5 Buy & Sell Side Myths, BUSTED.

Yo, the crowd is back. The lights are dim. Front rows buzzing. It's about to

dim. Front rows buzzing. It's about to begin. Ad tech ad talk to taking the

begin. Ad tech ad talk to taking the stage. Adam and Gareth on a knowledge

stage. Adam and Gareth on a knowledge rampage. Tech ad talk. It's a live Q&A.

rampage. Tech ad talk. It's a live Q&A.

Buy side sell side clearing the way. CEO

spitting truth. No delay. Chalice and

camera owning the day.

Welcome to AdTech Ad Talk. Last show of 2025.

What a year it's been. Uh if we're going to wrap up the year with some flaming hot takes on the worst parts of the buy

side sell side, we got Gareth Glazer with uh special special special takes from sellside as always. I'm struggling

today. The weather's terrible in New York. I'm very tired for the year. But

York. I'm very tired for the year. But

let us thank our sponsors. Pontiac

Intelligence, the transparent, lowcost, flexible DSP alternative. Uh if you don't know about sellside decisioning

with Pontiac DSP as the buy side um tech, then you are missing out. Uh the

early tests on this in a little place called the European Union are knocking out the advertisers who are trying it, seeing impressive growth fueled by the

incredible efficiency of Pontiac intelligence. the lowcost, transparent,

intelligence. the lowcost, transparent, and flexible DSP decision. They say, uh, Garris, happy Hanukkah. How's it going?

>> Happy Hanukkah, Adam. It's going okay, man. You know, I'm I'm not going to lie.

man. You know, I'm I'm not going to lie.

I was getting a little bored. Uh, if

you're on adtech socials, like, they were really quiet the past few weeks.

And, you know, something happened in the past 24 hours and >> and now they're blowing up.

>> It's popping off, baby. And I'm here for it. It's really messy out there and I'm

it. It's really messy out there and I'm really just having a good time right now. Um,

now. Um, >> yeah. So, for folks for for our viewers,

>> yeah. So, for folks for for our viewers, you know, we do 30 minutes of setup. We

do a run through, we prepare, we get our sound and our visuals right. And, uh,

this time there's there's posting action during the 30 minutes. Uh, there's

there's, you know, adtech uh, luminaries are arguing on social media. So, we we thought we would do a playbyplay, but then we decided we wouldn't cover it at all. We'll just talk about our own [ __ ]

all. We'll just talk about our own [ __ ] instead.

>> Yeah. And we have a game plan for this, but I think that uh I mean, it is a little bit of awe because I know things are good when someone messages me and I'm going to give Paul a shout out because Paul messaged me on LinkedIn and

was like, "Gareth, you're being too mean." And and I I that's that's a buy

mean." And and I I that's that's a buy sign for me, baby. That's that means things are going as they should be. Um

but >> I get that all the time. I just thought, you know, I prefer that it's not from any of my investors who are, of course, my bosses, but otherwise it's I don't mind.

>> Yeah. Do you have a board? Bosses.

>> Don't have a board. We do. You know, we have angels.

>> Angels.

>> They're alone for the ride, my friend.

They're looking.

>> I don't know, man. So, the chat, it's gonna be a little light in the chat because we're an hour early. Why Why are we an hour early today?

>> Uh, an emergency dental visit that I have to facilitate.

>> Oh, the dentist.

>> The dentist. Yeah,

>> I know. Um,

>> spot on with that, Mel.

>> This is life. This is life when uh it's about to be the holidays, but I feel like 2025 has been like five years. I

don't know about you, like I feel like there's there's been such an amazing trajectory of things that we've talked about like starting in January 2025, like thinking back to the beginning of

January 2025, the world was so profoundly different. It really was. It

profoundly different. It really was. It

feels Yeah. The CES last year's CES feels like a million years ago.

>> Yeah. I mean, it's great because it means that things are changing. I feel

like there was a five year there was a fiveyear period in adtech where the years were interchangeable. Like nothing

was really like shifting that much.

Like, oh, we added another ad exchange, we bought another thing. Um,

this is what was going on. Um,

>> I feel a little burdened by it. I feel

like I spent what by January I was, it was five years, right? I wrote I wrote the Chalice Business Plan on over Christmas break 2019. So, this is from my point of view, Charles's birthday is

this week. And last year at this time, I

this week. And last year at this time, I was on five years of telling people, hey, AI is coming to adtech and it's going to be really cool. It's going to add all these new capabilities. It's

going to bring more efficiency to what we do and it's going to be easier for brands to drive growth and predictable whatever. I had my pitch down five

whatever. I had my pitch down five years.

>> Yeah.

>> And then between then and now like literally like a hundred people have come in with the exact same pitch and some of them some of them being

gigantic businesses have come on saying very much the same thing. Uh, and you know, from my point of view, not not having done the the build to to back up

those claims, right? Not having the case studies or the demos or um, you know, even the credentials in some cases to be saying these things. Uh, and though my company has continued to grow, it was

actually easier to sell in the first five years than this year. And it's it's been a bit of a disappointing year for us in terms of revenue growth or our our uh, >> position has grown a lot. And we we've

had a great year. Chalice is doing great and we're growing and we're still a really hot company that's doing a lot of great things, but it's been harder to onboard clients to get them to try.

They're more tentative because they think we have to be very careful about AI and there's 100 companies that want to do a pilot and my agency says they're building something and uh it's made

people Yeah. It's it's just created more

people Yeah. It's it's just created more bar uh barriers to to trial and trial is the most important thing in AI. Imagine

if you hadn't tried chat GPT or something like you wouldn't know anything about it, right? If you were just hearing about the chat bots, ChatgPT and Anthropic and and you were afraid to use them. How much would you wouldn't know anything, right? You

wouldn't know what it is. You would

believe anything anyone tells you. Uh

it's through using these things that you understand what they are and what they can do and and how to use them. like

they're they're tools like right they're they're open-ended in ways that are fascinating and that let you bring your own creativity and ideas and and projects to them and and complete those

projects. So that's the way it is in

projects. So that's the way it is in adtech but uh it's it's gotten harder for people to to experience it because of uh the zone got kind of flooded with [ __ ] this year.

>> Yeah. I I mean it's I mean I guess that's a clap maldo.

I mean, it's it's hard, right, when >> Yeah.

