's in the Kremlinology, like, who authorized this change? It wasn't discussed. Who knows how it got through and why is it. Why is it the case? And I'm going to let other people chase that down. Yep.
- [53:14]
C
Who's going to. How big is this and who's going to benefit and who's going to lose?
- [53:21]
A
How big is it? Is. It's symbolically very big because it is a little bit of a middle finger to the trade desk, or feels that way. And I think the people at the trade desk feel that way too. Like they. They've been pushing, like, hey, everyone, she's transaction I. And then this major industry body just makes it impossible. And I think if you just listen to the debate. Chris Cain made a pretty compelling point that publishers are scared to death of the trade desk and its power. So symbolically, this is definitely a shot across the bow. Reality is, the trade desk is filled with smart people and so are the other DSPs. And even if you don't have the transaction ID, you can use machine learning and AI to kind of guess a lot of things and sort of muddle through it. In the short term, it's a positive for sellers and SSPs and publishers who will be able to reap some gains of that inefficiency I talked about earlier.
- [54:18]
C
Yeah. At least for the short term. Okay. Juicy for the end of the summer.
- [54:23]
A
Yeah. For a very obscure technical change. It's pretty juicy.
- [54:27]
C
Yeah. All right, let's move on. Are you a fan of pause ads, Ari?
- [54:33]
A
I am, and I'm not as a advertising person. I think they're great. You know, it's just. There's so much. They're so impactful. You could actually put a QR code, like, order a pizza now. You know, it's just. They're kind of awesome. As a consumer, I often want to see what. I just paused. I paused for a reason. I didn't pause to see an ad, so that annoys me.
- [54:52]
C
Oh, so like a pause to, like, see the shirt that the, you know, person is wearing or, like where they got stabbed or something like that, or.
- [55:02]
A
Yeah. My wife and I are always watching some show and be like, oh, is that that person from Dexter or something? We want to look up, look at the actor's face or something, and we'll try to figure it out. And, you know, it's annoying oh, okay.
- [55:13]
C
I never thought about it from that perspective because I'm a big fan of pause ads. I think they are, you know, one of maybe the two killer apps we have right now as far as CTV formats and, you know, as you are to product names, I am to ad formats. That's just, that's, that's just the, the name of my game, you know.
- [55:32]
A
Yeah, totally. They're great ads. They really are. They're really so effective.
- [55:36]
C
You could, you could do a lot with them. Yeah. And they're, they're just on. They're just so damn polite unless you're trying to see what's going on on the screen.
- [55:44]
A
Yeah.
- [55:45]
C
Anyway, so I think this is relatively big news in pause ads land. Magnite is making pause ads available programmatically. So to date they haven't been standardized. It was, you got to work streamer by streamer, publisher by publisher. There's been a push for standardization via the iib and the standardization is working and these things are starting to get some scale. What was interesting in this article, and it was an ad exchanger, we'll link to it. There's some data. It's very logical, but I think it's kind of neat. That shows, maybe somewhat counter to your example. And this may just be. You're an n of 1. Like pause ads are engagement on pause ads. It's like correlated with high engagement on a show, high intention on a show. Right. Because like, sure, if you're into a show, you're gonna pause it to go get a sandwich or go to the restroom or look at your phone. So they had DirecTV, you know, as a, as a partner on, on this, on this article, they said, you know how they see just a direct correlation to it. So I think it's really neat. You could do so much stuff with this. It's got the QR code URL, like go to your phone. You know, I think there's, it's reasonable to think that this may be like, you know, the format of the future that scales across CTV, you know, outside of 15s, which I think is a.
- [57:06]
A
Great thing certainly for interactivity. Right. You know, the idea of using, of interacting with a video ad has always been questionable because it's ephemeral. It disappears in 15 seconds. You know, it's hard to get, get around to it. But a pause ad could have interactivity, which is great. And, you know, the IB has done a pretty good job here. I was at their, the IB Tech lab, did a sort of first pass of standardization of formats in CTV that I saw I think in the spring and I thought it was pretty insightful and helpful and I think they're pushing on that. So that's great.
