Latest / Signal & Noise: Bringing Clarity to Media, Data, Tech and AI / The Cost of Keeping Quiet: Ad Fraud, Incentives & The Silence Protecting Billions, with Sarah Caputo & David Nyurenberg
Transcript
- 0:13Welcome everybody to another exciting episode of Signal and
- 0:16Noise. We are a no BS media franchise
- 0:21bringing clarity to data, tech and AI.
- 0:23And as usual, I'm joined by my Co host Ria Longacre.
- 0:27I'm Brett House. We've got some very exciting
- 0:29guests today, David Nierenberg and Sarah Caputo.
- 0:33Did I say that right? You did, yeah.
- 0:35I did great, great. The topic before I introduce
- 0:39David and Sarah, which is the first time we've had two guests
- 0:44at once, which is should be pretty interesting is the cost
- 0:48of keeping quiet media responsibility and the
- 0:51incentives that enforce silence. Now that's a juicy topic.
- 0:55So as you can imagine, this episode should should uncover a
- 0:59lot about ad fraud and related topics.
- 1:03But but the difference things between this and other episodes
- 1:06is that David and Sarah have written a really compelling
- 1:10article which should be available in this.
- 1:11This episode launches about their first hand experience sort
- 1:16of in the dark underbelly of adtech and mar tech, right?
- 1:20Seeing how incentives impact and affect people's decisions when
- 1:25it comes to media quality, ad fraud, and what you're
- 1:30representing to brands and how brands are thinking about a lot
- 1:34of these topics. So before we go any deeper into
- 1:37the theme, which we'll reel, we'll introduce in a second, I
- 1:39want to give Sarah and David an opportunity to interest
- 1:43themselves for those that don't know them, rather than me doing
- 1:46the introductions. David, why don't you start and
- 1:49then Sarah follow up. Yeah, absolutely.
- 1:51And by the way, Brett, kudos to getting my name right.
- 1:55That's usually a challenge for many.
- 1:58So I'm David Nierenberg. I am a performance marketer for
- 2:01about 15 years now. Started out on the publisher
- 2:04side, then bounced around to a few ad tech companies, then was
- 2:08on the holding company side, then had my own consultancy for
- 2:11four years. And then after that, I've been
- 2:14deep in the CTV space for the past five years working and
- 2:18leading up CTV innovation strategy of multiple agencies.
- 2:21Currently I'm the SVP of Digital at Intermedia Advertising.
- 2:25We're a 50 year old linear TV agency out of LA.
- 2:29We're on tasked with building up our digital and CTV capabilities
- 2:33and bringing over many of these legacy linear Dr. advertisers
- 2:37into the performance CTV space. So I've been naturally
- 2:42challenged with figuring out all the challenges in the CTV space,
- 2:46route transparency and driving performance and measurements.
- 2:48So that's where I've been having lots of fun in in the past five
- 2:52years. Sarah.
- 2:56Yeah, Thank you guys. Hard to follow David, He's very
- 3:00good and well known apparently as well.
- 3:03I have been in the industry 21 years.
- 3:06I always joke that my career is old enough to drink legally now
- 3:11and I have been in basically every role.
- 3:14But I did start at agencies and I actually started it on the
- 3:18brand side. I did a lot of brand positioning
- 3:21work, creative and development for large CPG companies,
- 3:26Hospitality. I moved into Allstate on the
- 3:29brand side as my inroad into enterprise.
- 3:32I was there eight years and moved around building
- 3:35capabilities from one department to another, ending up in housing
- 3:40Allstate's media buying program from a managed service model,
- 3:43oversaw programmatic, social and the tech stack.
- 3:46So quite a bit happened there and now I'm on my own as a
- 3:50consultant and advisor. I've been teaching a lot of
- 3:54workshops with Amazon, Nielsen and Viant, Roku cars, Walmart's
- 3:59coming up for their ad platform on attribution and housing, you
- 4:03name it, DSP competitives there. But I also flip on the brand hat
- 4:09and work enterprise side to try to help them with their
- 4:11ultimately their data and text act to ensure that they're going
- 4:14to be able to go to market adjacent to the goals that they
- 4:19have. Yeah.
- 4:21Well, and kudos, well, kudos to Sarah for saying 21 years
- 4:24because most of us stop at 20. And we talked about in our last
- 4:29episode, it's kind of. Scary, isn't it?
- 4:31Yeah, I'm like, I don't think I'm ever going to say 30.
- 4:33So I'm going to say 20 until I retire at like, you know, 7080
- 4:37years old. I'll be like 20 years in the ad
- 4:39tech Mar tech world. But anyways, I like, I like your
- 4:43story about 21 and in the drinking age that's.
- 4:46A point once it's holding up the drink.
- 4:48I will not say how many years it's been.
- 4:52So we love it. Awesome.
- 4:53Well, I was great intro for the two of you and this episode, the
- 4:57cost of being quiet. I guess the genesis of this
- 4:59Brett was we all went to dinner. It was a bigger group and Sarah
- 5:03did sit at the side of the table at the opposite side of the
- 5:05table. So we didn't get to engage with
- 5:07her as much as we probably would have liked to.
- 5:08But we all started talking about this this concept and then David
- 5:12and Sarah wrote this amazing article, which I encourage
- 5:14everyone to read cause to keep me quiet.
- 5:17And that's really where this episode came out of.
- 5:19And this, what I really liked about it is this, if anyone
- 5:23who's worked in the media, in the media space knows that it's
- 5:27kind of taps into something that everyone knows, but no one
- 5:29really says out loud. Namely, the system is broken
- 5:33because silence is not only encouraged, but it's kind of
- 5:38rewarded, right? Like we spend massive dollars as
- 5:40an industry and a lot of that value just gets leaked out
- 5:44through waste, low quality inventory, MFA, bad incentives.
- 5:48We all know the, you know the saying, show me the incentives.
- 5:51I'll show you the outcome as well.
- 5:53There's a lot of really like poorly aligned incentives, let's
- 5:56just say in the industry. And the core issue is that like,
- 5:59it's not that fraud exists. We all know it is.
- 6:01I mean, what you know, we know it's not 1%, We know it's a lot
- 6:03higher than that, right? We know that, you know, it could
- 6:06be 25, could be 30, could be more than that, right?
- 6:08Depending who you ask, who you ask, the Foo, he'll tell you
- 6:10it's 8090%, right? But there's a lot of fraud out
- 6:12there. We know it exists and somehow
- 6:15the market structure tolerates it, mostly because so many
- 6:19people benefit from the spend continuing, continuing to flow,
- 6:23right? The more that's spent, everyone
- 6:25is taking, you know, the way ad techs typically prices a take
- 6:28rate, right? There's ADSP take a percentage,
- 6:30SSP take a percentage. All these other vendors take a
- 6:32percentage. And so there's their incentives
- 6:36to keep the spend high and there's a disincentive to really
- 6:39call out the fraud that we really all know is there.
- 6:43The buy side, you would think is the last line of defense for for
- 6:47the market or protecting their budget, right?
- 6:49And it's only if buyers create like demand transparency, insist
- 6:54on control when they observe dark patterns and bad behavior,
- 6:58calling that out, you think they would be doing that.
- 7:01But because of these incentives, it's very, very difficult to
- 7:04solve. As the Foo himself put it, ad
- 7:06fraud is an incentives problem. In this episode, we helped to
- 7:10discuss that and see how we can maybe talk about breaking that
- 7:13loop. So with that said, we know ad
- 7:16fraud is an incentives problem. We know it's not 1%.
- 7:19We know it's more than that. Brett, first questions over to
- 7:21you. All right, so, yeah, so I'll
- 7:23start with, you know, you guys have some very interesting
- 7:26stories to tell from your own experiences.
- 7:28And that's really what the article is about.
- 7:30And the article, by the way, for the audience is on Signal and
- 7:32noise dot AI, which at the point of this episode being published
- 7:36should be live. It's not live yet, but it should
- 7:39be live when this this episode goes.
- 7:40But so, so why does why do you think?
- 7:43And I'll start with you, David, and then we'll go to Sarah.
- 7:44Why do you think the system keeps rewarding silence and what
- 7:48we talked a lot about incentives and misaligned incentives, but
- 7:52tell us from your first hand account what you've experienced
- 7:56and what you really think the driving, the driving.
- 7:59Behavior is, yeah, well, I, I think there's multiple factors
- 8:02here, one of which being that just the financial models on top
- 8:06of which this industry runs, right.
- 8:09In many cases, most media, but most vendors agencies are taking
- 8:13some type of margin, right. Sometimes it's transparent flat
- 8:16margin. Many media vendors are operating
- 8:19on non transparent margins, right.
- 8:21So they're depending on the arbitrage model, right?