>> people start selling using your language and they're selling [ __ ] So, people try it and then it burns you, right?

Because now you have someone who's disillusioned with your pitch and you have to be like, "Well, no, no, no, no.

The way that we do it is real. the way

that they don't. And of course, like everyone you're pitching, every sales meeting, they hear that and they're like, "Well, yeah, of course you're going to say that you're the real one."

Like, everyone is going to say that.

>> Um, >> yeah.

>> Yeah. It's it's

it was a high quality rant.

>> We're all me and all the AI adtec competitors are swimming in a pool together. And I would just say, "Don't

together. And I would just say, "Don't poop in that pool. We're all swimming in it. Don't lie."

it. Don't lie."

Well, the good news is in my experience so far in 2025 the people seem to care about how things are going and like if I think back about

the trade desk's moves this year. The

reason that I I I'm I remain surprised by the trade desk moves. I remain

surprised and I don't mean any offense by this to anyone, but some people might find it slightly offensive. Um, I was under the impression, and I've said this

to death in my articles and wherever else, that adtech is really a slush fund for agencies. And no one in adtech ever

for agencies. And no one in adtech ever really cared about performance. Well,

not no one, but many, many, many participants in adtech never actually cared about performance. And it was a place to allocate budget where they could get rebates. And because of that,

a lot of adtech was built in the image of performance is a commodity. We can

make anything look like it's performing.

And what is actually important here is how well do these tools allow our agencies to recoup margin. Um, that was

my very cynical take of adte from >> that's before you come on. That's that's

a that's some that's someone ignorant of the buy side. You you've learned more since then. I have. Well, and not only

since then. I have. Well, and not only have I learned, the adtech has put its money where its mouth is. The trade

desk's moves are performance moves.

Like, say what you want about them.

Their moves help them with performance.

They make campaigns perform better. They

were completely unnecessary if they didn't care about performance. Uh, and

that's like really cool. I've been in adtech a long time. I haven't seen a lot of DSPs actually make the right performance moves. And that was where my

performance moves. And that was where my skepticism came from was because my exposure to the buy side was DSPs. And

DSPs whenever I talked to them on the phone about like, hey, this is going to work better if you do this this way or hey, this is going to work better if you do this. I know the pipes. I know the

do this. I know the pipes. I know the plumbing. Like you don't um they would

plumbing. Like you don't um they would be like, we don't really care.

Yeah. It's like, yeah, yeah, yeah. But

we work with Pubmatic and, you know, >> that's enough for us.

>> Specify what moves you're talking about.

>> Yeah. So, I think the creation of open ads and the pressure on T um the cynics on the sell side say that those were self-s serving moves by the

trade desk to maximize their margin. I

don't think that's the case. I think

that they could maximize their margin without those things.

>> Also, trade desk margin is public. So,

if they were doing things just for their margin, like it would show up. it'll

show up. I think that those moves were real moves towards transparency and how do we make auctions more efficient? I do

not think that they're full of [ __ ] I really don't. And like those moves are

really don't. And like those moves are not always politically popular with a lot of their clients. That's the other thing that that astounded me about it is

any agencies who are in this for the kickback part of it, which I know you swear up and down is not widespread. I

think there there is some of it though.

It's definitely there. I I hear enough stories, Adam. I hear like even today I

stories, Adam. I hear like even today I still hear them. Um but like a lot of those agencies are like, "Hey, can you stop adding all this transparency into this like like we don't need this. We

have enough. We want to make sure we don't serve to bad websites." Right?

That's our the the amount of transparency we need is we don't serve to bad websites and make the campaigns look like they perform well. Um, and so I was really like 2025 was a year when

adtech when programmatic took probably the the largest steps towards performance that I've ever seen and they were big steps

>> at tag on there besides trade desk there's indexes index exchange allowing really sophisticated sellside decisioning right modular AI in their

servers which of course creates uh more competition in the DSP market because now you have Pontiac the transparent flexible a lowcost DSP as a perfect partner for that sellside decisioning

that index has has uh invented and then brought to standards now launched on OpenX. It's clear that this thing is is

OpenX. It's clear that this thing is is going to grow and it's there's there's no there's no way to see this as a cynical grab. Like this is a this is a

cynical grab. Like this is a this is a move toward technical efficiency that you know is is is it's right there for advertisers to take advantage of. Right.

>> Yeah. The only time it gets dicey is if agencies spin up their own version of it that just the deduct margin. Um, but I don't think that they will. And I think that all of your clients would attest

that like, hey, this is actually driving performance because how could it not?

Architecturally, it's great. Like

everything's dduplicated in a single exchange. And this was a brilliant

exchange. And this was a brilliant conclusion by index. They were like, hey, we know that we have direct access to all the publishers that you could want. The era of a DSP thinks it needs

want. The era of a DSP thinks it needs to listen to 20 exchanges is over. And

it started with SPO, but SPO was always soft, man. Like when when SPO first came

soft, man. Like when when SPO first came out. So this is this is all right.

out. So this is this is all right.

Here's more of my cynicism. SPO as a term probably rolled out what, four or five years ago. That's probably when it became a popular thing for people.

>> You should save this. This comes up in our my five scams. Five scams that still exist. So

exist. So >> Oh, we'll save it. We'll save it. We'll save

>> Did you do your five? Did you do a list of five as well?

>> Um, I'll I'll pick the ones that are all Let's do the segment. So, everyone,

>> Adam and I, five buyside and sellside myths busted.

>> Yeah, look at that. Sherlock and Holmes >> and I I feel really good about this.

>> So, I had a list of 10 and Adam made me skinny it down because he's like, Gareth, these are some of these are too in the weeds. They're too small.

Augustine Fu talks about them. You know,

there's no reason for us to mention these that that charlatan. Um, but I definitely >> Dr. Augustine Fu, thank you. He didn't

spend five years at fraud college getting a PhD to be called mister.

>> Is it his PhD from MIT?

Um, but that aside, um, it is I mean the the fraud department at MIT is very strong. Um, so I originally had 10, but

strong. Um, so I originally had 10, but I'll winnow it down to five. And part of what I want to reflect upon here is how far we've come. A lot of things have

actually gotten much better. So, I'll

pick five things. I'm going to pick two old ones that I'm going to tell people about because you guys seem to love it when I talk about adte like scams. And a scam is too strong a word.

>> I love it because I take a nap.

