Episode 43: Chris Kane on what makes an MFA site. Plus bid request shenanigans at Freewheel

Welcome to the Markitecture Podcast. I'm Ari Papparo. I'm joined today by Eric Franchi and Chris Kane from Jounce Media. So Chris is our first time ever repeat guest. He's like the Steve Martin of the, of the Markitecture Podcast. And we have him on because this is a special episode to focus on MFA, made for advertising, of which Chris is an expert at, which is a very hot topic. So we have a lot of great, really great things to talk about. After the interview, we have... Justify your existence with Finestra, which is a bid optimization company. We spoke previously with Chalice. So it should be interesting to compare and contrast what the various companies in optimization have to say. All right. So with that said, Chris, thanks for being here.

good to be back.

All right. So, why, I guess let's start with like, why are we all suddenly obsessed with MFA? MFA's been around forever but it's like the season of MFA.

the season of MFA. I think it's been a slow boil and I think that boil boiled over. When the ANA pushed out their first look report in June there was, you know, a group of people, three of us plus a bunch of others, who are kind of like close to the center of the industry and have been generally aware of this business model for a long time. But all of a sudden, through the ANA report, every agency, every brand is starting to ask questions about this category of supply. And so I, you know, that's, that's my simplistic interpretation of sort of how we've gotten here. And by the way, that's going to happen. Again in another week when the ANA pushes out its final report.

Right, so the ANA report said that 15 percent of all ad spend was going to MFA sites, is that accurate?

15 21 percent of impressions capturing 15 percent of

And do you think the next report's gonna be worse?

And do you think going to be any new numbers in that next report, just for clarity, we're not, we don't, we're not part of the team that's writing that report. We didn't, you know, we weren't hired by the ANA for this. They've reached out quite a bit. We've been, you know, as collaborative as we can be. They've provided some, I think, really thoughtful questions. But I don't have a good sense of what's going to be in the

Right, so the MFA, so the ANA site, excuse me, the ANA report came out, and then there was a Digiday article saying, or was an ad exchanger saying, like, actually MFA is really not so bad, and we roasted them pretty hard on this

I recall, yeah, I remember listening to that one.

at agencies said we won't buy MFA anymore. But it all sort of begged the question, which is what is an MFA site? How can you determine? And I think that's where you have a lot of insights. So how do you define an MFA site?

well, and I, and I also think maybe that's the other answer to your first question, like, the reason this is starting to become a thing is that there's now some specificity around it. Like there's been, there's been a lot of anecdotes and you know it when you see it sort of stuff for many years. But you know, you can't really take action on, you know it when you see it. And I think what has also happened over the course of the last several months is sort of like some consensus being built around like how do you sort of like reasonably draw a circle around the pool of supply that we're going to call MFA. And frankly, there still is ambiguity there. And the reality is these are judgment calls, which makes everybody feel uncomfortable. They should be very data driven judgment calls, they should be very transparent judgment calls. But they are judgment calls. But they're much more, there's much more specific guidance now from the ANA, ISBA, 4As, I think the WFA signed off on it as well. There's this, there's this reasonably specific definition floating around the industry now. Here's my interpretation of it. We basically are looking for three characteristics of a website to call it made for advertising. And they, one leads to the next. Sort of characteristic number one is paid traffic. We should debate this maybe in some more detail, but We're looking for publishers that are highly dependent on paid traffic as sort of like a core component of their business model. Second, as a result of paid traffic, somehow you have to overcome those traffic acquisition costs, and one way to do it is to very aggressively monetize the landing page. And so factor number two is a very aggressive, aggressively monetized page view through some combination of high ad density, rapid auto refresh, and so on. And then third, is what we call superficial KPIs. Evidence that, while the site might exhibit some attractive media metrics, like high viewability, or brand safety, or low IVT, that it doesn't drive business outcomes for marketers. And, that's not exactly the same as the definition put out by those industry trade groups, but it's sort of, it's our interpretation of like, how do you take their guidance, and turn it into some

Well, that's how those sound like good criteria, but they are a little slippery when you really dive into them. Like,

It's extremely slippery. It's extremely slippery. BuzzFeed would be called an MFA site when they launched. Any media startup that doesn't have a brand, that doesn't have any awareness of their business how are they going to get the word out about what they're doing? They need to buy traffic. It's like super concerning. That this not only penalizes startups and, you know, other, you know, not, not, not well known media companies, but there's no specificity to it. And I realize that it's a problem and I realize MFA, you know it when you see it, particularly as you talk about ad density and not designed to drive marketing KPIs. But the whole thing, like when I read it this morning, it was like really concerning.