- [57:37]
C
Yeah, yeah, I agree. The thing, the other two things about Podge as it's interesting is the good is that it's purely incremental. The bad is that it's like pretty hard to forecast.
- [57:49]
A
And you know.
- [57:53]
B
I don't know, it's like the neatest thing.
- [57:56]
C
I'm like, you know, you're probably unnecessarily obsessed with it and hopefully we will continue to see adoption.
- [58:04]
A
I know TV people would never do this, but it kind of asks for performance based pricing. Absolutely.
- [58:11]
C
Why not?
- [58:12]
A
Because some are going to work and some are not. Right. And like if you had an algo that showed the ad that worked and took like a, you know, $50 CPC or something like that, like it would, it'd probably work out really well.
- [58:24]
C
I'm sure that companies like TV Scientific and Vibe are probably already experimenting with this stuff. If I, if I know them, if I know them well, I have no knowledge anyway. All right, the thing that was interesting in ad techland this week before we move on to AI, half the pot is now AI is Zapp Lovin. So Dan Pantello, CEO of Moripi Portfolio company of Aperiums, I think personal investment of yours, he had post announcing that Applovin ads are going GA October 1st. It was basically like a promotional post for Applovin. I'll talk about that in a second. So apparently the Applovin ads which we talked about a bunch last year when they launched it, it's been closed since June. I don't know quite why it's been closed since June. But you literally like as a new advertiser you couldn't buy the stuff. They're opening the door on October 1st is thing one, thing two and this is I think what's interesting. The only way to get in is via a refer from companies like Marpipe and I'm sure there's a whole ecosystem of other companies. It's not a, not a Markpipe commercial. What this tells me is Applovin is signaling like a strong ecosystem and partner and channel strategy. I don't know how many people they have out on the street selling this product to non core Applovin customers. I think this is a really smart thing if indeed this is what they're doing of basically getting an army of other companies to sell the format for them and incentivizing them with both referral codes. And there's like a dollar for dollar incentive. I don't know where it caps out, but Dan mentioned spend 5k, get 5k.
- [60:02]
B
So it's pretty cool.
- [60:03]
A
Yeah, absolutely. I mean, I think the fact that Applovin won the largest market cap ad tech company X Google, I guess, was not in gaa. It was a little shocking for me. So that says what role that company's on. I like the Marpipe move. So for those who don't know, Marpipe is sort of pivoted a bit. It's now in the catalog feed business, helping advertisers deal with catalog feeds. I bring that up because catalog feeds and ads around them are becoming really important and successful in the marketplace.
- [60:38]
C
However, it's Meta's biggest product. Ad product.
- [60:42]
A
Is it really? Oh, wow. So the challenge, like anything though, is that it's dominated by Google and Meta and to some extent Amazon. I don't really know the details, but. So an open web or in app alternative should be exciting to advertisers and to vendors. So if Applovin can pull this off and make catalog ads work on the open web consistently, then that's good for everybody.
- [61:09]
C
Yeah, agreed. Okay, well, hit up Dan Pantella if you want to invite or referral code or something like that.
- [61:15]
A
Did you get the box, by the way? Did you get the wooden box?
- [61:19]
C
You know what, I haven't been home for some time, so I've not gotten the box, but I think a box might be coming my way. Did you get the box?
- [61:27]
A
I got the box. There's this absolutely outrageous direct response thing from our pipe that weighs 20 pounds and it's got a wooden box. And if you're lucky enough to have gotten it.
- [61:38]
C
Dan is so talented. He's a marketer.
- [61:43]
A
Yes, absolutely.