>> I like to be able to take a break during the show.

Well, I'm just going to tell you what you were buying when you were at Horizon seven years ago that you had no clue about. Um,

about. Um, >> that reminds me of a joke.

>> Why don't Jews have necks? You ever

heard of this one?

>> No, I've never heard this one.

>> Thanks, Maldo.

>> But, um, I think we've solved a lot of things. like a lot of things are fixed

things. like a lot of things are fixed and what's left to fix is when I'll shamelessly pitch Camra. Um

but we've solved so much. Um so do you want to do your five first or should I do my five?

>> You go first.

>> Okay, we'll do mine. Um

let's talk about All right, let's do five BS ones. I'm

going to start out with one that people talk about on the internet a lot that if you're a buyer, you've probably heard about and that's domain spoofing.

>> So domain spoofing is >> that's ads. TXT fixed this.

>> You look angry.

>> I don't know that ads.txt fixed it because people became very promiscuous with who they let get into their ads.txt txt file. Um,

txt file. Um, and it meant that, you know, a lot of people get to say they're selling publishers inventory.

>> I was afraid you're going to extend that sexual metaphor.

>> Well, I mean, their ads somehow, >> but let's let's domain domain spoofing is is crazy ad tech history. like there

there was no there was nothing preventing you from saying your shitty website was msn.com. This was a huge gap that was widely exploited became a huge

problem and a new technology came about called ads.txt that supposed to solved it but you know it's kind of on publishers to make sure that their uh their authorized sellers aren't spoofing

their domain >> spoofing. And so I think the real fix

>> spoofing. And so I think the real fix for this, the real fix was the middlemen in ads in in a domain spoofing world actually didn't have a lot of control.

Um so prior to header bidding domain spoofing was rampant. And that's because when one exchange called another exchange server to server and said this is msn.com there was no way to know that

they were lying. There was just no way because like whatever they say is what the JSON packet is. It's very similar if an >> bid request was not on the page anymore.

It had been >> once it's serverage, >> you can't tell what's going on. And the

same DSPs had the same problem, right?

They could only discover domain spoofing after the fact. They had to buy the thing, get to the page, and then be like, "Hey, this is [ __ ] This is not where it's supposed to be." Header

bidding largely fixed domain spoofing.

It largely did. The reason for that is the exchanges who actually turned out to be relatively trustworthy when it comes to domain spoofing, right? Rubicon and

OpenX and all these guys are now on the page and they can see what domain they're actually serving to.

>> That's because every publisher put got some header bidding tags so they could get more demand. So now attached to the page.

>> So the fact that domain spoofing is a big problem in 2025. If anyone tries to tell you this, they are wrong. We have

largely solved this and I'm sure Augustine Fu is screaming at that he's watching. We're talking about it because

watching. We're talking about it because in 2025, this is something that's fixed.

When you buy an impression from a domain in 2025, you are probably serving to that domain. This is a clean, happy

that domain. This is a clean, happy thing. And do not let people convince

thing. And do not let people convince you otherwise. Um,

you otherwise. Um, >> so 2025 domain spoofing fixed. The fact

that it's a problem, it's a myth.

>> Oh, the fact that's a problem is a myth.

Isn't it a problem in CTV?

H well so this is why I never buy CTV inventory. Um

inventory. Um >> it could be it could be.

>> So the problem in CTV and this is a this is a good one. The problem in CTV is that everything is server to server like we were talking about right? What fixed

domain spoofing in the web is the fact that ad exchanges are serving into the actual website. They can see where they

actual website. They can see where they are. That's that's where everything

are. That's that's where everything originates. In CTV it's servers talking

originates. In CTV it's servers talking to servers. And so a publisher or an

to servers. And so a publisher or an intermediary who is sketchy can say that they're selling whatever it is they want to say.

>> Yeah.

>> And and that's really hard to deal with.

Um because in the CTV world, the way that you verify is actually much harder.

>> Um >> so next myth, domain, it's good news.

Domain spoofing not really a thing.

>> And CTV, I would add, if you use Trade Desk, they're really good about policing that kind of spoofing in CTV. like

they're a a market maker that's does a lot of policing for you.

>> Yeah, this is one where the old school solution is the right one for CTV which is no know your seller even though a lot of sellers bundle [ __ ] in. Um well let's talk about another one and I want to

talk about ad stuffing. This is still a thing. Um, so if you are a buyer, one of

thing. Um, so if you are a buyer, one of the risks from buying from resellers is that I have seen a proliferation of

resellers in adtech that take one ad slot and turn it into five. And the way that they do that is normally a magic word called carousel.

So, a carousel is a quote unquote rich media unit that serves five ads into one and then rotates them. And for some

reason, people have described this as acceptable.

I do not think this is acceptable. I

think that is a a sneaky way around. How

do I shove five ads into one and sell all five?

Um, so if you are a buyer and you are ever buying from a reseller, this is a real thing and you need to understand from your resellers exactly

how and why they're serving ads to the page and what their utility is. Um, so

as we've talked on the show before about good resellers and bad resellers, there are certainly good resellers out there who have unique inventory that are worth buying from and then there are bad ones.

And so in 2025, I would say this is a problem that remains. There are people who figure out ways to turn one ad slot into many via rotation or via refreshes.

And if you're buying and there's a reseller in there, you better talk to them about this.

>> Yeah, the Chalice Deep Sea solution to this is is uh pricing on the number of network calls and ad density, both of which get spiked by this behavior.

>> That's interesting.

>> And refresh rates.

Yeah. So,

>> all our metrics are married to a page, a full URL.

>> It's hard to get hard to get away with things.

>> The sneaky thing that these guys will do here is they will not always search the page.

>> They a sneaky reseller will buy one impression and turn it into five, but they'll do it with a frequency cap and they'll do it like not consistently.

It's not a unit that sits on the page.

It's someone buying a normal ad slot and turning it into five.

>> That is the thing. And it's

>> it's a this is a know your seller situation too. This is one that we do

situation too. This is one that we do not have a programmatic solution for other than other than which even even we would like have to really eyeball this.

We'd have to really think about this. Um

this is something that would show up as rapid refreshes to us.

>> Um but it's this is not a trivial thing to catch.

>> Um yeah, something so something to look out for. Another reason like if you're

out for. Another reason like if you're buying from resellers and you're so listen listen buyers listening to this these are real issues. If you are buying from resellers you better talk to them.