Well look, I think, of course I agree with, you know, your general concerns. But to talk in specifics, I think it's very important to also recognize that like, there are thousands, many, probably tens of thousands, certainly single digit thousands. of independent content creators that are doing the hard work to publish really high quality content and gradually, slowly, over time, build an organic audience. And for every publisher that takes the shortcut of just buying heaps of paid traffic, it's not just that those ads aren't effective for marketers, which they're not and we can talk about, it's that they're taking budget away from the media companies that are doing the hard work to gradually, methodically, over time, build an

right. So let's talk about that criteria. That's criteria number three, the doesn't produce real results. Is that a consistent criterion? How do you find that out? Like, basically, if is that you just have to be running, you know, a platform at scale where you have enough data to tell that it doesn't produce results?

a very important sort of like, corollary to this is like a blind spot for publishers. Our, the basics of our business is resourcing data from a whole diversity of sources. One of the most important sources is signing data sharing agreements with marketers, who say like, hey, you can study, in aggregate, anonymously, our campaign data to inform your research. And so we currently monitor 50 million per month, or 600 million a year of spend. Which is big enough and diverse enough to allow us to measure some of these things. And we can quantify for any website, what we call a conversion score. Which is, skipping some of the details, like does this site generate more sales than you would sort of naturally expect or fewer sales? And so we regenerate that every day for every site on the internet. And we are not willing to call a publisher MFA unless the data substantiates that the ads on this site are very ineffective. Driving sales may be the wrong word, but being attributed with sales by,

So does that look like high view through and low click through attribution? Or am I simplifying?

It's, it's, it's like none of either of those things. It's like you find that like, the sites on this website are almost never followed in

But I thought the whole point of these websites was to grab view through cookies.

I think what these sites are good at... is taking budget from campaigns that are optimizing to media metrics, and that are not measuring sales.

Yeah, view, viewability, clicks, all that

It's the consumer packaged goods companies, the other upper funnel, mid funnel campaigns that are, have a very hard time measuring sales, and instead optimize to a metric like a cost per viewable impression,

real humans seeing these ads? Or are the people seeing the ads also fake?

real humans

but they're having no effect on their purchasing activity.

Correct. And to me this sort of like checks out. I've clicked on plenty of clickbait ads. You go to the site, you might even stay there a long time. There's one that's floating around Twitter recently about, like, some story about like digging the deepest hole that's ever been dug or whatever. And it's like, man, I could, that is kind of interesting. I'll stay there. But like, I'm certainly not paying attention to like the insane flanking of the viewport of, you know, these, you know, high density auto

How

those, those ones you're referring to on Twitter, and by the way, Twitter X is littered with this stuff right now. It's becoming, I think, the place to do MFAR. That's a you know it when you see it type of site. Right?

I think that's right but I also think publishers rightfully are sort of demanding that like, we're not making a bunch of like, on the fly note when you see a judgment call. So there has to be some like, systematic, you know, even handed approach here.

How I was gonna ask on the first criteria, the paid traffic where do you get that data from?

ask on the first criteria, the paid traffic where do you get that data from? So, the best signal is the shape of bid requests real, like publishers that have real organic audiences have very predictable weekly patterns of bid request volumes. This week looks the same as last week. Publishers that are buying absurd amounts of traffic see incredibly jagged patterns of daily bid requests. And so that is, that is directional, that's not like a deterministic, like, oh, I observed you buying traffic on Facebook. But it's a very, very good predictor. And it's also a prohibitively expensive predictor to to avoid. You know, like it costs a lot of money to inflate your traffic volume. And so it's cost prohibitive to kind of like fake it forever.