- 8:25And you know, in a world where there are multiple mouths to
- 8:29feed, right, digital introduced this ecosystem of tech players
- 8:34that are all trying to extract value out of the media dollar,
- 8:38where in the past, you know, you would have the agency and you
- 8:39would have the publisher, right? And the agency would pay the
- 8:42publisher. And there was a very direct, non
- 8:45complicated pit line of payment. Now you have ADSP, you have an
- 8:49SSP, you have the publisher, you have the ad server, you have the
- 8:53measurement partner. Everyone is trying to extract
- 8:57some value out of the CPM. While at the end of the day, the
- 9:01publisher still needs to get paid and still needs to make
- 9:03enough enough money for to sustain themselves, right?
- 9:07So, and then at the end of the day, also the media buyer and
- 9:10the agency also needs to be able to pay their staff and get some
- 9:14money out at the same time, right?
- 9:16So with all these different mouths to feed, you know, you,
- 9:19you're going to need scale, right?
- 9:21You're going to depend on moving as much money as possible to
- 9:23what Rio said. So you're not going to want to
- 9:27kind of kind of do anything possible to hurt scale and raise
- 9:32questions about where money's going into, right?
- 9:35You're going to want to just keep keep everything just
- 9:37flowing nice, nice, nice and happily between all sides.
- 9:41Yeah. And when and when you say, when
- 9:42you say you in this in this topic, are you talking about the
- 9:46brand, the buy side, the brand and their agency representative?
- 9:49Or is that more of an agency problem where it's like use it
- 9:53or lose it type of issue, right, where they're getting budget,
- 9:56they're required to spend that budget and then they.
- 10:00Get paid out of cut of that budget, right?
- 10:02Yeah, and they can. And they can throw out some
- 10:04vanity metrics, right? Some reach metrics and R and row
- 10:09as metrics that don't necessarily prove definitively
- 10:12outcomes and yet and yet continue to defend their their
- 10:15budgets, right? And what do you think?
- 10:18I I think it's very convenient. We love pointing fingers at the,
- 10:22the, all the tech companies in the space and agencies, But in
- 10:27many ways, brands are kind of responsible for this world that
- 10:30we live in these days, right? They keep pushing their agencies
- 10:34and partners down on margin to the point where it's, it's not
- 10:37sustainable. They view media as all
- 10:40commoditized, right? So they just want the lowest
- 10:42possible CPMS, right? And naturally, when you go after
- 10:45the lowest possible CPMS, you're going to have to sacrifice
- 10:48something there. And often that is media quality.
- 10:50But the same time brands don't want to know that they're buying
- 10:53poor quality media or, or MFA or anything, something like that.
- 10:56So yes, the, the tech companies, the some agencies vendors, they
- 11:05might not be they, they might not be operating above board and
- 11:09trying to keep things under, under the water, so to speak.
- 11:12But at the same time, they're kind of having to deal with the
- 11:15hands that the brands are dealing them, which is lower
- 11:18CPMS, lower margins. And naturally, you know, people
- 11:22need to get paid and companies need to keep growing revenue,
- 11:25right? So something has to give there.
- 11:28It's funny you mention that I met.
- 11:29I once had a client, we're pitching him and then they told
- 11:32us have with you to pitch. So we've decided to outsource it
- 11:34to India. This is like, this is by media.
- 11:39This isn't like you're, I don't know, this isn't like business
- 11:42process outsourcing, right? This is like, you know, I just
- 11:45thought it was crazy, right? And they said, oh, they'll do
- 11:47for half the price. So crazy stuff.
- 11:50So Sarah, I mean, love to get your take.
- 11:52I mean, I loved in the article you talked about your experience
- 11:54of one of the big hold codes about you receiving some
- 11:58backlash for actually questioning some of these
- 12:00things. Maybe you can start to dig into
- 12:02that a little bit. Yeah, I think you guys have
- 12:05covered quite a bit of ground and I'll go, I'll go further
- 12:09upstream since David has already hit on quite a few points.
- 12:14They don't. I think we've made a problem
- 12:16with programmatic that it's more science than art.
- 12:19I think it's problematic for a few reasons.
- 12:21One, the creative is just getting lost in translation and
- 12:24we're over indexing on position and science.
- 12:27It's just not the reality nor is it logical from a marketing
- 12:30perspective. But the second part of that is,
- 12:33is you are putting pricing pressure way upstream,
- 12:37especially in the larger companies with procurement
- 12:39involved. We're looking at it as a cost,
- 12:44you know, situation, like you said, front end CPMS versus an
- 12:48outcome situation and they don't understand the complexity of
- 12:52what's being bought. So Rio, to your point around,
- 12:54you know, outsourcing to India, not saying there's anything bad
- 12:57with certain skills and capabilities, but usually you
- 13:00want to sit close to those partners and relationships
- 13:04because when you're spending hundreds of millions of dollars
- 13:06and your business is relying on those outcomes, sitting closer
- 13:10to those teams and proximity and security I think is really,
- 13:15really important. So I think we're looking at the
- 13:17wrong things from a brand perspective.
- 13:19And then the second thing I'll say is structurally, brands are
- 13:22not built for this sort of complexity.
- 13:26And what I mean by that is if you're really doing the right
- 13:29thing on the brand side, you're focused on product, customer
- 13:31experience and user value and also driving those revenue
- 13:35outcomes. So when you get stuck into the
- 13:37nuances of having to triple track and go under every single
- 13:43nook and cranny that the industry has made so incredibly
- 13:47complex without a doctorate degree, it has created a problem
- 13:53and pressure on the brand side where it's difficult also to
- 13:56keep up. So it's just kind of full circle
- 13:58to be honest, on all sides. And brands aren't usually
- 14:01reputationally allowed to talk about their partners or where
- 14:05they may have gone sideways. So it becomes a reputational
- 14:08problem to boot. So it's interesting.
- 14:10So David, on one hand, you're saying like it does go up, that
- 14:12should be the brand's ultimate. They should be accountable for
- 14:15this, let's say, right? But you know, sorry, you're
- 14:17saying that it's super complex. Maybe it's unfair to expect the
- 14:19brand to know it. And we had Foo in this pod a few
- 14:22months ago and he even said he doesn't even think programmatic
- 14:25should be handled by he says he thinks he should be handled by
- 14:28brands, which is, which is an interesting argument because he
- 14:30said really, it's an auction, right?
- 14:32Why would you need someone doing it for you?
- 14:33So those are all interesting. So what do you think though?
- 14:36Like, who's really does the buck stop at the brand?
- 14:39Should it, should we expect them to be better informed about
- 14:42like, you know, should they know, for example, that more
- 14:45than $0.50 out of the dollar in open exchange is not going to
- 14:47the publisher, It's being scraped off by all these
- 14:49intermediaries? Like love your thoughts on that.
- 14:51Yeah, I think procurement who is not technically savvy and
- 14:58doesn't really understand what's happening in the in the industry
- 15:02really has an outsized role in all this where I can't tell you
- 15:07how many pitchers I've had where, you know, come in with a
- 15:10great media strategy, great great research, great pitch
- 15:14brand, loves it. And then at the end of the day,
- 15:16it still comes down to just procurement asking for your CPMS
- 15:19and they're just comparing your CPMS to the other agencies.
- 15:22And that's where the final decision or a big part of the
- 15:25decision ultimately comes from. So to Sarah's point, we've made
- 15:29this. More, it's a race to the bottom
- 15:31right it. Is right.
- 15:32So Sarah said we it's more of a science.
- 15:35We've kind of made this more of a science and less of an art
- 15:37where advertising historically has been both part science and
- 15:40part art, right? And the art has part of it has
- 15:43been like the psychological aspect of of advertising.
- 15:46How do you influence people? What are the best channels to
- 15:48influence your target audience? What unique out-of-the-box ideas
- 15:52do you have to stand out from the crowd that aren't all gets
- 15:56lost at the end of the day when what the brand really wants to
- 16:00know is how much are you getting this publisher for versus the,
- 16:04the, the competitor agency, how much you're getting that same
- 16:06publisher for? What are your rates on the DSP?
- 16:08Right. All these pricing factors
- 16:11completely disregard ultimately what really makes advertising
- 16:14special and what many people who work at advertising got into
- 16:17this industry for. Has it always been that way or
- 16:20is it has it intensified in recent years like the
- 16:22procurement driven, let's say like as Brett, you said, the
- 16:24race to the bottom in terms of just trying to find the lowest
- 16:27CPMS? From what I understand, once
- 16:31McKinsey and Deloitte and all those consultancies started
- 16:35getting into into the advertising space and working
- 16:39with the brands and really influencing procurement, I think
- 16:42that was in the early 2000 tens around or about around that
- 16:46time. That's when things really
- 16:47started started being more price focused.
- 16:49Yeah, consultants did invent procurement just yeah, I confirm
- 16:53that I didn't, but the consulting did.