If there's a reseller in your buy and you haven't talked to them you have like done yourself a disservice in my opinion. Um and a lot of re do a lot of

opinion. Um and a lot of re do a lot of buyers still buy from resellers Adam? Is

that a thing?

>> Yes.

>> Yes. There's a lot of pressure on CPMs and there's if you have if you want low CPMs and you're pretty good at measuring outcomes then you know it's it's I I don't argue with someone who says we're

not going to worry too much about it.

>> Oh, Adam, think about what I just said.

They're taking one ad slot, they're turning it into five. Guess what those five are?

>> Cheap. They're inexpensive. They're on

the same domain, on the same page, and they look like cheaper cheaper places to buy than the direct slots for this very reason.

>> I hear you. We actually right when we met, >> we never talked about this. Me and

Gareth only had a few conversations before we took our our our dialogue to the show, but one of the topics was this, like we found really cheap reseller uh placements.

>> Yep. that we couldn't find anything wrong with the page. And you told us to look into the placement IDs and see if those were actual placement IDs that were usually on that page. And the

answer was no. Like they were extra placements that somebody was squeezing into a page. It's hard as hell to detect. And yeah, you're right. We

detect. And yeah, you're right. We

should stomp IT OUT. BUT

>> THE noise of them guys.

>> It's hard to go micro on this one.

>> Yeah, I I think it's just you got to have a DSP should have reseller off by default. That should just be a thing.

default. That should just be a thing.

Gosh, if I if I had a DSP, that's how I would build it. But then everyone would [ __ ] at me about my CPMs being high. So

maybe maybe that wouldn't work.

>> Yeah.

>> Yeah. God, I just I learned from a client, big corporate client, that um that employee bonuses on media teams are often calculated

uh with the unit cost with the unit cost of an ad. So the CPM is is like the I don't know the numerator or denominator.

I'm not good enough at math to to know, but the the bonuses are structured so that the lower the unit cost of the ads you're buying, the bigger your bonus is.

>> Adam, didn't you tell me earlier in the show that, you know, like the buy side was like getting all super clean and everything was getting super nice? This

is >> I guess I guess you were talking about agencies.

>> I don't think I said that.

>> No. God,

>> I have a lot I think I mean the advertiser who's telling me about this was fixing it.

>> Yeah. I mean, power to them. But this

means that it exists. There are

definitely other people who do it this way. There are definitely agencies who

way. There are definitely agencies who do this, too. There are definitely agencies who do this who can get the cheapest [ __ ] Um, which brings us to our next one. That is I actually think this is this is probably my number one

for this year. Um, video placement misdeclaration. And what I mean by video

misdeclaration. And what I mean by video placement misdeclaration is for many many years everyone who's been on the internet has seen these little players that go to the bottom right of your

screen the little outstream players for many many years they sold that inventory as instream and this is because instream was hard to define right how do you actually define a YouTube placement on a

page what makes YouTube YouTube is it the fact that there's content behind the ad is it the fact that the aspect ratio of the screen is a certain thing like

and the poor IAB had to deal through this where they were like wait what actually makes like from a technical perspective what makes YouTube YouTube

>> um and how can we put this in a spec and it wasn't easy it's actually a really hard thing to do >> um and because the second right they wrote a spec to try and make things like

YouTube youtuby um people would find the edge case where they're like oh okay well here's how we can satisfy these requirements while making it a little thing that minimizes to the bottom right of the page.

>> So, I think this is still a thing. I

think if you go if you go into your DSP and you buy instream today, you're not buying YouTube still.

>> You're still buying a bunch of stuff that's saying that it's instream and it's actually outstream. Um, and the video placement misdeclaration, a lot of people have come a long way on this

where they're making people describe this as outstream. Um, but I don't think we're there yet. And I think 2026 is the year when this really goes away. I think

this is when this when this really croakkes and people stop having those little players on their pages because I've also I've also heard feedback from agencies that they don't want to be on pages where those things are present

because they're worried they're going to cover the ads.

>> Um they're like, I don't want something floating on top of my ads as they render. Um,

render. Um, >> yeah, I think a lot of times when you hear anonymous sellside entities screaming anonymously in the press about some some crackdown, you know, that

they're screaming that publishers can't make money. It's it's often

make money. It's it's often so-called publishers who who have no prayer of ever producing original video content. Original video content is the

content. Original video content is the requirement to have quality video ads, right? You have to either license it or

right? You have to either license it or make it. So, it's a higher level of

make it. So, it's a higher level of publishing than just having some information site with a bunch of slop on it and then be, you know, for years they've been able to show video ads on that page and they're all going to get

caught. Like, if there's no original

caught. Like, if there's no original content, if you're running pre-roll and there's no content, it's in a banner, it's in a popup. Uh, Charles, for one, is going to sniff you out. We're going

to stop buying it.

>> Well, and I definitely think there's a world for outstream video, right? Like

outstream video can work. It can be a thing. Um, like I buy it. I think, in

thing. Um, like I buy it. I think, in fact, I think that a lot of like Instagram ads are like a form of outstream video. They're like an

outstream video. They're like an interstitial outstream video. There's no

video content behind them, but you can run a 5-second video ad while someone is consuming nonvideo content. That's

possible.

>> Um, I think video stream. Yeah, it's I don't know. We got we got a question.

don't know. We got we got a question.

Good old RL7115 is asking if the IAB's overhaul of the inream outstream classification solved the problem and what did the IAB do exactly? Yeah. So

the IB worked really hard to try and solve this thing I just described where they were like how do we make sure that instream is actually instream and they got into the details. I think they did a

very good job and that's why this is changing. And the problem the hard part

changing. And the problem the hard part wasn't just the definition. It was

adoption and I think that they got Google to adopt it. I think that Google actually came out and said okay we're going to start labeling these things properly in GAM. Um so we've made a ton

of progress on this. I think that we're just at that point where there's like a tipping point in the world of progress where first we overhauled the standard now things are starting to be labeled properly and like in a lot of

circumstances they are and the next step is okay well what's the actual utility of the unit do people know that they're buying it and that level of education is what comes next where people actually

know what they're buying because they just check the video box in their DSP and they surf a bunch of videos into these and everyone starts to go oh wait that's not what I want to a little more pessimistic, but everything you're saying is true.