What's the right amount of non organic

That's the question. That's the hard one, right? Because what you'd want let me say a couple things on that. First, I think. I would have been too timid to say this maybe a year ago, but we've done a lot of research on this, and I think it is very clear that a pageview that's the result of paid traffic creates substantially inferior outcomes for marketers than a pageview that's the result of organic traffic. It doesn't matter whether you're a clickbait site or the most premium publisher on the internet. Paid traffic does not perform as well as organic traffic. And so, like, conceptually, you'd love to say Publishers that buy paid traffic get blocked, but then you block the whole internet. It's very hard to find a publisher that doesn't buy traffic. And so it's like, okay, so we're now saying that if a publisher is totally dependent on paid traffic, there's good demonstration that that's a bad business product, business outcome for marketers and we should block that site. But we're acknowledging that even the most premium publishers on the internet buy some paid traffic. And so now we're having a very difficult debate about like, well, how much is too much? If 5 percent of my traffic is paid, is that okay? What about 10%? 15%? What if I get to 40%? 50%? Like, where is that line? And I don't know the answer to that

So you're saying the top premium publishers pay for some piece of the traffic?

100%. Very hard to find a household name publisher that isn't right now arbitraging clickbait ads that run on Facebook. I think

Why are, why?

there are basically three reasons. One is they really want to build their audience. You buy traffic in order to build new loyal site visitors. Two is you want to just simply arbit through programmatic advertising. You want to just like boost that. You're, you're, you're coming up short on your Q3 revenue and so you got to inflate it all and you're going to arbit. And we can talk about how that works. And then I think the other is you've oversold a campaign. That, you know, you thought you were going to have 100 page views, and you actually only have 90, and so now to, to deliver this campaign in full, you're going to buy the extra 10.

Right. Makes sense. So you send me, let's talk about false positives, or, or false negatives, or the you know, sites that are legit that are being swept in with this anti MFA activities. You send me a really interesting article from Digiday about black owned publishers who is being, who is being labeled MFA.

Yep.

going on there?

What's going on there? Well, I guess I, I don't know how to speak to the specifics, because that publisher, I wouldn't disclose who they, you know, who they were. But I have a couple thoughts on this thing. I think the biggest one is publishers have should be concerned about the risk that they are incorrectly being labeled as MFA. I see this all over the place when we work with brands and agencies who built their own lists. And Eric, you were sort of touching on this earlier, although maybe in a different context. There are a lot of publishers that, You know, publish ridiculous clickbait that should not be sort of like swept in with this category of made for advertising publishers. And so, you know, the tests that I described earlier I think are pretty thoughtful tests. And there's just like, there are a lot of false positives. We try really hard to avoid those false positives. I think there are also publishers that are not false positives. They're sites that really are clearly meeting the definition of made for advertising that I described. That you feel bad about labeling made for advertising. You know, you want black owned media to succeed. You want Eric, to your point earlier, like, fledgling media companies to succeed and grow. You want a competitive landscape. Having said that, this is not a matter of, like, who we want to succeed. This is about having a systematic, standardized way that is... That is robust, and, and, just the reality is, like, again, I don't know who that publisher was that was referenced in the Digiday article, but, yes, there are black owned media companies that are entirely dependent on ad arbitrage, and, shh, it would be irresponsible to not label them as made for

I don't know. I think the African American community should participate in the Chum Box. Economy.

I think there's, this is sort of like the perfect example of why this is such a thorny topic. It's just, it's so multi tentacled, like, there's just so many sort of angles on this thing. One important consideration here, as it relates to, not necessarily black owned media, but like diverse owned media more broadly, is back to the point I was making earlier, like all the other companies that are doing the hard work to build an organic audience. There are budgets that are earmarked, 5, percent of budget, at major brands and agencies, that, at least in the U. S., that are earmarked to be deployed to diverse owned media businesses, in order to foster the growth of a whole bunch of media owners that have been sort of like, pretty structurally sidelined in the industry. That's a very noble cause. Okay, now those budgets get deployed to a media company who's totally dependent on ad arbitrage. And two, I think, I guess three very bad things fall out of it. First is, paid traffic doesn't work and it's a bad use of the marketer's funds. That's actually kind of a secondary consideration here. The other two things that are really important when that happens is, first, It's not like that budget gets thrown away if it doesn't go to the Ad Arbitrage publisher. It would go to thousands of other independent, diverse owned media companies that are doing the hard work to build an organic audience, and Arbitrage takes money away from those publishers. And two, the flow of money is the agency pays its DSP, the DSP through the supply chain pays the diverse owned media company, and then the diverse owned media company sends 90 percent of that money to Facebook. That's not supporting the growth of diverse owned media. That's putting more, that's spending more money with Facebook.