- 16:56Yeah, yeah. Well, it's an interesting point
- 16:58about you guys are saying science, but but I think there's
- 17:02something profoundly unscientific about like media
- 17:06ineffectiveness and the fact that that brands may not have a
- 17:10full grasp or the, or even the agencies of how much of their
- 17:14actual media investment is actually being seen by end users
- 17:18and how much is actually going down.
- 17:20The the either it's an impression that's being
- 17:23triggered but not shown like in the case of CTV or it's a, an ad
- 17:28fraud play where it's going to an MFA based site.
- 17:32To me, if you're scientific about this and you're, and
- 17:35you're really focusing on outcomes, right, shouldn't you
- 17:39be able to connect the dots and understand the inefficiency
- 17:41built into your total impression served your total dollar spent
- 17:46and, and go to cut that out? And it wouldn't that be the more
- 17:48scientific approach if it's really about driving
- 17:51performance? And is that even possible when
- 17:55you're, when you're in a world where you're, you know, where
- 17:57you're being forced down the sort of, you know, the, the,
- 18:01the, the cheaper approach to, to buying media.
- 18:05Well, I'd say try explaining that to procurement teams.
- 18:08I, I love procurement by the way.
- 18:09I worked really well with them at Allstate, and I'm not just
- 18:12saying that they definitely understood where there was
- 18:16expertise to lean on. Some of your best friends work
- 18:18in procurement, right? Yeah, I love Marie Marie Colby.
- 18:22I'm going to shout out there's a couple others.
- 18:23They were really amazing at allowing the expertise.
- 18:26I'm just giving you a hard time. Like I have to, she's retired
- 18:30anyway, but I think it's less about it's more about the
- 18:35science is more about what people feel can be definitively
- 18:38qualified quantified and comparing the apples to apples.
- 18:43And that is the common denominator that they come up
- 18:45with versus the art part versus that science quantifiable part
- 18:51is actually about logic and intuition.
- 18:53And somewhere along the way, we have forgotten that we all have
- 18:58instincts and you don't need to quantify everything to the NTH
- 19:01of the degree and at 100% definitive and deterministic in
- 19:05our language in order to make a logical business decision.
- 19:09We were able to do this before the Internet really, really
- 19:12well. And we lost sight of that
- 19:14because everything needs to be put on a deck on, on a
- 19:16stakeholder report or on a QBR where everything is, you know,
- 19:21to the dollar. And, and so this is where I
- 19:23always guide brands to take a step back and to really thorough
- 19:27outside platform and inside platform testing to kind of show
- 19:32what trends are over time and versus relying on every single
- 19:36penny because it's impossible 1-2 it's quite inefficient to do
- 19:40so. So I think it's pivoting to more
- 19:42of the testing methodologies as a gold standard in order to come
- 19:45to those conclusions and making sure you have that testing in
- 19:48place before you assign your life away, you know, to one
- 19:52partner or a set of partners could go in a crawl, rock, walk,
- 19:56run and and test into these things in a way that's smart and
- 19:59logical. And when you test into these
- 20:01things, you mean you could test, for example, a top of funnel
- 20:04brand awareness play that might have a Halo effect over your
- 20:08more direct response lower funnel advertising approaches,
- 20:12right. Would you say test those things
- 20:14to see the, the, you know, with the control right, exposed
- 20:18versus unexposed to see how those perform so people can be
- 20:21more convinced? I mean, I saw this a lot in the
- 20:23financial services industry where financial services 70 to
- 20:2680%, this was JP Morgan, Bank of America, etcetera spent 8070 to
- 20:3180% of their total budget was spent on direct mail marketing
- 20:34for the most part direct mail and direct response digital
- 20:37extensions. And there was, there tended to
- 20:41be, and I've talked about this before in the pod, they, they
- 20:45tended to look negatively at the brand spin that was being spent
- 20:49on bigger, broader Reach channel, CTV, television,
- 20:53etcetera. Because they couldn't attribute
- 20:56the Halo effect as cleanly as they would have liked versus
- 21:01what they were doing in direct mail marketing and indirect
- 21:03response marketing, which was much closer to sort of an action
- 21:07conversion event. Physical address based on your
- 21:10customer file if you have accuracy there and offline wise,
- 21:14I get it. But you're right, it's the
- 21:16equivalent to me of last touch attribution.
- 21:19Just because someone had an impression and a piece of mail.
- 21:22You have to look at everything as an impression.
- 21:24I look at a billboard the same way.
- 21:25Everything is an impression to the consumer's eyes.
- 21:28It's just eyeballs that you want and attention over time.
- 21:31So it's exactly right. The physical address combination
- 21:35is a one to one match, so you'd kind of tend to default toward
- 21:38things, even though if you took a step back your logic and
- 21:41intuition would probably say, hey, maybe we're over indexing.
- 21:44Yeah. And if you scientifically look
- 21:46at it, yeah, for like we would do these pretty advanced models
- 21:49that were sort of a they're MTA, but they're a combination of MTA
- 21:52and MN kind of A and, and what they proved was that the Halo
- 21:56effect actually had a direct impact on the lower funnel
- 21:59performance, right. And, and when you compared
- 22:01apples to apples between a sort of an 8020 split between lower
- 22:05and upper and like a 5050 split, the 5050 split one every time,
- 22:11you know, and we would simulate these programs with, with, and
- 22:15we had like 10 years of historical data.
- 22:18So it wasn't like we're, we're operating with a, with a small
- 22:20data set, but it's hard to prove because it's not tangible.
- 22:24You can't really show it in a spreadsheet.
- 22:26And people tend to trust what they see is that last click or
- 22:29that that final action committed by the consumer, right?
- 22:33Exactly. We talk about a lot of this
- 22:35being the age of outcomes, right?
- 22:38Sorry, David, I cut you off. We talk about this being the age
- 22:40of outcomes, right? And you know outcomes.
- 22:42You know, Brett, you just mentioned all the different ways
- 22:44we measure things, the last click being one.
- 22:46And you know, and I think you can make the argument these are
- 22:48just vanity metrics and they're widely abused.
- 22:50So David, let maybe start with you love to hear like how can
- 22:53better like better metrics, better understanding of what
- 22:57they mean? How could that maybe drive like
- 23:00likes, better targeted media, better planning and maybe help
- 23:03those brands be better informed and like how what they should be
- 23:06looking at in terms of how they value its success?
- 23:09Yeah, absolutely. Well, I think one, it starts
- 23:11with understanding that not every impression is created
- 23:13equal. We've really commoditized media
- 23:16to the point where, you know, a reach number on display for many
- 23:20people will will almost be similar to a reach number on
- 23:24CTV, right. And because display's so much
- 23:27cheaper, they'll be like, all right, let's let's put more
- 23:29money into the display than it should probably get.
- 23:32But ultimately, you really need to have an understanding of the
- 23:36dynamics of every single channel and the various pitfalls
- 23:39involved and how things are gained.
- 23:41Because the thing about digital is you could spin up an almost
- 23:45unlimited amount of impressions and there's a giant ecosystem of
- 23:49bad actors who are constantly evolving because it's just and
- 23:54finding ways to extract more and more value out of the ecosystem
- 23:57because it's just so easy to do right.
- 23:59So like understanding those pitfalls, valuing channels
- 24:03accordingly, and also knowing that ultimately what you get,
- 24:06what you pay for, right? So if you're going to pay for
- 24:08cheap impressions, those impressions themselves, maybe
- 24:10there's some value there, but it's not going to be the type of
- 24:13value that you might, you might think that reach number probably
- 24:17isn't, isn't as real as you want, want it to be, right when
- 24:20when you're looking at it. So, and you also need patience,
- 24:24right? And set, you need to set
- 24:26expectations with, with your, with clients too.
- 24:29Advertising has always been a law, a long game, right?
- 24:33How many people watch an ad on their TV and automatically
- 24:35convert when they see that ad? No matter how much we want to
- 24:39push this performance CTV narrative, I think it's
- 24:41ultimately hurting us as an industry because it's creating
- 24:44these false expectations with clients that yeah, if you put in
- 24:47AQR code and creative, people are going to scan the creative
- 24:50and they're going to convert right Now those, those units
- 24:54often come at pretty expensive premiums, right?
- 24:56So the brand is going to pay that expensive premium thinking
- 24:59they're going to drive lots of outcomes.
- 25:01Then they're going to get their reporting back and see that not
- 25:04many people actually scan that QR code and then they're going
- 25:06to be turned off from CTV for the immediate future.
- 25:10So I, I think for us, right, you, you do need to measure
- 25:14things in, in multiple ways, right?
- 25:17The last touch, the direct response actions many times,
- 25:21especially in CTV are more directional because there's so
- 25:24many different measurement gaps and signal loss happening in
- 25:26the, in the industry. So, so there's that.