>> I hope so.

>> But yeah, that was my big one.

>> This was also the death of MFA in 2025, by the way. Not that I think that MFA was a sellite's fault at all, but um I think that >> like inventory quality. I mean, I I do

run an inventory quality company, so I'm a little bit biased, but um inventory quality became a real thing in 2025.

>> Like it was definitely starting in 2024.

Um, in 2024 it like hit started to hit really hard, >> but I feel like in 2025 it's like table stakes now. Like this is like this is a

stakes now. Like this is like this is a part of the industry. Um, it's crazy.

Like it's it's amazing. It's wonderful.

>> Yeah. Let me let me let me invite you onto your soap box and talk about how like we could save it for the 2026 preview episode, but let's put some of it here that the uh open internet is going to have fewer

better pages, right? like they hear a lot of crying that there might be fewer pages but fewer better pages is a good outcome and make us more like Facebook.

>> I totally agree. Um so my world the purpose of GA is to make the internet

perform better for advertisers.

Um in doing so it is my core thesis that publishers will make more money. I had a call yesterday, Gareth, I'm not going to reduce my ad density because I won't be

rewarded for it. And they were correct.

They will not currently be rewarded for it yet.

>> Um, I live in a world now where Gamera Game's goal is to get everyone to buy

with these sorts of things on every single campaign. Um, so we will have

single campaign. Um, so we will have simple segments that people can target that will re that run everywhere, right?

And that will reward publishers for low ad density. And my mission in life is

ad density. And my mission in life is going to be to get those segments slapped on every single thing possible next year >> where it's like, hey,

>> this is a low ad density viewport segment. Layer it into your buy. Just

segment. Layer it into your buy. Just

layer it in. I promise performance is going to improve. You can tell me how what performance you want to improve.

I'll improve your viewability. I'll

improve your CTR. I can focus on that within that. But also, we're going to

within that. But also, we're going to reduce ad clutter.

>> How are buyers going to get the signal?

>> So, I can send it two ways. I have raw data which Chalice will get and Chalice will be able to buy against and that should be live in Q1. And then I have simple segments. Um, so simple data

simple segments. Um, so simple data segments.

>> Ah, so I'm really trying not to really trying.

>> I'll give you a rupture. Well, no, no, that's that's my plan. So, my actual plan for this, I like if you look at my business plan, the primary person I'm planning on selling these segments to is

curators. Curators and DSPs. So, I want

curators. Curators and DSPs. So, I want to get the segments live inside of DSPs for people to adopt in DSPs. I've heard

mixed reviews on how well that goes. Um,

I'm placing a lot of faith in curators.

I believe in curators, actually, to a large degree. I think that they'll do

large degree. I think that they'll do whatever they can to make performance better for their campaigns, to get them more budget. I I'm not gonna speak for

more budget. I I'm not gonna speak for the fact that a lot of them take very very very high margins. Like I I'm I'm gonna let competition solve that one. Um

but I I'm gonna lean really heavily on curators to be like, "Hey, every single curator should be using one of these segments. Everyone, I'm going to price

segments. Everyone, I'm going to price them super cheap and I'm going to just make things work, right? You want high CTR, you get high CTR. You want high viewability, you get high viewability.

We'll crank down the ad density on all of it and publishers will be rewarded for this now." And that's actually a wonderful thing for publishers. They get

to reduce their ad load and they make more money because buyers >> pushing in the same direction. Yeah.

Penetration's very low to like uh >> there's over a 100 million running through Chalice and we just got to our first 100k month of uh running through

the quality PMPs with metrics by deep sea. So you know a pace of 1 million or

sea. So you know a pace of 1 million or 1% of the dollars running through Chalice now have this incentive toward better pages. But expect we're looking

better pages. But expect we're looking at we're expecting a big year.

>> You're in charge. Why is it so low?

>> It's new. It's new.

>> We had to scrape the whole web. We were

building it with Sincara and Trade Desk pulled Sincara away from us. Partnered

with Deep Sea and we took months to uh scrape the whole web and get the the data flowing through it. It was a huge project.

>> I mean, that's crazy.

>> Just about done.

>> I I I I've crawled the entire web in the past month. um

past month. um >> in a month.

>> Yeah, you get big domain lists, man. It's not

hard. Um

but yes, I think this is the future. I

think this is the future. I really,

really, really do. And uh I because I think if we can fix this part of programmatic, right? If we can

programmatic, right? If we can simultaneously fix quality and performance because it has to perform.

Um and if people are asking for CTR, then we have to give them CTR. If

they're asking for high viewability, we have to give them viewability. selling

addition selling new ways of them measuring performance while also selling them quality I don't think is going to work and I think that that's what a lot of people have been doing and I think a lot of people have been going in and

saying oh well you should change your goals and write target and we fix quality at the same time and I actually want the inverse to happen I think it should be fixing quality should allow

advertisers to hit their goals more cost-effectively and if that happens that's the tipping where like a this data is every >> yeah how do you do I mean how if the

CPMs are higher and the measurement is Google's which doesn't care about anything but CPM in their own mysterious attribution how do you get better performance with higher CPMs >> which measurement like postview

conversions >> yeah so the way that we do it is we measure inventory against a taxonomy

that no one else has access to. So right

now, >> but Google is the official measurement of the advertiser. You're saying you have a competing measurement.

>> Ah, it's not competing measurement. So

any measurement that Google does, they're going to correlate back to impressions that they purchase.

>> Yeah.

>> And I'm willing to bet most Google measurement has a viewability requirement. So most people, even if

requirement. So most people, even if they're doing post view, there's an active view requirement on there.

>> Yeah. I can deliver active view impressions more efficiently than Google can.

All right. More cost-effectively, more accurately. And I know this because I

accurately. And I know this because I have a direct sales product that publishers use where if their buyers are asking them for viewability, it can be hard to optimize to a buyer's viewability goal when you're targeting your Google Google ad units. Even if

you're using key value pairs, even if you're using fancy stuff, it's a pain in the ass. Um, we are really good at it.

the ass. Um, we are really good at it.

And the beauty of this is when you're buying against a signal that no one else has access to, it ends up being really cost effective.

>> Of course, it's moneyball.

>> It's moneyball. Now, I I may one day become a victim of my own success. This

is possible, right? If like everyone is it, it's it's okay. The numbers go up, but that's at a it still is at a more efficient place.