So what should we be doing about this? Should we, should there be a universal exclude list at the DSP level and you just shouldn't be able to buy them? Or it should be brand to brand? Or some accommodation?

We field a so we write monthly research. Our December report each year is based on a survey that we field to all of our clients. And so we just started fielding it. One of the, one of the questions in there is, How should DSPs handle made for advertising? What's the right thing to do here? There is overwhelming consent that the answer is not that DSPs should sort of like defer their responsibility to their customers. It's like there's a point of view broadly across everybody we work with. They're like, no, the DSPs need to do something here. And there is similarly like really clear point of view that DSPs cannot be too heavy handed here. That it would not be reasonable for a DSP to make it impossible to buy made for advertising. The debate is, is this an opt in or an opt out? Should DSPs, should the default setting of a DSP be to include MFA and make it easy to push a button to say, no, I want to turn this campaign off? Or should the default behavior be, I want to block MFA, I expect my DSP to block MFA by default and then I want a button to turn it on as needed.

And there's an incentive issue here because the MFA sites have a huge amount of QPS because they have so many ads.

we've tracked this pretty closely. As a percent of web auctions, this is a web concept, as a percent of web auctions in June, MFA supply was 30 percent of all web auctions. So it's an absolutely astronomical amount of QPS. That is coming down. There's actually some interesting stuff there. But that is coming down. Maybe it's 20 percent of the bid stream now. It's still an absolutely enormous, multi million bid request per second category of supply.

And so there's also a carbon linkage here. So we talk about carbon a lot lately. Scope 3 got funding, 20 million last week. We had GoodLoop on here. This is a lot of wasted bandwidth and wasted CPU power.

Right, so I think there's maybe two angles to that one. First is, If you just think this is a waste of everybody's time and money, then 30 percent of the bid stream is a waste, and that's 30 percent of the industry's carbon emissions. I think that's not a reasonable point of view. The more measured point of view is what Scope 3 put out in their State of Sustainable Advertising Report, which is, even if you think the quality of made for advertising is not a problem, just on a you know, carbon per impression basis MFA sites are 26 percent more carbon intensive than the rest of the internet.

Chris how much do DSPs use your methodology to determine an MFA site or not an MFA site because I think you're like three categories that's really good right because Everybody buys traffic. So I think you know using that in a vacuum is bad But then when you start going down to, you know, the appearance of the page, number of ads on the page, how this then performs for real marketer KPIs, all that stuff ends up, you know, I think coming together into a black or white is this MFA or is this, is this not MFA? So. It would appear to me that blocking MFA as a result, if it's based on good methodology, is the way to do this. And then, you know, publisher by publisher, working with brands and agencies that aren't on, you know, some sort of designated list there's a process for getting them included.

I think that's all right. To answer your question, you know, nearly every ad tech company works with us and has access to our data, buy side and sell side. They do different things with it. But I guess the, the other thing that I think is worth highlighting here is, I think publishers are sort of like rightly frustrated that things like brand safety, for example, are a total black box. And this can't be like that. Like publishers deserve clarity without having to write a check into like how are you currently being classified. And should have a clear process, I think you were sort of getting at this, to sort of like dispute these classifications. And so we try really hard to make that easy, you know, like the data that we classify is available to thousands of companies that work with us and available for free to any publisher that thinks that we might be making a mistake or is just generally curious about how we classify their supply.

So, so, Chris, do your customers, what do they actually say? Do they say, based on your data, I want a blacklist, everything? Or do some of them... have more subtle, subtle points of view.

There's an enormous range. There are DSPs we work with who make it literally impossible to buy MFA based on our data. There are DSPs we work with who turn it off by default, but allow buyers to opt in. There's DSPs we work with who do the opposite, have it on by default, but make it easy for buyers to block it. There's DSPs who take our data and just are, it's too early, they're just using it to audit. Like, am I spending 1 percent of my budget, 10%, 50 percent of my budget on MFA? So there's, there's a huge range. I think this has gone from being like a slow build over many years doing research to all of a sudden an emergency. And I think everybody in the industry, buy side and sell side, is sort of like struggling to move fast enough to sort of like, address a problem that has a lot of spotlight on it, but, but, but address it responsibly, you know, I mean, the conversation, the points, Eric, that you were making at the very beginning, I think are very valid, and like, I try hard to like, listen to the market, like, we don't have, this isn't fully baked, we're like, figuring this out as we go, and I think a lot of DSPs and SSPs, and especially agencies, are trying to take a measured approach, saying like, Of course we want to look into this, but we don't want to make, like, a snap decision and do something that's incredibly aggressive that we're then going to regret. Like, we have to be measured in how we're going to address this, and we have to constantly be listening to feedback from the market.