- 25:31And then there's the long term approaches via MMMS and market
- 25:36match market tests and all, all those type of more extensive
- 25:39testing, right? It's easy for us to look at the
- 25:43last touch model and just depend on that because it's fast,
- 25:47immediate, it often times it gives us the metrics that we
- 25:50want to see. But if you if you put all your
- 25:53eggs in that basket long term, you're going to see diminishing
- 25:56returns and you're going to go into what many DTC brands have
- 26:00experienced over the past several years after the DTC hype
- 26:03cycle died out, which is what I call the the the death loop,
- 26:08where they're just constantly chasing those direct response
- 26:10metrics. They look might look good, but
- 26:13they're not building long term brand awareness and they're
- 26:15ultimately probably converting against people that were going
- 26:17to convert regardless and wasting lots of culture.
- 26:20A lot of those DCC brands are now gone, which is which is
- 26:22interesting or they or they've scaled back significantly.
- 26:24You mentioned CTVII saw you posting about this a lot and I
- 26:28really like this and love to maybe get your thought before,
- 26:31before I before we hear from Sarah on the vanity metrics
- 26:33thing. So you admit like legit CTV in
- 26:37stream high quality inventory, right versus crap, you know,
- 26:42let's call it bogus CTV where it's out stream, let's say a lot
- 26:47of times below the fold even right that you know, they might
- 26:49be lower CPM. So I think I think a lot of
- 26:52clients are confused about that. Is that safe to say about like
- 26:55what is CTV? Well.
- 27:00The thing is that I think we've kind of overstated how much
- 27:04available CTV inventory there is, right?
- 27:08We love talking about how many people are watching CTV right?
- 27:10But I rarely see people go into how many people are actually
- 27:14watching ad supported CTV compared to the entire CTV
- 27:18universe which also includes non ad subscriptions.
- 27:20Yeah, which would be the fast channels, right?
- 27:23Right, right. So like to me like we still
- 27:25haven't really like we like to think that FAST are lower
- 27:28quality for publishers, right? But we also haven't really
- 27:31defined what quality in the CTV space is, right.
- 27:34So for me it's like ad load, right.
- 27:37We really need to start tracking ad load against publishers,
- 27:39right? Because if there's five ads
- 27:42every two minutes in a CTV stream, that's going to piss
- 27:45people off. And they're also not going to
- 27:46necessarily be put put as much attention into the ads as they
- 27:49might otherwise if there was only like one or two ads every
- 27:52so often, right? So that's like a proxy metric
- 27:54for me that we need to start looking at for CTV that
- 27:56currently we're not doing. Additionally, there's a lot of
- 28:00actors in the space because CTV is so expensive because there's
- 28:04not that much as much inventory as we'd like to think.
- 28:07They're mixing an online video and to your point, those kind of
- 28:10those out stream units, which is like the bulk of what's out
- 28:12there in the online video space and and is.
- 28:16That being labeled as CTV within the buying platforms in some
- 28:20cases where it's off the television glass.
- 28:23Sorry you're laughing. So I think.
- 28:27Because I think David has an opinion based on a lot read.
- 28:31I'm like anticipating how he's going to phrase this.
- 28:34Yeah, because because do you tie, do you tie CTV to the
- 28:37device which would be the television glass, which is your
- 28:39sort of sight sounded motion experience that's different than
- 28:42any other digital? Device.
- 28:44Ideas, right. Or do you tie CTV to the type of
- 28:47content, right. So you could argue that that CTV
- 28:51content, YouTube might be an exception to this, but is
- 28:54generally longer form, it's generally more narrative driven,
- 28:58whereas the stuff that's off of the television glass tends to be
- 29:01shorter form, less narrative driven like the TikTok feeds and
- 29:06such. But how do you, how do you like,
- 29:08is it a content thing? Is it a device thing?
- 29:10You're talking about the ads being served, but what is CTV?
- 29:16Well, you know, there's OTT, which is streaming, streaming
- 29:20content over an Internet connection, right?
- 29:22And then there's CTV for me at least where you're actually
- 29:25watching that OTT content on ATV screen.
- 29:28Now a lot of buyers will be buying what they think is CTV,
- 29:32but ultimately they're getting who they're also buying Hulu,
- 29:36which is being watched on people's iPads or on their
- 29:38phones or on their computer devices, which to me that's OK.
- 29:43But like the fact that CTV, like the OTT streaming space is
- 29:48highly non transparent. There's not a lot of signals
- 29:50being shared in the metadata at the same time, right?
- 29:55You're not really seeing what you're delivering against.
- 29:57It's all blinded in many ways as like 1 bundle package that
- 30:02leaves a lot of space for bad actors to slide in.
- 30:06In addition to legitimate CTV, right?
- 30:10They're also sliding on desktop and mobile devices, media
- 30:14running on their own and operated sites, for instance,
- 30:16right? And because it's all bundled
- 30:18under one business name, right? You really can't tell, OK, is
- 30:24this desktop mobile tablet impression actually in the
- 30:27streaming app or is it an out stream impression on this
- 30:32publishers own and operated website, right.
- 30:35And there's a good amount of that happening, which I think is
- 30:37diluting the quality of CTV these days and further not
- 30:41helping things. On the flip side of that, those
- 30:45desktop mobile impressions are also a bit more trackable,
- 30:50right? They're often they're in a
- 30:53cookie based environment. So that helps on the direct
- 30:58response side of things too. So one of the reasons vendors,
- 31:03publishers might be mixing in those those types of online
- 31:08video units is because it helps them.
- 31:10Not only does it help on the CPM side of things because they're
- 31:13naturally cheaper, but it also helps on the direct response
- 31:16performance side of things because it's in a more
- 31:18measurable signal rich environment where often times if
- 31:22it's a downstream unit, someone went to that site, they probably
- 31:25didn't even see it. They didn't even hear it.
- 31:26Shout out to Eris Levin. I know this has been his crusade
- 31:29for for a while, but it makes the numbers look better and the
- 31:34media buyers are none the wiser. Sarah.
- 31:39I mean, I'm going to just go philosophical because I feel
- 31:42like they're found again. We've made such complexity out
- 31:47of everything that we've touched, starting with display
- 31:50ads and now this is like PTSD 2 point O with CTV.
- 31:55Like, yeah, well, you'd go through an agency or through a
- 31:58regional buyer to buy your linear.
- 32:00You'd say what demographic generally you're after and then
- 32:04you'd pay a certain amount, you reconcile it on the back end,
- 32:07bada Bing, bada boom, we're good.
- 32:09And now they have made it so complex, this industry in terms
- 32:14of what am I buying, where am I buying, where is it showing up,
- 32:17what's getting a fraudulent clicks, what are the type of
- 32:19media? And ultimately it is so hard to
- 32:24track as a brand. This is where I don't think
- 32:27agencies will die or go away. Like having that advisor is very
- 32:31smart to keep up with this. And to David's points like where
- 32:34to turn and where to look. But I'm a bigger fan of being
- 32:38closer to the inventory, aligning it to the target
- 32:43demographic, looking at your reach that you need to make sure
- 32:46that's really nice and clean. And trying to use some of that
- 32:50logic and intuition we talked about before on where should I
- 32:53show up contextually relative to this audience And use the data
- 32:57where you can overlay testing and go direct to the source as
- 33:02close as possible. And then use bidding
- 33:05environments or auction environments as like remnant or
- 33:08an augment in a secondary strategy, not your leading
- 33:10strategy. And even then try to ask the
- 33:12right questions of where is this, you know, by going?
- 33:15But I love contextual and it's for the same reason you buy news
- 33:19on the morning and linear. So why are we thinking about it
- 33:22differently and programmatic? It should be about your
- 33:26audience, who you're trying to reach and tying logic.
- 33:29Like I said, back to shows programming.
- 33:32I could go on and on on the. Top yeah, Well and aligning your
- 33:35advertising to the medium that it's being served in, right.
- 33:39And your top of funnel brand awareness plays should be going
- 33:44to change to mediums that are reach mediums where there's a
- 33:47ton of engagement, Super Bowl ads, things like that, you know,
- 33:50big screen television, whereas there are things that are more
- 33:54appropriate for direct response, deeper funnel engagement like
- 33:58your e-mail channels in your in your direct mail channels,
- 34:01etcetera. So to me, part of it's, it's,
- 34:03it's an Audi, it's a where are you reaching your audience and
- 34:06then and then appropriately advertised to them in a relevant
- 34:09way within that environment, right?
- 34:11Like that's why I, I I sort of think AQR code on CTV that's a
- 34:17non that's a television screen CTV by seems to me a kind of a
- 34:22miss may work and maybe you guys can prove me wrong, but it seems
- 34:26like it's a, a misapplication of the.