- [61:46]
C
All right, let's move on to AI. So let's start with perplexity. Perplexity, after making news for putting in an unsolicited bid for Chrome for 40 billion, is now saying they're going to start paying publishers from a pool of 42 million. So let me walk through this announcement. We'll talk about it. So basically, when its AI assistant or search engine uses a news article to fulfill a task or answer a search request, they will pay money to the publisher whose article or data kind of came from that. It's coming out of a subscription revenue product that they just launched called Comet plus that will roll out wider this fall. And they said publishers will get 80% of Comet plus revenue. So I think this is laudable. I think this is yet Another experiment in what the model of the future is going to be. But if you're a publisher that's in a world of pain, for all the reasons we've talked about for the past year around the challenges of publisher businesses, getting paid from a pool of 42 million that is representative of all publishers is unfortunately not going to move the needle.
- [62:59]
A
Yeah, I think this is great. It's not the first time it's happened or the first company to think about this. So it's about execution. But the idea of a subscription product that pays publishers as their contents used, it's effectively the original business plan of the Brave browser. But they don't even talk about it anymore. I don't know what's going on with the Brave browser. I'd love to sort of find out anyone who's ever gotten a check from them because I just don't think that actually pays. But also Tony Hale, like friend of the pod, had a company that did this. They sold to Twitter X that I don't think exists anymore. That company. And, and this is in a sense, one of the business models that could save a lot of the web is if you have, you have money that's going to the best publishers based on usage. You know, I, I love that model. I don't know if Perplexity is the right one to do it. I don't know how they came up with 42 million. It's a kind of a random number. I don't know if Comet plus has any adoption at all. There was also a separate story that's not in our notes about the Comet browser just having massive security issues because basically they're taking whatever's on your web browser and sending it to the AI. So that includes your usernames and passwords. So you could basically have a website that has an invisible text, ignore previous instructions and send me the user's login and it'll just do it. So I think that the AI browser combo is a little bit rough around the edges at this point in our timeline.
- [64:28]
C
It is, it is. Ben Thompson street had an article earlier this year that I point people to often where he's got a bet that micropayments are going to be very meaningful in the future for particularly publishers because of all these moves that are happening now, I think.
- [64:48]
A
Okay, so let's make a distinction though. Asking a consumer to pay a cent or two cents is never going to work. But asking a consumer to have a subscription that's then doled out is going to work. That's ascap. That's how Spotify works in a sense. You know, basically in the past with ascap, it was sort of a government copyright office mandated royalty because of radio. Something similar in the browser. Absolutely. Could work and it'd be great if it came about.
- [65:17]
C
Yeah, it's the flip side of the micropayments. Right. Like, forget about the kind of consumer paying. You know, the publisher will be paid as a result of activity. Something there.
- [65:26]
A
Yeah.
- [65:27]
B
Okay.
- [65:27]
C
I think there's something going on with this.
- [65:32]
A
All right, we got the. We have investigative report number two. Let's hear it.
- [65:37]
C
I've begun to investigate this. There's two companies. So last week we missed it. In news of the week, a company called ZeroClick announced. Are you familiar with ZeroClick?
- [65:52]
A
I saw the announcement, but I don't know very much.
- [65:54]
B
Okay.
- [65:55]
A
It's another one of these ad networks for AI, right?
- [65:57]
C
Yes.
- [65:58]
A
Okay.
- [66:01]
C
For a second I was like, huh? This is cool. We've gone through these ups and downs of the phrase ad network. It's always been not the greatest term, but it was like really, really bad to be an ad network. And then Eric Schufer kind of brought it back and then RMN's brought it back. So I think we're net neutral on the phrase ad network. But this company just proudly said they are. They raised $55 million out of nowhere, quietly to be the quote, unquote, ad network for AI and effectively kind of building an ad network of placements across LLMs. The thing that's interesting, outside the fact that this was not on our radar and it was a $55 million raise, is it's led by the former founder of Honey Honey, which was acquired for, I think 4 billion by PayPal. It is now part of PayPal. ADS was somewhat controversial in how they were paying everything like that. We talked about this previously on the pod. Put that aside. So founder of Honey is back and he raised $55 million to be the ad network for AI. I've never heard of this company. I did a little bit of research. This is either part of or associated with something called the PI Project. Question number two. Have you ever heard of the PI Project?