And we've talked about this on a previous episode, but effectively publishers, mainstream publishers cause this problem in some sense, because they run the content recommendation boxes from companies like Outbrain and Taboola, which are the source of the paid traffic to those sites that then compete against them.

And to 15 percent of traffic to MFA publishers comes from premium publishers. The huge super majority of traffic to MFA publishers comes from Facebook.

Interesting. So that's paid advertising on, on,

Yeah, it's clickbait ads running in, you know, clickbait ads can run in content recommendations which is at the bottom of CNN, clickbait ads can run in the newsfeed on Facebook. Both of these, both CNN and Facebook are sending traffic to MFA publishers, CNN's a drop in the bucket.

got it. And Facebooks are presumably better targeted because they're good at that. So one thing that came up in this discussion was the CNAME subdomain trick. I love a good trick. So what's going on with the CNAME subdomain trick for those of us who aren't on top of it?

Yeah, I don't know that I can speak much to the technical details, but I can explain the business backdrop here. Brands and agencies and DSPs and to some degree SSPs are blocking MFA domains. Those MFA companies can then essentially rent a subdomain from a reputable publisher that isn't on a block list. And so, you can move your MFA business. To operate on someone else's domain and so MFA publisher goes to some premium publisher and says hey, let me Get a subdomain from you I'm gonna drive traffic to it. I'm gonna populate it with content. I'm gonna monetize it all and we're gonna split the revenue and Block lists Generally speaking don't work that well at the subdomain level And so now you got to make a business call of like well if I'm dealing with a premium publisher and for 30 percent of their traffic is this subdomain that's ARBed. Either I block the whole domain, or I open myself up to 30 percent of traffic being MFA.

So are they hoping that people just don't notice that like, you know, there's some subdomain called like content. legitimatesite. com is actually garbage, whereas www. legitimatesite. com is good?

Yeah, I mean, I think that's the business bet. I think premium publishers are increasingly aware that this could turn out to, this could turn around and bite them.

It just seems so, so dirty. I mean, if you were outside of this industry and you heard about, you know, we don't want to name names here, but a legitimate publisher, like one that your mom knows is doing this, it's just pretty dirty stuff.

Yeah, that's my view. I mean, look, I think the other way to think about this though is, you know, I think the MFA companies deserve a lot of culpability here. I think a lot of premium publishers, without realizing what they're signing up for, have signed up for what's a pretty underhanded business practice. And as those companies listen to this podcast, I think they, you know, and there's a lot of, as these companies generally get, More educated. I sus, I'm optimistic that the subdomain loophole starts to fade away because, because it will, it really does put the overall business at

It's also fairly easy to block so if you know about it, you can block it. So this is a great conversation we're going to take a quick break, and then we're going to come back with the news of the week, as well as a new game we want to play, should be fun, called Explain This Press Release. So we'll be right back. All right, we're back. All right, so we're playing a new game. It's called Explain This Press Release, where we're going to read some excerpts from a exciting new development in ad tech, according to the press release, and see if we can understand it at all. The first we're gonna roast my good friend, Jonathan Mendez. So, Jonathan and I have known each other for quite some time. I consider him a friend. And he put out a press release about his new company called Neuralift AI. And the three brains on this podcast are going to see if we can tell what they do. Okay, I'm going to summarize some stuff here. Neuralift is the first applied AI built to optimize consumer conversion rate and KPIs. Okay, that kind of makes sense. It redefines customer intelligence and it's a step change in the use of first party consumer data. Any thoughts so far?

I'm stumped.

All right. Built for marketing leaders and the innovation and analytics teams that support customer growth, Neuralift surfaces actionable insights contextualized within the framework of your business. Eric, would you invest in this? What's going on here?

I feel like I need to recuse myself from this conversation because I talked to Jonathan about it and I can actually explain what the business

No, you don't have to recuse yourself. You'd be helping them out here. What does this mean? What does it mean? Actual insights contextualized within the framework of your business.