- 34:29I can never see the option. Go ahead, David.
- 34:33Yeah, you've never seen it work. I've never seen it work.
- 34:36Maybe it's nice to have the option, but to me it's like, OK,
- 34:39you're going, you're buying CTV because you want to be on the
- 34:42full screen, You want to get the most attention possible, right?
- 34:45So why are you making your video asset, which is the most
- 34:48powerful form of advertising, in my opinion, smaller, harder to
- 34:52look at, and then also drawing attention away from that asset
- 34:56to AQR code to some words on the in the overlay.
- 35:01And then you're also asking the person to start fumbling for his
- 35:03phone trying to scan. Now he's not hearing the
- 35:06message. So to me, I, I, I think it's
- 35:09again, us trying to force a channel to be something that
- 35:12it's not. Oh, totally, totally.
- 35:15What do you think? Amazon?
- 35:16Amazon has some really like they have those little buttons now,
- 35:19like I'd say it's an ad for those like those are like to
- 35:24say, because in a plus, Amazon knows who you are, right, They
- 35:27know who you are, but because it's within their within their
- 35:30within their own. You know this is content owned
- 35:32and up owned and operated content, right?
- 35:36That they're able to serve up within their own?
- 35:37Charge you more for those types of types of units.
- 35:41They're the same prices on the other unit, which also I like
- 35:47most publishers for like those QR code units, they're asking
- 35:49for pretty high minimums and they not CPMS are higher.
- 35:52So again, that goes back to my earlier point.
- 35:55It's, it's you're trying to, you're creating this perception
- 35:58that humans operate in this psychological way where they're
- 36:02going to scan the QR code and the advertiser should pay more
- 36:06for that. And that's just creating bad
- 36:08expectations that often they'll come to fruition.
- 36:11So the title of your article was, you know, cost of keeping
- 36:14quiet and you know, you had some great examples in there of I
- 36:18mean, we talked about the the perverse incentives right to
- 36:21continue spending more, not calling out maybe some of the
- 36:24shady behavior. I'd love to hear from the two of
- 36:26you for the for the sake of the viewers here.
- 36:29Maybe we can talk about like when you did try to bring up
- 36:32some things you found that didn't look right when you tried
- 36:34to do like David as a media buyer, Sarah, both as an in
- 36:38house planner buyer as well as agencies.
- 36:40I'd love to hear maybe from each one of you that you're an
- 36:44example of like, is there a cost for keeping quiet?
- 36:47Like like what happens when people try to speak out?
- 36:50Maybe Sarah start with you. Yeah.
- 36:53I mean, the examples I have are are there we, you know, at
- 36:57Allstate when I took over the program didn't own the ad
- 37:00account that's in the article of the social media ad account,
- 37:03which is important. So it's run by the agency,
- 37:05right? The agency, yeah.
- 37:08And they recently had gotten acquired at that literal same
- 37:11time that I told them we would in house this social media
- 37:14program by a top consulting firm under their digital arm.
- 37:18And we were definitely a big piece of business.
- 37:22And they would. They refused to hand over that
- 37:24ad account to me. Why would they not do that?
- 37:27Why would they not want to hand it over?
- 37:29Yeah. Aren't they contractually
- 37:30obligated to to you know? No.
- 37:33Like there's terms and conditions and limits with your,
- 37:35your MSA with with these agencies.
- 37:37Right. In terms of technically, no, if
- 37:39your MSA doesn't give you explicit ownership or they're
- 37:42not working out of your AD account.
- 37:44So that's one of the pieces. You have a poorly written MSA,
- 37:47yeah. Yeah, blame the lawyers for this
- 37:49one. They should have had
- 37:50protections. But this?
- 37:52Is going back to procurement not understanding the nuances of why
- 37:55you should work out of one account versus another?
- 37:57Because it all looked like Meta. Facebook was the name at the
- 38:00time. So I actually, it wasn't the
- 38:03consulting firm that did this. It was the contact who lost the
- 38:06business that I was working with, who refused.
- 38:09And the finally I just said screw it, somebody's working
- 38:11with this company because it's a very large firm and another
- 38:15means. And I went through procurement
- 38:17and said, who else owns this contract on the consulting side?
- 38:21And inevitably the first e-mail shocker that I sent to say we're
- 38:26holding up work over there until I get my ad account, They, they
- 38:31released it. So it took me a couple of weeks
- 38:34and back and forth and some contentious conversations, but
- 38:37ultimately I was like, I'm, I'm not getting anywhere.
- 38:40And only reason I was able to talk them into it and to release
- 38:44those keys over to us was because we were a big company.
- 38:48So if you're a mid size company and you don't have that
- 38:52leverage. Or no leverage at all.
- 38:53Then then you're gonna have to get super creative and you might
- 38:57end up paying literal dollars in order to get that power back.
- 39:00I mean, it's, yeah, that is just, it's that's, that's a bad
- 39:03act, right? Or a bad actor.
- 39:04But what? What data, what assets did they
- 39:07actually have that they were holding hostage in a sense?
- 39:10Yeah. So if you think about any time
- 39:12you start a new campaign, you've got a, you've got all the
- 39:14flighting you've got. I have multi millions running in
- 39:17this one platform with all the campaigns which we're relying
- 39:21on. The second.
- 39:22Year, yeah, you've got a 1 1/2 to three-week ramp up time.
- 39:27You've got all the historical data, the connections into, you
- 39:31know, the buying signals. And in order to have this
- 39:33performed, so Plan B would be creating your own account from
- 39:38scratch. And you're going to create a
- 39:39disruption in the business because of the time that it's
- 39:42going to take to create a new account.
- 39:44But also that onboarding time in order to get that campaign off
- 39:47and running and producing the way that you expect so that
- 39:50there isn't disruption. I mean my media, because
- 39:53hopefully it performs and you don't want a four week break in
- 39:56that process. So it was not a great problem.
- 39:59It's interesting, I actually it was asked by client to media
- 40:02audit. Another agency was by media for
- 40:04them and like they refused to give over access to the DSP and
- 40:08their and their excuse was well, there's multiple clients and
- 40:11they were just fine, right? I mean, you can't prove or just
- 40:13prove that, but you they couldn't give us access to other
- 40:15clients data and they then they refused to give us the markups.
- 40:17It took months, months of fighting legal letters from from
- 40:20the client eventually to get that what was legally theirs
- 40:22right. So I think that is that's
- 40:24interesting. Not super surprising to hear and
- 40:27that's yeah, it must have been frustrating.
- 40:29David, love to hear from you. I mean, what, Like you had a
- 40:31good example in the article about, you know, I think raising
- 40:33your hand and saying like, maybe you shouldn't be buying his
- 40:35inventory, right? Yeah, so I'll actually give 2
- 40:39examples. One was a very early on in my
- 40:41career formative experience. So I started out at a, at a good
- 40:46publisher, Gameloft. They create great mobile games.
- 40:48I was part of the team that rolled out their initial
- 40:50programmatic offering. So that was a really good
- 40:52learning experience about how publishers operate and how they
- 40:55build out their their programmatic stacks and all
- 40:57that. And you know, I, I learned what
- 40:59I needed to learn there. And then I went to a vendor,
- 41:02right? And this vendor operated on an
- 41:05arbitrage model. That's how, you know, we would
- 41:07sell the media to XCPM. And then the goal was to buy was
- 41:11to maximize margin and buy it as at a lowest possible of ACPM as
- 41:16possible while hitting the advertiser's goals at the same
- 41:19time. So I was trying to find that
- 41:20perfect mix of hitting the goals while maximizing the margin.
- 41:25Very early on, I noticed that the sites we were running on by
- 41:29and large and where the majority of spend was going to were sites
- 41:32that I'd never heard of that if you went to these sites, they
- 41:35were your classic templated, you know, not legit content type
- 41:42sites. You could call it MFA or
- 41:45whatever you want. This was back in the mid two,
- 41:482010. So MFA still wasn't was a name
- 41:52people were throwing around yet. So like I noticed that I'm like,
- 41:57you know, I don't think anyone would want to be on these sites,
- 42:01but at the same. Time.
- 42:02So is this principle based media that was it like where they
- 42:04actually are buying and then selling keeping?
- 42:07Yeah. But it's not like we weren't
- 42:08getting the inventory upfront. We were just basically a just
- 42:12buying through ADSP through. So it was.
- 42:15Just classic arbitrage, right? But my bonus is were based on
- 42:19how high my margins are, right? So I was getting the I was
- 42:23driving the highest margin business and getting the highest
- 42:26bonuses out of all the traders and I felt really dirty and bad
- 42:30about it. Yeah, you were buying like the
- 42:31cheapest inventory possible, selling it at the highest price
- 42:35possible. And then you lo and bold realize
- 42:38the cheapest inventory you know was being produced, but you
- 42:41know, in sweatshops overseas with WordPress sites that keep
- 42:45generating more and more versions to show fake ads, real
- 42:49ads to to fake people in a way. Exactly.