>> Um, and this is what's coming. This is

2020. My my world in 2026 is getting our data segments implemented on every single campaign on the internet.

>> Yeah.

>> We're competing.

>> We're not competing.

>> Inconceivable >> because I I don't do optimization. That

is allow me to say this. I am not going to optimize your campaign for you.

That's uh that that is the that is where I draw the line. I am but a data provider. A simple

provider. A simple data passer through her. Um, and that's a core distinction. I actually think that

core distinction. I actually think that those things get muddled. And you posted about this the other day with the OpenX thing. Is this measurement or is this

thing. Is this measurement or is this optimization? And good data can help you

optimization? And good data can help you optimize, but it doesn't optimize.

>> Like because the data is the data >> and you need to change what you're buying based on what the data says. And

data will not do that for you.

>> Um, >> thanks guys.

It's true. People get confused about this, man. They're like, "Oh, so your

this, man. They're like, "Oh, so your optimization." I'm like, "Oh, this is

optimization." I'm like, "Oh, this is not really like >> I'm not running stuff for you. I don't

I'm not in there every day running reports."

reports." >> All right, I got to bust my myths.

>> Bust your myths.

>> We do have a All right, I'll bust my Did you guys make graphics for my five?

I got to get my envelope. I got to get my envelope. All right, hang on.

my envelope. All right, hang on.

Yeah, Jonathan. I think I mean it has its issues. I mean, who was I talking to

its issues. I mean, who was I talking to about this? Post view assuming it was

about this? Post view assuming it was viewable. Ah, nice. Um, the idea of

viewable. Ah, nice. Um, the idea of tracking post view conversions on non-viewable impressions is just so astounding to me that anyone would ever posit that as being a thing. Um,

>> but the market didn't tell you anything.

uh sorry >> Amazon kind of do a similar thing where they only count attribution for viewable impressions but the IAB definition for all post view but there are flaws to

that also >> I know that >> you should throw out the non-viewables >> how could who doesn't >> it should be an even higher standard

yeah Google at all yeah it's All right. What do you say?

right. What do you say?

>> I I'm in a much different place at at I went I went much I'm speaking to the advertisers. Maybe I was influenced by

advertisers. Maybe I was influenced by this conversation with a big client yesterday.

>> I'm I'm I'm pointing myths right at at advertisers mostly about the things they hear from their teams that are just uh not true

or or uh yeah, not true or or not true anymore. All right. So,

anymore. All right. So,

and just we talked about pages. I kind

of I kind of went this is going to be a US- ccentric discussion because I really am talking about mostly issues around IDs. Take advertising to first

IDs. Take advertising to first principles. There's a set of people you

principles. There's a set of people you want to reach with advertising. Right?

The whole idea of chalice is there's an optimal set of people to reach at any given day. Right? And but there's an

given day. Right? And but there's an optimal spend of this budget. The people

who you would get the most benefit from advertising to, right? And that could be households or individuals or devices.

And then pages is a bit of a proxy. But

the the unit of advertising is the person, right? So if this was

person, right? So if this was understood, I think I think AI gives a chance to reset of that and say and say, "Hey advertiser, start from this." And

the first thing how many people is it?

How many people is it right? How many

people do should be targeted by your campaign? And then you could get into

campaign? And then you could get into cost. But the first thing you should

cost. But the first thing you should know is your US audience is how many households are eligible for this advertising. There's eligibility and

advertising. There's eligibility and then there's optimal, right? If you're

selling face raisers, people who shave their faces is the optimal audience, right? And the and in the past these got

right? And the and in the past these got turned into all kinds of crazy proxies that are not people who shave their faces and therefore need a face razor, right? So it's it's actually you could

right? So it's it's actually you could get to the actual number now uh through the there data is lying around right how many razors are purchased you could get really close to the actual number how

many people you know stop shaving and grow beards >> that's crazy whatever >> that's so interesting >> you can get all the numbers so the first thing I want to attack as is as as the

myths and and this is I I kind of blame Tradees for it I don't know if it was their fault the first thing the first but this is where I first heard it is when you run programmatic campaigns and

the the operator says it's not scaling.

We don't have scale. Scale is this huge myth that there's some other concern besides the size of your audience and the ability to reach them, right?

They're using a software word to imply that some that there's some software issue. The actual issue is either you're

issue. The actual issue is either you're targeting too narrowly or you're targeting somewhere where the audience is not or you're trying to pay too low, right? like you're just

avoiding the issue and and almost all the time that people say scale, it's because of targeting too narrowly or they've made an SPO decision that's preventing them from being where the

advertising is. Uh but the worst thing

advertising is. Uh but the worst thing about this view of scale is that it's pushed the cell side to fake IDs, right?

to expand IDs to stuff segments to do all kinds of ID bridging so that people have this experience of like well I targeted the segment of razor buyers and it's not scaling so then the sell side's

like all right we'll put every household in America in razor buyers >> yes >> now it'll scale and it's just >> you want razor buyers I will I will give you a probabilistic slew of razor buyers

>> no it's like now you've bl now you've blown the first principle which is don't waste the razor ads people who don't shave their face or know anyone who shaves their face.

>> This is a like primary actor issue where if there were one company defining the audience and delivering it, you wouldn't have the

issue. But the problem is you have one

issue. But the problem is you have one company defining and trying to deliver and another company who is involved in the actual execution of the delivery and

it creates a perverse incentive for the second party where the the second party is basically getting a signal if you tell me this I will buy more.

>> You're being nice about it but there's there's there's no reason there can't be a unified reach metrics. There's no

reason there can't be entirely private safe measurements of reach. And that

brings us to number two, GDPR and privacy. Any privacy law that prevents

privacy. Any privacy law that prevents the measurement of reach is [ __ ] Right? This is this was a boon for

Right? This is this was a boon for YouTube for Google to say, "Hey, Google has the best reach, so we should have the most money." And you can answer the

question, how many YouTube watchers who also buy razors, can I reach somewhere else, somewhere in a better environment, with a bigger screen, with a more engaging piece of content? Right? You

can you cannot find out because Google got away with saying that their IDs are first party private, can't be shared, can't do any reach matching. You cannot

find out how many of your ads went to children and how many went to men who shave.