They are taking first party data. So, so customer data. They are creating I believe they're sort of like creating unique cohorts of users and then deriving insights from those unique cohorts and then using AI, because this goes a little bit further on in the in the release or the, or the blog posts to drive conversion rates, to create new marketing messages based on what they learned from unique first party audiences.

Okay. I'm going to trust you on that.

It's an application of AI for customer data. I think it's actually super

I'm just gonna read one more sentence because I enjoy the the verbiage here results dwarf pre existing methodologies and superpower marketing operations So that's that's really something. All right, Jonathan. Congrats on your new startup. Sorry for making fun of you a little bit We're gonna go to the next one which we got a hat tip to a friend of the pod, Keith Petrie, who put us onto this one. So this is a a press release from IPG about their RealID solution. And some people online on Twitter tried to make some sense of this. So I'll read some of this, which is you know, IPG, which is a big holding company. Oh, man. As I'm reading this, I can't actually read it because they they put a pop up ad on top of the content. Okay, here we are. I figured out how to hit the cl I figured out how to hit close the button. IPG today reveal unveiled RealID in the cloud, an identity resolution application built by Axiom and designed to seamlessly integrate with a brand secure cloud ecosystem to power growth in a cookie less world. Chris, you wanna give it a shot?

is. But you remember in like 2011 12 when like all the agencies had their own DSPs? But they weren't their own DSPs, they just like, bought a DSP seat and called it their own thing. This sounds a lot like

Right. Yeah, it's like, it's like an ID, but it's, it's like a white labeled ID.

It's an I, it's a trading desk. It's an ID for trading desk. Trading desk IDs.

It's a real ID in the cloud. What does that mean? Aren't all IDs in the cloud? Where else would the ID be? Identity resolution application to seamlessly integrate with a brand secure cloud ecosystem. We're stumped here. I'll just, one more sentence. Real ID enables marketers to recognize the real people behind the scattered data points, swiftly creating accurate, holistic customer views that tap into Axiom's best in class data assets. I'm really, I'm really struggling here. I mean, I drank my coffee this morning and I'm just struggling.

So it's a new, so it's not a new identifier because you know what the industry probably doesn't need is a new identifier. It's a new way of stitching together existing identity identifiers.

Yeah, I think the only thing that's giving me a clue here is where it says when it integrates with the brand secure cloud ecosystem. So it sounds like it's a graph.

identifiers.

I think, you know, I've worked at big companies and a lot of times press releases are entirely meant for morale boosting internally. It's a lot of times it's just like, hey, we did it. It's out. Here's the, here's the link. Okay, we've made fun of this enough. I have no idea what real ID in the cloud means. I have a better idea of what Jonathan Mendez's new startup does. So congrats again, Jonathan, for the startup and sorry to roast you a little bit. Maybe come on, justify your existence. Okay, news of the week. So we have some breaking news. So, this is like the Steele dossier of ad tech at this point. So, there is an anonymous blog post. that was put out that has internal documents from Comcast's Freewheel division, where I used to work that claims that Freewheel is doing something pretty shady where they're increasing bids, they're basically duplicating bids in an algorithmic way meant to Essentially fool DSPs into bidding more on the same inventory. And and it's causing a little bit of a stir in inside circles. I haven't really seen it published that much. But it's there, it's public, and I can send, I can put the link in the show notes. So, Chris, this is actually kind of a common activity in, in display, isn't it?

I can put the... So Chris, this is actually kind of a comment,

it.

There's just a variety of techniques that are all built on this basic learning that like, if a publisher can send lots of duplicate requests to a DSP, it makes more money. And more duplication equals more money. And I think we have sort of tricked ourselves into believing that that's like a display phenomenon. And that for some hard to specify reason, CTV is different. And I don't think CTV is different at all. And I think you see... Publishers working with Spring Serve and Publica so that they can run auctions through a lot of SSPs. I think you see publishers working, CTV publishers working with ad networks so they can deploy resold auctions through a lot of SSPs. This all feels consistent with like that broader theme that like, CTV's not special, it's subject to the same basic market forces, and CTV publishers will make more money if they can send more bid requests to DSPs.