- 42:51So like on average I was getting 80 to 90% margins and the
- 42:55advertiser was none the, the, the brand was none the wiser.
- 42:58Because, you know, of course we weren't transparent.
- 43:01So we couldn't share the site lists even when asked, you know,
- 43:04the, the blanket excuse what, that our publisher relationships
- 43:07forbid us from sharing those site lists, even though we
- 43:10didn't have any direct publisher relationships.
- 43:12We were all buying, we were just buying SSP volume, but we could
- 43:16provide a top ten list, right? And I always wondered why, why
- 43:21the publishers in that top ten list were OK to share, even
- 43:24though that top ten list was completely fabricated and just
- 43:27had sites that, you know, would look good to the layperson.
- 43:32So Needless to say, like I didn't feel good doing that.
- 43:35So I didn't stay too long at that company.
- 43:37But that was one of my, and also, I also realized just how
- 43:41easy it is to hit these vanity metrics that advertisers want,
- 43:48even if you're running on just BS inventory at the same time,
- 43:51right? So that was my early toe in the
- 43:54water and the incentives and just the way things work game.
- 43:57I went to another media vendor, media vendor after that.
- 44:01That one was actually worse in many ways.
- 44:03And we could get into that on another podcast perhaps where we
- 44:05share just more war stories. Holding company side, right?
- 44:09And when I went to the holding company side, that was my
- 44:11attempt at trying to make a difference and trying to, you
- 44:15know, use the knowledge that I had accrued doing things that I
- 44:18wasn't proud of to protect brands and do right by them.
- 44:23And, and, you know, I was still naive and I thought, you know,
- 44:26the agencies, they, they just didn't, you know, they, you
- 44:29know, they want to do right. And they just didn't know.
- 44:32They just don't know better for whatever reason.
- 44:34And they just need more knowledgeable people coming from
- 44:37the, the other side because typically people go from the
- 44:40agency side to the vendor side. I don't think it's as common for
- 44:44that for people from the vendor side to go to the agency side,
- 44:46right. So I was a bit naive and I
- 44:48thought that, you know, I could drop make a difference and help
- 44:51educate brands and drive value from there.
- 44:55But lo and behold, you know, even at the holding company, I
- 44:57saw that our trading desk was still buying all the same crap
- 45:01that I was buying on my past rolls.
- 45:03And you know, I tried to do something about it.
- 45:05I tried to speak up. I tried to, I actually pitched
- 45:09the client on a what at that time I called a curated
- 45:14marketplace strategy. And this was before curation was
- 45:17even a thing. But my whole idea was why do we
- 45:19need to go to these middle men? Why do we need to buy 10's and
- 45:2210s of thousands of sites? Why can't we?
- 45:25We're a holding company, We have the direct publisher
- 45:26relationships. Why don't we just onboard a
- 45:28bunch of? Just trading allow list, right
- 45:30of like premium publishers, right?
- 45:31Yeah. Well, not even in allow list.
- 45:32I was like, let's just get P and PS going right.
- 45:34Let's just build up a marketplace of P and PS, layer
- 45:37on audiences on these bundled P&P marketplaces and then we
- 45:40could still scale. Premium, yeah.
- 45:43You know, it's going to be a bit of a higher CPM at the same
- 45:45time, but like we'll still be able to scale all our audience
- 45:47targeting. You know, we could bring in more
- 45:49publishers, but we'll have more control over the over the
- 45:52inventory set. We'll know exactly where the
- 45:53money's going. We'll know what we're buying.
- 45:55And I thought that was a great idea.
- 45:57And, you know, the client thought it was an amazing idea.
- 46:00Well. The client liked it.
- 46:01Client liked it. Client loved it.
- 46:03Like this makes a lot of sense, but my leadership didn't like
- 46:07it. They didn't like it at all.
- 46:09Yeah. And the margins?
- 46:10Lower margins were lower on the on the other inventory, is that
- 46:13it? It was so it was a combination
- 46:18of margin, non transparent margins, kickbacks via the
- 46:23different suppliers and all those, you know, all those
- 46:26things that go and. So it economically didn't make
- 46:28as much sense to the agency to be fully transparent.
- 46:32And the brand, just so we know, from a brand perspective, was
- 46:34this a, you know, kind of a big premium brand that should, you
- 46:38know, yeah, this is Microsoft. So.
- 46:42So they should. Yeah.
- 46:43This is not like a Temu, you know, cheap product direct from
- 46:48the warehouse type of situation. This was a real brand that wants
- 46:51to get in front of real buyers, Right.
- 46:54And, you know, talk about engineering relevance into the
- 46:57advertising, sort of the bid stream or the advertising
- 47:00ecosystem. It doesn't seem to me that
- 47:02you're achieving the goals of Microsoft by advertising against
- 47:06shitty content. That is, largely the audiences
- 47:09that go to these places are probably largely irrelevant to
- 47:13Microsoft's interests. Well, right, you think about the
- 47:15brand supported David's strategy.
- 47:17David as the planner and buyer is recommending a strategy,
- 47:20right. I mean, and then yeah, it makes
- 47:22sense like the brand, you think they'd be supportive right
- 47:24anywhere, which is great, but then the agency didn't like it.
- 47:27It sounds like an. You'll love it.
- 47:28No, and you know, this is actually right after I got
- 47:31promoted to a bit of a higher level of role where I could, you
- 47:34know, set strategy and call some level of shots.
- 47:38But you know, the end result was I was put on a, on a pit and
- 47:41basically told, you know, we like what you do.
- 47:43We like you, you know, we, we want you to stay here, but if
- 47:47you keep rocking the boat, you know we're going to have to.
- 47:49Quit being so honest. Quit listening to your clients.
- 47:52Yeah, quit things. Yeah, quit.
- 47:54Quit doing much right for them. Right.
- 47:56So it was like, yeah, yeah. So it was like, shut up and
- 47:59we'll all be good, or you keep talking and there's going to be
- 48:02problems. And my ultimate decision was I
- 48:05quit and then I went to do my own consultancy for four years
- 48:08where I got to, you know, actually work directly with
- 48:10brands and agencies that one, do right by their clients and help
- 48:14them out in the ways I thought best and drive value.
- 48:18So. So yeah.
- 48:19So like on one hand, there was experience where, you know,
- 48:23keeping quiet actually made me more money, right?
- 48:25Not doing necessarily right, drove my bonuses.
- 48:30And then there was another just like, I think, you know, a lot
- 48:33of people on the holding company side deal with this type of
- 48:35situation where they just, you know, come in, they do their
- 48:38job. They don't, they know they can't
- 48:40rock the boat because they'll jeopardize things And, you know,
- 48:43they come in for a paycheck and and then leave.
- 48:46But for me at least, I just couldn't.
- 48:47I couldn't do that. I mean, do you think it's
- 48:49changing? And Sarah, I'm going to throw
- 48:50this at you. I mean with with the spotlight
- 48:52being shown on on on these dark corners, these dark recesses of
- 48:56of misaligned the dark. Underbelly I love the.
- 48:58Dark underbelly of the ad ecosystem.
- 49:02I mean, do you think that that things will change and how, how
- 49:05do you better align incentives? Because you could argue that
- 49:07there's certain incentives that could be driven by the brand.
- 49:11In this particular example, it was not the case.
- 49:13They, the brand seem to want to do the right thing that the
- 49:16agency wanted to maximize margin, right.
- 49:19And so that that became your kind of breaking point.
- 49:22What, what do you think? What do you think needs to
- 49:25happen for, for the, you know, things to improve?
- 49:28Do you think it will change? I'm reluctant that it's going to
- 49:33change. I see different companies.
- 49:37I call it the shell game there. Whenever a word gets out about
- 49:41one thing and where the bodies are buried in the dark
- 49:44underbelly, which is the brand, start understanding what
- 49:50questions to ask. So now you know we're open.
- 49:53Path for Trade Desk is a good example of that for me.
- 49:56I love going direct to publishers, I like having a
- 50:01closer throughput. But in the case of open path,
- 50:06they're changing the algorithm to wait inventory and push the
- 50:11inventory delivery into open path, which they're making a
- 50:15higher margin on. And the problem with that is not
- 50:19that it's competing with SSPS on the CPM, which is every argument
- 50:24I keep getting. It's that they've created a new
- 50:26supply path. They've forgotten that
- 50:29everything starts at the top of the waterfall and now they're
- 50:32competing against APG deal. That's the apples to apples
- 50:35comparison where brands already are buying that upfront and now
- 50:40they have to unravel and detangle an inventory mess that
- 50:45Trade Desk has not been proactive in communicating that
- 50:48are making higher margin on that.