I will say I see a lot less people belly aching about users and their privacy preferences at the end of 2025 than I did at the beginning. Every time I saw one of those posts,

>> everybody hates the fabs and they're so much worse in Europe and England. All

right, that's it. I mean, all obviously all that shit's [ __ ] but I think what doesn't get mentioned enough is that advertising reach is like an important thing for the economy. Every

company needs to be as efficient as they can in reaching their target audience.

And these laws that make it impossible or at best extremely difficult are total [ __ ] They should be eradicated and there should be there should be reach measurement.

>> I agree. It makes it it makes it so much easier, right, for the big tech companies to win this battle.

>> What is this? Is this the company he works for?

>> No, he's he's independent now, man.

Jonathan used to work at Kepler, right?

Oh, this is >> research our listeners chat. I I look at who all of you are and I love all of you. Um

you. Um >> that's good >> aila. What's that?

>> aila. What's that?

>> That's an initiative. I think uh I think my friend Tina Denells is on that. It's

uh Yeah, the ANA is doing a project to selfreach. The Association of National

selfreach. The Association of National Advertisers.

>> Yeah, I like the ANA.

>> They they do interesting things. I find

I find like the way that they approach problems always fascinates me.

>> Um, >> all right. My third myth is that you have to pay one or two companies, I'll call it double verify is the worst for

viewability, fraud, and brand safety protection. This is insane. Like I once

protection. This is insane. Like I once I once did a bake off where before they started the bake off, they talked us all down to the lowest possible price. This

company got us down to 6% of media for their custom algorithm. Extremely low.

And then we found out we got in market they pay 6% of media to double verify for doing nothing. Like we made them a custom algorithm. Double verify is

custom algorithm. Double verify is giving them a check mark for for fraud and brand safety which they've been exposed a zillion times that they're not very good at. They don't have the

technology. Things get through. It's

technology. Things get through. It's

just like a it's just a tax on the whole industry. There's much better ways to

industry. There's much better ways to measure these things and monitor these things. There's companies that you could

things. There's companies that you could pay by the month. You could you could work without them and and get partners to do the filtering and give you all the feedback. Like these percent of this

feedback. Like these percent of this percent of media scam has got to end.

Yeah. I mean it's just one of those things where all of these are considered separate features and topics when in reality they all stem from one thing.

Know the website you can't be.

>> Yeah. Don't buy blind. Don't buy blind.

If you do that, all that other stuff is solved. Like, yeah, sure, pay someone to

solved. Like, yeah, sure, pay someone to avoid bots. Like, that makes sense to

avoid bots. Like, that makes sense to me. That's real anti- fraud. Like,

me. That's real anti- fraud. Like,

detecting bots is requires some sophistication and is a real thing for like really fancy bots. That's that's I buy that. Um, but like Yeah, man. It's

buy that. Um, but like Yeah, man. It's

crazy. I I do this view I if you want viewability, I got it for you better than they do. Yeah, viewability is I mean since what's it thing called that cross that cross-section thing there is >> intersection observer my friend

>> intersection observer it's free free tank >> yep >> track view >> people should wake up about that all right number four this this is a big one and bit of a hit on my agency friends so

the digital ones know that I'm what I'm talking about the digital media planning process is a total waste of time and money like if you get like they they're

still producing these reports being like, "Oh, the adult shaver, you know, spends times on websites where men enjoy sports and talking about entertainment topics enjoyed by men." It's it's so

[ __ ] stupid and wasteful. You could

get in and market and find these ideas in a couple of days and start to have really good projections on what it costs to reach them and the best pace to reach them and the most effective uh attributes that are associated with

this. So to to do it all in advance in a

this. So to to do it all in advance in a plan is a total waste is a total waste of time. The thing is that you should

of time. The thing is that you should decide in advance or what are we trying to do here? How much are we going to spend and which channels should be which channels should be

>> is that Marcus Zagorski.

>> That's uh the comment is please stop calling us a scam. We prefer the term sham. Thanks Mark.

sham. Thanks Mark.

Um yeah, so the the planning process is way too overloaded with with BS that should be found out in market, right?

It's the you should have a good strategy and then should have good optimization, but most of what's done in digital media planning uh is is extraneous.

>> Isn't there a lot of offline or online too? Like, can't you like take the

too? Like, can't you like take the people who have bought your thing, upload that, show them ads, figure out where they are, and then just buy those places for people who aren't in your list.

>> Is that not a thing?

>> Yeah. I mean, make a model like a and like the best possible model of of them, right?

>> Yeah. That's how we're going to market saying you have you you have a customer list and you're just your plan is to just send this customer list to every platform to train a their AI that they're going to use for your

competitors. That's insane.

competitors. That's insane.

>> Don't do that. Don't do that.

>> That's insane. You should hold that data. Train AI in your trusted execution

data. Train AI in your trusted execution environment and then send instructions down to the platforms. Pay this for that. I'm not telling you why. I have my

that. I'm not telling you why. I have my own AI anyway. It's a great pitch for the enterprise, Adam. I love your pitch.

It's a good one.

>> Little advertisers benefit, you know, a rising tide lifts all boats. Is not big enough to move the needle. But like for big advertisers, 100%.

>> It's money on the table is what I keep telling them. And this brings us to my

telling them. And this brings us to my number five scam sham of the year.

>> Aentic, right? Like every time you see this aentic pitch, >> ask what is the value for me the advertiser? How do I benefit? How did

advertiser? How do I benefit? How did

Butler Till's client benefit from this pomatic thing that got all the all this press? Right. If you're if they cannot

press? Right. If you're if they cannot quickly say the advertiser benefit, move on.

>> Thank you.

>> No more salazo tricks.

>> Exactly.

>> Like I like I read that and that makes perfect sense to me. And if the answer wasn't, hey, the me the planning and execution process was 65% cheaper in

terms of like manpower that that seems that's the benefit that was the benefit like that was >> the publishers who benefit the publishers benefit >> do they what did the publisher benefit

from these fewer FTEEs to set the campaign up that's did anything other anything that was not

possible like anything else I mean, as I just said, you don't have to media plan anyway. You could always go to Pmatic and say, "Give me razor buyers."

buyers." >> Oh, god. I There's not even savings there. I mean, I'm I'm sure that they

there. I mean, I'm I'm sure that they pay some, you know, person two years out of college $75,000 a year to like write this up and then email their PubMatic

rep. Um, that that seems automated now.

rep. Um, that that seems automated now.