Yeah, so in this instance, what we're talking about is that the SSP gets a add slot. There's an add slot opportunity. And the SSP might send more than one bid request to the exact same DSP for the exact same path, supply path, basically. And maybe they'll change the data a little bit so it's less obvious that it's totally duplicative. Is that your understanding of what's going on

Yeah, so in this instance, what we're talking about is that the SSP SSP DSP with the But working with 25 SSPs and integrating them each through pre bin and TAM and OB is okay. Like those are, those are not different things.

Yeah, I made the analogy. Someone asked me about to explain this to them in plain language. And I said, imagine you're selling like stocks and bonds and you're at Goldman Sachs. You could call up the the buyer at J. P. Morgan and say, I have these bonds for sale. Great. But what if you and three of your colleagues call up different desks at J. P. Morgan at the exact same time and tried to sell them the exact same bond? You might have higher odds of closing a deal. It's not exactly, it's not immoral, but it's not something that is a good trade practice.

I guess I just, it's a bad trade practice for sure. It's broken the basic unit economics of ad tech businesses. I just always sort of find myself puzzled about like, why are we saying certain ways of creating auction duplication? Our normal best practices in the industry and others border on, you know, bad behavior or breaking industry protocols. It's all the same to me.

Yeah, I had the same reaction that that you did are, which was you know, this feels like just sophisticated finance trading strategies brought to programmatic advertising. And the more programmatic gets sophisticated, the more we're going to see things like this, quite frankly, unless there's some sort of agreed upon standards within the industry to not do this type of thing for the benefit of.

Yeah, I think the agreed upon standard in the industry is that the supply team at the trade desk yells at you until you stop. The, that it's in writing, so whoever wrote this anonymous blog post has internal Freewheel documentation explaining how it works, so that kind of shook things up. I reached out to some friends at Freewheel to just give me the informal scoop, and their basic reaction was like, hey Publica and SpringServ do this too. So I'm not sure that's good or bad, but that was part of the reaction I got. That's an informal, non official reaction, I just, like, have friends there. So, and... And we want to thank Publica as usual for being a sponsor of this podcast, but church and state. Okay so other news this week, the New York Times is bringing back Open Programmatic. So after making a pretty big deal out of not doing programmatic, they're bringing back programmatic. I guess you gotta make money sometimes.

To be clear, they're bringing OpenProgrammatic back to in app. So it seems to me that that's probably a really good place for New York Times to be playing in open programmatic, given that it exposes them to budgets that they might not be seeing on their traditional, you know, sort of like core mobile web and desktop business, IE app installs, performance, advertisers, all that type of stuff.

Yeah, monetizing in app for a news site has got to be a nightmare, because in app is so overwhelmingly games that I'm not sure that there's going to be a lot of juice in that squeeze, or squeeze in that juice, whichever it is but, you know they probably are under monetizing pretty significantly in app. But if they put a, if they put a, like, video ad in front of my New York Times crossword, I'm going to walk down to Times Square with a pitchfork like that, well, it's unacceptable. Just saying, I just want to put them on notice that I will, I will become violent if there's an ad in front of my crossword.

thing that you've said this week. The funniest was your reaction to Brian Morrissey's tweet about Advertising Week being, you know, the bane of Adweek's existence. I felt that so deeply.

While we're on that, did you guys do any advertising week activities?

Just some things around it to support portfolio companies and see some friends. Nothing on the main stages.

You're such a good investor, you'll even take one for your portfolio companies during advertising week.

Always.

Last news of the week. So Google launched like 10 years too late CMP, DMP. So you know, I, I will say, I worked at Google from 2008 to 2010, and they were working on the DMP then. So you know, 15 years later, they launched it. So, this appears to be... a method for marketers to upload first party data into Google and then have the, I guess, the matching of that first party data inform the use of a lot of different Google products. So Google ads, the DV360, a bunch of others. So it's, it's not really a DMP. It's more like a, I, I see it as an enhancement to add DataHub, which may be the right way to think about it, I'm not, I'm not really sure either of you guys have a thought on this one?

Well, the release seemed to talk quite a bit about Google's owned and operated properties, which suggests to me that, like, this is sort of like underpinning this whole initiative is a strategy to sort of like squeeze as much juice as they can, to borrow your analogy, from their authenticated user base in a way that is sort of like structurally different from what you could do with sort of like standalone ad tech platforms.