- 50:50Q4 earnings are coming out soon, so I'll be curious to see when
- 50:53they're dipping in there what that will look like.
- 50:57But ultimately that's one example of Trade Desk burying
- 51:01bodies on their fees, their platform fee, which is
- 51:04compounding fees. They mark up on their inventory,
- 51:09they're marking up on data, marking up on everything and it
- 51:12it's doesn't show up on an invoice, so.
- 51:15And and they're eating into the direct buy plays by the
- 51:19publishers, right? That's correct.
- 51:21And so the the end point here is they've now gone in and created
- 51:26a new mode to get rid of all the fee stocking, quote UN quote,
- 51:31and that's called the advanced mode.
- 51:33Now. They did not.
- 51:35This is when they labeled all s s PS as resellers in the
- 51:38platform. So in September they actually
- 51:41started making these changes. They did not.
- 51:43Did they label every SSPA reseller?
- 51:46And that's de prioritized in the algo, but also they have
- 51:48something called the DQI, which is the TV quality index.
- 51:53I bet you it's de prioritized on the score as well.
- 51:57So you know, you're seeing this, this inverse effect with
- 52:00delivery now over indexing and competing with large brands that
- 52:03have direct direct buy. So my argument here is not that
- 52:07it's wrong. I don't like dislike the
- 52:09strategy and going, you know, close to make sure that we have
- 52:13great inventory regardless of where it's coming from.
- 52:16It's the execution. They're charging publishers 5%
- 52:21as a fee. They're undisclosed markup on
- 52:23top of that. And now what they've done, this
- 52:26is my articulation of an example, is they've changed fees
- 52:29in one area and pulled it into a different area and, and, and
- 52:33buyers are not really aware of this change.
- 52:35A. Lack of transparency around
- 52:38that. It's, it's, it's a little bit of
- 52:39a game gaming the system in a way.
- 52:42And I have to give them credit because like, you know, there's
- 52:44been this whole narrative over the past year or so of like
- 52:47direct versus resellers and resellers are bad.
- 52:50And that narrative is largely seeded and started by the Trade
- 52:53Desk, right? And they started pushing, you
- 52:55know, they, they were, they started putting these like S
- 52:58pretty good, these S chain reports for clients showing them
- 53:01how much resold inventory they're buying and how they need
- 53:04to optimize into direct and to turn off resellers.
- 53:08And it's a, it was a long game because then they rolled out, of
- 53:11course, they like they already had open path.
- 53:13And then they started rolling out open path more and more in
- 53:17their, in their position, in their market positioning and
- 53:19pushing it harder to clients. And then they have that SP 500
- 53:23and then, you know, they forced Kochai on everybody.
- 53:26We're now open path and SP500 are on by default when someone
- 53:30creates a new line item in, in Kochi.
- 53:33So I got I, I do have to give them some level of credit.
- 53:36It was some good social engineering there, not that I.
- 53:39Disagree. Push.
- 53:40Yeah, push. Push from a positioning
- 53:41perspective to direct and then actually eat into the direct.
- 53:45Margin, that's right. Become a direct supplier in.
- 53:48A yeah, which competes with the agencies they work with.
- 53:51So think about hold coves. Part of the value proposition
- 53:54there is they're buying massive amounts of inventory.
- 53:58I'm going to flip what David said on its head earlier and you
- 54:01get economies of scale for brands.
- 54:04Well, this is essentially what The Trade Desk is now doing.
- 54:06So how does that work with the whole Co model?
- 54:09TBD on what that will do to cannibalize that.
- 54:13Plus they're going to wreck with many big brands.
- 54:15I had a big pharma client trade desk was there going direct and
- 54:18you know, they're still working through agencies.
- 54:19But you know, how much longer do they if they don't have to, they
- 54:22won't, right? I mean, it's clearly like
- 54:23they're, I mean. Yeah.
- 54:25That's the direction they're going in, I would argue.
- 54:27Yeah. So that's the optimism.
- 54:30That's one example. It's very recent.
- 54:32It's it's taking something and turning it into a different
- 54:36problem. So I think ultimately, what
- 54:38should people do? I would like brands to hold
- 54:41themselves accountable, to ask the right questions, hold
- 54:45companies to margin instead of CPMS, like to make sure that
- 54:49they ask questions on going beyond the invoice, not just
- 54:54what's buried in the CPM. Tell me every dollar quarter by
- 54:58quarter how much you took of my media dollars so you get to the
- 55:01true working and non working. Would it be possible to get an
- 55:05answer though? I mean, like it's so it's so
- 55:06opaque, right? Not if you don't push them in
- 55:09the programmatic realm, absolutely.
- 55:11I've run those audits and I've terminated contracts as a result
- 55:14of the findings because you can't lie on the data.
- 55:17This is where I do like quantify.
- 55:19But you would need to see the markups, right?
- 55:20You would need to have access to in platform to see them right?
- 55:23Well, I it was a manual pull on the behalf of this one DSP that
- 55:27I'm thinking of and it was over 55% fees.
- 55:30So it was $1,000,000 that they were taking in profit on this
- 55:34one campaign a month. And I ran one audit Sarah, and
- 55:38like I found some markups up to 400% for no reason.
- 55:42You know, it's crazy. I mean, I think that's a little
- 55:44extreme, but you know, the 15% is what they're supposed, you
- 55:48know, let's say contractually supposed to be taking if you're
- 55:50seeing markups for if extra work is going in and it's justified
- 55:55maybe, but as a lot of the markups I remember finding were
- 55:57not. So keep going.
- 55:58I'm sorry I cut you off. No, that's OK, but it's
- 56:00egregious and I think that there are less outliers than they are
- 56:04the norm if you don't know what you're asking.
- 56:06So go beyond the CPM and go in the margins and pull manual
- 56:11pools. The third thing is coming from
- 56:15the buy side, I had really great partners that were very honest
- 56:19and their sales structures are meant to promote their own
- 56:21products. Can we actually turn that on its
- 56:24head? I don't really know that we'll
- 56:26combat that fully, but I always align myself with those that
- 56:29made recommendations or suggestions and went outside of
- 56:33their role and outside of their company because those are the
- 56:35good actors kind of want to like Shadow and, you know, ask for
- 56:38those POV. So do you think they exist?
- 56:41I think they should be speaking up more and ultimately, you
- 56:45know, that's kind of some of the advice I'd give, but there's a
- 56:48lot more where that came from. This is where I think training
- 56:51comes in. Yeah.
- 56:52And do you think and Speaking of training, do you think that
- 56:54there needs to be, is there a re skilling here?
- 56:56I mean, are you're, if it, let's say we're just folks of the
- 56:59marketing department for a second, we could go to other
- 57:01departments. But like, do we have the right
- 57:03people in place that know the questions to ask that are
- 57:06technical enough to understand the nuances of the ecosystem for
- 57:10digital in general, not just programmatic or CTV and
- 57:12otherwise? And, and do we have to reskill
- 57:16in that department or, or in that area?
- 57:18Or do we think that they can that, that, that people have
- 57:22the, the knowledge, they're just not asking the right questions?
- 57:25I mean, what? What do you think?
- 57:26I I think that legacy CM OS are largely grown up in brand and
- 57:30product positioning like where I started.
- 57:33What you're seeing is the convergence of AI and hack and
- 57:37machine learning and all sorts of different complexities that
- 57:41is starting to flip that, but it's a slow flip.
- 57:44You're getting people like me that are starting to get
- 57:47ingrained and T shaped more in that analytics realm.
- 57:51And my feeling is that there's still a lot, a lot of brand and
- 57:56legacy brand. I'm not saying that's bad, but
- 57:58what you need to make sure you're hiring right on your team
- 58:01for those gaps and be aware of those gaps.
- 58:05But this is where I personally love that I'm leaned into
- 58:08analytics and tech, but also have the business strategy side.
- 58:12So the consultancies and advisories like mine will pay
- 58:15off dividends if you just do sounding board gut checks and
- 58:19also don't rely on your own team because then you're just talking
- 58:22to yourself within your walls. So there there's definitely, I
- 58:26think in the future going to be this turnover of talent and
- 58:31where they grew up. But right now, it's still
- 58:33heavily in the brand realm from what I see.
- 58:35How are you, David? I know we're getting a little
- 58:37short of time. What, what are your like if you
- 58:38were to give some advice to let's say people on the brand
- 58:41side like what, what, what coders should they be doing in
- 58:44order to get ahead of this to create a better environment or
- 58:47maybe or is it not their responsibilities?
- 58:49The agencies love to hear your advice.
- 58:51No, I think I think they need to take more ownership over their
- 58:54house, right. So like 1 you know, you could
- 58:56bring in Sarah. I think that's that's always
- 58:58going to be a. Good decision, always a good
- 58:59idea. Thank you, David.