That should be automated. Should be

automated.

>> Yeah, I I think it is for the seller. I

think automation for the sales side and for smaller sites to be able to sell guarantees without a sales team is a great benefit for them. They should they

should try to express a benefit for the for the buy side. I also like I had this conversation earlier and I went hard on this on LinkedIn and Twitter like

guarantees are just I don't understand in a programmatic ecosystem that has to

compete right with the walled gardens who don't really do guarantees like why do we need them like who who are they

help are they are they helping campaigns perform better like cuz there's no way you can know that every single impression on a given site is going to perform for your campaign. Now maybe you

can get down to the level of granularity in your guarantees one day. Maybe this

is a problem of maybe this is a promise of agent. Maybe there's a world here

of agent. Maybe there's a world here where you have you go to a publisher and you say I want precisely this from 6 to

7 a.m. every day

7 a.m. every day

>> and I know ex Oh no.

hunting >> and my kids are little. Um,

>> maybe that's maybe that's the world where guarantees become really interesting where if you know there's a there's a piece of inventory that I want to lock up because I know it performs

for my campaign and it's page level like we're talking really deep stuff here.

Um, maybe that would be an interesting thing to have automated guarantees for that advertisers could really benefit from. Um,

from. Um, >> yeah scarcity. Yeah, it's scarcity is the Yeah, >> thing you need.

>> Jonathan Duza Rato, is that how you say it? Rado, I don't get the value of it

it? Rado, I don't get the value of it for the use of CTV. You do not need AI to plan execute that pretty well. It

would have been much better to do it on the open web. What is what's the it there? The chalice pitch.

there? The chalice pitch.

>> The planning and the executing I think.

>> Oh, or agentic planning and executing.

>> Yeah.

>> I mean, there's still there's still a lot of decisions to be made in CTV.

They're not giving us the data on the user and the show, but you know, if the all the user and show data is available, there's going to be a huge variation in

value per campaign, per advertiser, and to the old job of matching buyers and sellers is yet to be done. And when it's done better, both the advertiser and the

publisher will benefit.

>> That's the trick.

>> It's the truth. We live in guarantee world right now.

>> It's uhic. You want you want that reporting? You guarantee us money at

reporting? You guarantee us money at this number that we pulled out of a hat.

>> So, let's let's do uh let's doin in our last uh last two minutes here.

>> Last two minutes.

>> We got to hammer LinkedIn because they they did those year end wrap-ups. Let's

see.

>> Typical Lincoln.

>> Look at that subtle off-white coloring.

The tasteful thickness of it. Oh my god.

>> I love that scene so much.

Madam, >> that was that was really wonderful. Like

really, Adam and I didn't get to see that. That

was a surprise for us today. That was

>> trash cans.

>> I love it.

>> All right, LinkedIn, the worst social media network that execution of everything they try to do.

Uh, we're celebrating them this week with their year end wrap-up. A lot of people made fun of it. This one is Kenny from Chalice

who made a mock post about having 1592 offers to revolutionize lead generation and the like. They might as well have just done this. Like that's how useful

the year- end wrap-ups were.

It's crazy to me that there's like a spam folder that a bunch of stuff gets into and my normal folder still gets filled with this [ __ ] but the spam

folder like I never even worse. I know the

even worse. I know the >> I got so much spam in my normal filter.

God help the spam holder. That's that

that thing is just toxic and it's just the the lowest quality. Um

>> and then you know I' I've kept you know just for this segment I've I've I've taken every connection request. I've not

been fil I've had like a thousand in three weeks from people I don't know and letting them pitch me and it mostly sucks. There's been nothing interesting

sucks. There's been nothing interesting to use on the show until yesterday. I

got this one. And can you can you put it up? The screenshot I got.

up? The screenshot I got.

>> Please join me in inviting President Barack O.

>> And this person, >> sorry, wrong button.

>> This one said, "Hey, Adam, enjoyed your conversation with Janna about curation transparency. The fragmolidation

transparency. The fragmolidation discussion was spoton." And then it goes on to a regular pitch for whatever he's pitching. So I think this is AI. I think

pitching. So I think this is AI. I think

they got the AI to actually say something relevant >> about our show.

>> Even fragmolidation getting in there.

>> The AI is getting good.

>> It's getting good out there.

>> It's getting good.

>> I've gotten one or two of these where I was like, "Wait, does this person actually know me? Like, do they know what I do?"

>> And then I'm like, "No, no, never mind.

This is not actually a person, but >> weird.

>> Eventually, >> our new employees always get AI, fake AI emails from me, and the one that fools people is is like, I need gift cards now. And people, you know, I've come

now. And people, you know, I've come close to buying the gift cards.

>> No, they haven't.

>> Yeah, it's like their first day of work.

>> Oh, but that's a classic scam. Oh, no.

>> Everybody knows. But if the AI is is there was one that was like too nice and they could tell it wasn't me. Oh, it was a fake invoice and I was thanking the coach for the coaching. The coaching was very insightful and I'm really happy to

pay this invoice. Here it is. I'm giving

it to my assistant. My assistant looked at and was like, I wouldn't be thanking a coach.

>> This is amazing infosc that you guys do.

You basically do penetration testing all the time to all of your employees with AI and then you're like, "Hey, you fell for it. Don't do that."

for it. Don't do that."

>> Yeah. It's scary because I'm so cons consistent. They might they might

consistent. They might they might actually get get me dead on eventually.

They'll have Marley making like a 50page deck or something based on one idea that I had in the middle of the night. I

hope it doesn't happen, but >> Oh, this just gave me a really dystopian idea.

>> Um, well, thank you so much, guys. This

is wonderful. Thank you, chat.

>> Thanks. Thanks. Thanks, chat. Thanks,

fans.

>> Ask us questions. Keep it fired up.

We'll be back January 20 January 2nd, 2026 to preview the year in advance of the year. And what's that? What's that?

the year. And what's that? What's that?

CES. In advance of CES, we'll be back to preview everything that will happen.

It's CES and in 2026. So, don't miss that episode.

>> This be a fun one. Thank you guys. This

is wonderful. Have a wonderful new year, everyone. Be safe.

everyone. Be safe.

>> Thank you. Happy Happy holidays.

Yo, the crowd is back. The lights are dim. Front

dim. Front