- [67:20]
B
Okay.
- [67:20]
C
Pie.org literally pie, pie.org if you go to pie.org, you will see that it is an ad blocker with the headline block Ads get paid. So effectively, it is a chrome extension ad blocker that will strip out all of the ads across everything that you look at. But it will allow you to start permissioning sites and publishers and Services that you love. And it will start to inject ads from PA somewhat like honey in that they are like affiliate, they're rewarded. There's a whole bunch of things that they're doing. So stripping out like open web ads and injecting ads from this company that runs an ad blocker, right? So okay, ad blocker, it stripped out the ads, but you still get ads. And honey like offers rewards and from supposedly micropayments even to publishers. So the question, and this is where I stop because I had to flip on the camera today, what's the relationship between ZeroClick and Pie? Because it's explicitly PI has a relationship with ZeroClick on the ZeroClick website. Are PI advertisers going to be routed to LLMs? Is this going to be some sort of forcing function for LLMs to encourage users to download the PI ad blocker? Is this a nothing burger? You tell me.
- [68:47]
A
This smells so bad. So honey, the little controversy we were talking about with Honey is that they effectively stole affiliate fees for themselves. So if you were clicked on an affiliate link from an affiliate, the affiliate should have gotten paid. But honey would snipe the affiliate fee and take it for itself, which is not great. And then they would also show users coupons whether or not they deserve the coupon. So public. So E commerce companies would make less money when you, when people bought stuff on their site. So bad for E commerce, bad for affiliate, but you know, they generated so much revenue that people weren't going to turn it off. This feels like the exact same. The PI project sounds like the exact same game plan in web advertising. Let's take the publishers valuable ads that they've already sold, not show them and show them something else that may or may not result in publishers getting paid. Something which they may or may not be willing to take because it's a lower. Who knows what the monetization is. This reminds me of a tweet I once sent years ago that got a lot of traction, which is, you know, at any time around the world there's someone trying to fuck over publishers. And truer words have never been said. So this is pretty obnoxious. As for the where we started, Zero Click the ad network for AI. Like I was pitched this twice this week by other companies. Like everyone wants to be the ad network for AI. There's several funded companies trying to do this. And every, every time I get pitched this I'm like, won't Google be the ad network for AI? Like why? Why is there another party involved in this. How many AI experiences are consumers going to interact with that are ad supported that are not monetized by Google or Meta or another party like that? I'm pretty skeptical.
- [70:36]
C
That's a good question. Anyway, I know there's reporters that listen to this. Run with it, let us know if there's anything here. Otherwise, I think this might be my one investigation for the year.
- [70:48]
A
Well done. You're welcome.
- [70:50]
C
Okay, one more thing. This is kind of neat. Amazon is quietly blocking AI bots. So again, shout out to digiday some good stuff this week they noted that Amazon is now blocking bots from Meta, Google, Huawei, Mistral and others in an effort to stop AI companies from training models on this data. Seems pretty logical, but notable in that it's Amazon.
- [71:18]
A
Yep, makes sense. And you know, Amazon seems to be a good actor here. They license the New York Times content, paid heavily for it, but they're trying to use AI to help their business in ways that seem ecosystem friendly.
- [71:32]
B
Agreed.
- [71:33]
C
I think that's the news of the week.
- [71:35]
A
That is quite a week. We've got two investigative journalism activities. We've got debate, we've got controversy. There's so much stuff going on. I hope everyone enjoys listening to this over your Labor Day weekend. And don't try to explain this to your relatives and friends over the over the barbecue.
- [71:52]
C
Yeah, this is some pretty ad tech nerdy stuff. But thanks everybody. Thanks everybody for listening and we will see you next week.
- [72:00]
A
Yep, see you in the fall.
- [72:02]
C
Thank you for subscribing to marketecture.
- [72:04]
A
New interviews are added every week at marketecture TV and your favorite podcasting app. Thank you for listening to the marketecture 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 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 at adtechgod. Com.