Right, right. They have this enormous consumer base so they can identify users presumably using an email address pretty accurately. And then that data, they don't want to, they don't obviously want that data to leave their, their environment, their ecosystem. But if you're using if you're using one of their app products or analytics products, you can get insights and A marketer could say, Hey, you know, ari at aripapero. com. That is a you know, one of our high valued users. And then you can see whether it was exposed or not in various contexts.

So it's not a customer data platform because it's not connecting to marketing platforms outside of Google. It might be closer to like a purpose built clean room solution because, you know, they're taking customer data, they're taking marketer data and trying to do cool innovative stuff, you know, throughout the Google ecosystem.

Yeah. It's like a clean room, except it's not really clean. It's you put all your data in and it never comes out. It's like, it's like the the storage attic of clean rooms. It's a one way valve, just put more and more data into this room and it never comes out and no one ever sees it except for Google. Sounds great. Okay. On that note, let's, let's call it. This was an amazing episode. Chris, thank you for being here. So stand by for Justify Your Existence with Festra. And and we'll be with you next week as well. Chris, Eric, thank you so much for being here.

Thank you.

Thanks, guys.

Welcome to Marketextures, justify your existence, where we ask early stage ad tech and MarTech startups to tell us why we should care about what they're building today, we have Ashley McKenzie, who is the CEO and founder of Finestra. Did I get that right, Ashley? Yeah, that's bang on. Thanks Eric. All right. Tell us where's the company based, how many employees and what funding have you taken? It's queued in the UK, but we've got sort of customers all over the world. We currently have 10 FTEs and a couple of consultants. To date, we've got about 2 million US dollars of equity funding into the company. Oh, great. What do you do? I've never, I have to admit, I've never heard of you. That's okay. Yeah, we provide, uh, an independent and I think that's really important, independent platform and ecosystem that allows programmatic media buyers to understand through analytics and then automatically Improve their performance in, in their programmatic advertising spend. And we do that through a deep analytics platform. And if you will, marrying log level analytics with optimization and bid modification APIs. Right, so, so you're not like a buying platform yourself, you take the data from the trade desk or other systems like that? Correct, yeah, so we don't intermediate in the buy, we simply ingest data, we get read write access to the DSPs, we ingest data, we analyze it on our side of the fence, if you will, and then we use bid modification APIs to change, change bid pricing, and we do that on a multi dimensional basis, right, so subject to the APIs that the DSPs provide. Right. And how fast is the turnaround time? It's not a real time optimization, right? No, it, you know, the, yeah, well, the cadence depends on how often you get the data, right? So, you know, with the trade desk, they have the Reds feed. So we work with that for some partners. And that's, that's a stream data set. That's, that's, that's getting aggregated and batched on our side. But, you know, data transfer files from DV360, you know, these come out every 24 hours. So we'll resolve. Uh, however, however, often we get fresh data to, to analyze and what's special about your optimization. Are you working like some of your competitors say that they work on like more nuanced KPIs or, you know, they really understand what a conversion is better. Let's say I'm just throwing that out as an example. What, what, what's your special, uh, differentiator? Yeah, you know, I guess we go over way on CPAs and performance and conversion So we're basically putting incremental effort into optimization that human beings can't do and the DSP bidders aren't doing because they've got so many other things They've got to take care of we'll take sales data customer lifetime value data From up from our partners from our advertisers and we can pull those into our algorithms as well But really, you know In terms of optimization, we're often just simply doing things that human beings just cannot do, and then taking it to the next level. Cool. And do you have any customers you can name? Any case studies? For sure. Yeah. So we work in the States. We're with a fantastic agency called Media by Mother. And their clients, Amica Mutual Insurance, Premier Bank. We work with Kindred in various parts of the world. They're a big gaming company called Unibet. Unibet's their big brand, but all of their brands. We work with, commonly we work with Deutsche Telekom, various subsidiaries in Europe as well. Great. And what's your pricing model? We, we, well, CPM. And then the optimization, we're, we're effectively going from value discovery to value delivery, right? We're actually enacting these optimizations. We charge that on a percentage of media and we can do various fixed pricing as well. All right, great. Well, I think you've done a pretty good job justifying your existence. One last question. If your company was an animal, what animal would it be? I'm going to shoot for an owl. Owl. Okay. Why is that? Yeah, I think we're, we're pretty observant. We're very deliberate in what we do and and we're generally relative to this industry quite a quiet organization All right, very nice. Well ashley. Thank you so much for being here. You're very welcome. Ari. Thanks so much for the time