- 59:01But. Two, I mean you do need to bring
- 59:03some technical level of expertise in house, in house in
- 59:06some way. I know, you know, I know there's
- 59:09companies that have completely announced programmatic
- 59:10operations for some that they can do that others can't, but
- 59:15you should not be fully trusting and dependent on your media
- 59:19partners and agencies and just let them let them do their thing
- 59:23and unquestionably right. I've seen that not work out way
- 59:27too many times. Just just middle last year I
- 59:31remember auditing ACTV campaign for a law firm and they spend a
- 59:35lot this up. This law firm spends about 50
- 59:38million on TV all all in but their CTV vendor and they didn't
- 59:42have the agency. They had a vendor and I service
- 59:44sold them on the Super Bowl. They sold them CTV impressions
- 59:47on the Super Bowl when those are were not able to be bought
- 59:52programmatically. Right?
- 59:53I know 120 this. Is not a Peacock ad.
- 59:56This is what you're saying. No, How could you even?
- 59:58You couldn't even buy it, right? It's just, it's not even.
- 1:00:00Available. You can even buy it.
- 1:00:01Yeah, yeah. On demand later not.
- 1:00:04Live. I don't.
- 1:00:05Yeah, no, not the only. Argument I think.
- 1:00:08Yeah, so like all this brand got was just a spreadsheet that said
- 1:00:11Super Bowl placement, you know, X CP/M, X impression and and
- 1:00:15that's it. And it doesn't take rocket
- 1:00:17science to figure that out, to ask the right questions.
- 1:00:20You don't have to be radically technical or analytical or a
- 1:00:22data scientist to figure that out, right?
- 1:00:25Well, you need, you should, you should have some curiosity.
- 1:00:28You should do your homework on it.
- 1:00:30But a lot of people just don't have the time or they just
- 1:00:33think, why would my vendor RIP me off like this?
- 1:00:37They, yeah, they don't realize that, you know, ultimately this
- 1:00:40is unfortunately an ocean of sharks that we're all swimming
- 1:00:43in. So have someone check your
- 1:00:47agency's homework. It sounds like that's probably a
- 1:00:49good idea, whether someone in house or maybe a third party,
- 1:00:52like bringing a shower, right to do that.
- 1:00:54I mean, it sounds like that's probably a smart thing to do
- 1:00:56periodically, safe to say. I'd I'd say so I'd say so.
- 1:01:00I know there's, you know, there's there's third party
- 1:01:02auditing companies as well, like an ambiquity.
- 1:01:05I hear good things about them. I just highly cautioned not to
- 1:01:09fully trust anyone party with your media spend because there's
- 1:01:14just too many bad incentives around that might make things go
- 1:01:16awry. And how about about untangling
- 1:01:18all of the markups and take rates and different fees within
- 1:01:22that, say within ad tech rate? I mean, I've worked a lot of the
- 1:01:24publisher side and you see just, you know, they're getting
- 1:01:27pennies in dollars sometimes, right?
- 1:01:29And I've, I'm working with one on cable company now they're
- 1:01:32talking about, you know, like traditional dollars versus
- 1:01:35digital pennies, right? Because it's just, it's such a
- 1:01:37different environment. How can they start to untangle
- 1:01:39that as well? I mean love your thoughts on it.
- 1:01:45I mean, that's, that's a tough one.
- 1:01:47I mean, you need to have that, that ad tech expertise in house,
- 1:01:50right? And you need to have really
- 1:01:53strong contracts. I mean, all honestly, at the end
- 1:01:55of the day, it really just starts with your contracts with
- 1:01:57your partners. If you're not contractually
- 1:01:59protected, then there's really nothing you could do to untangle
- 1:02:03those fees. And this goes for the
- 1:02:04publishers, this goes for brands, right?
- 1:02:06You need to have full audit rights.
- 1:02:08You need to make sure that you could, you're able to check
- 1:02:12every single dollar and where it went and also have a full
- 1:02:15understanding of who these vendors or, or tech companies
- 1:02:20where they're sending your money to, right?
- 1:02:22And how the dollar breaks down. But I believe that the best
- 1:02:26thing they could do is start with having airtight contracts
- 1:02:30because then at the very end, at the, at the very least, they're
- 1:02:33not going to bring in a Sarah to come in and, you know, set
- 1:02:36things right. Lawyers are very, very.
- 1:02:41Scary. It's interesting to say that
- 1:02:42because I remember I, I was brought in by health system to,
- 1:02:44to audit some of their, their work and we found, we found it
- 1:02:47was around 22% ad fraud, right? Which was, it just seemed it
- 1:02:50didn't seem like it was intentional.
- 1:02:51Seemed to us more just sloppy. Like there was a couple
- 1:02:53affiliates that started, I mean, a lot of the traffic was
- 1:02:56actually going overseas, right? And then, you know, it started
- 1:02:59performing well, lo and behold, and so they optimized and
- 1:03:02actually sent more traffic there.
- 1:03:04And then we couldn't even get a make good because contractually,
- 1:03:06to your point, they didn't know them anything, right?
- 1:03:09So, so we had them rewrite the contracts when they, when they,
- 1:03:12when they, they ended up reviewing with the same agency,
- 1:03:14but they ended up saying, OK, if this happens in the future, you
- 1:03:16least owe us a make good, right? We have we have audit rights if
- 1:03:20we, if we find egregious fraud that we, that we, you do owe us
- 1:03:23at least and make good for that for that budget amount.
- 1:03:25So I think that's a good point about contracts.
- 1:03:30Well, I see we're getting short on time here and I know, David,
- 1:03:33you have a hard stop. So maybe before we go love,
- 1:03:36maybe if you could tell our viewers how they, if you know,
- 1:03:39sour, for instance, maybe start with you viewers want to get a
- 1:03:41hold of you. If they have any questions about
- 1:03:43the work you do, if they want to contact you, if they want to
- 1:03:45bring in an expert, let's say to take a look at these things, how
- 1:03:47would they get a hold of you? Yeah, thank you guys also for
- 1:03:51your hospitality and this amazing show.
- 1:03:54Here I am on LinkedIn primarily Sarah Caputo and I do follow
- 1:04:00quite a few people. So if you want to look under the
- 1:04:03ducks and crannies, if you're starting out, I think that's a
- 1:04:06good place to start in terms of who and what companies I follow.
- 1:04:11I also can be I cured through you of digital.
- 1:04:15I ran about 50 workshops for them last year from March to
- 1:04:18December, so yeah, so I think I was almost 50.
- 1:04:22I ran last year, my first year working with them, so I'm.
- 1:04:25Shiv was on this pod. He was one of our early guests,
- 1:04:27yeah. Yeah, I love Shiv, so I can also
- 1:04:30be found there. Nice.
- 1:04:33David, how about you? If I knew you want to, but I
- 1:04:34know you post a lot in social and you, yeah, you're very
- 1:04:37active. And if someone's get hold of you
- 1:04:38like how? Where's the best way to find
- 1:04:40you? Yeah, I mean, I'm terminally on
- 1:04:42LinkedIn, unfortunately, but that's probably the best place
- 1:04:46you can reach me. And I'm always happy to respond,
- 1:04:49share notes, happy to always look at data campaigns and just
- 1:04:54let you know my feedback. So Linkedin's probably the
- 1:04:57easiest place to reach me. And you know I'll I'll never
- 1:05:00leave you hanging. Right, pop up a QR code because
- 1:05:04David. This may not be the right
- 1:05:07medium, the medium of choice for QR codes, but but yeah, for
- 1:05:11those watching on YouTube, we're going to show a QR code, no.
- 1:05:15But the bounce around the screen right?
- 1:05:17Just like that. Yeah, exactly.
- 1:05:18Pick up your phone and. It's I missed my neon light in
- 1:05:22the back. Yeah, yeah.
- 1:05:23So this was a great episode. We're definitely gonna have you
- 1:05:252 back. Thank you so much.
- 1:05:27And and and Sarah and Deb are gonna be contributing to to SIG.
- 1:05:30Brett maybe we can do an episode on like the real dark underbelly
- 1:05:33and we can all pick our shadiest stories, right?
- 1:05:36Yeah. Yeah, exactly you.
- 1:05:38Can bring in Foo for that one we can.
- 1:05:40Bring a Foo. Yeah, yeah.
- 1:05:43And so, so wherever you're listening, subscribe YouTube,
- 1:05:47Apple Podcasts and Spotify, we've got lots of episodes
- 1:05:52coming down the pipe and we'll see you next time.
- 1:05:56Thanks. Everybody and check out The Cost
- 1:05:57of Keeping Silent by Sarah and I.
- 1:06:01Thank you for doing that AI dot AI.
- 1:06:04See you guys. Bye everyone.