← AdExchanger
AdExchanger
The Outcomes Era Is Dead. Long Live The Quality Era
00:53:25
Emet Advisory’s Erez Levin explains why the digital ad industry should transact on media quality signals like attention instead of optimizing to outcomes, conversions and flawed attribution models. The former Googler also weighs in on the ruling...
Loading summary
Transcript60 lines
- [00:13]
Anthony Vargas
Welcome to Ad Exchanger Talks, the podcast devoted to examining the issues and trends in advertising and marketing technology that matter most to you.
- [00:24]
Allison Schiff
This podcast is sponsored by Data Axel. We where actionable data fuels connections. For over 50 years, they've helped businesses turn information into insights, igniting meaningful relationships between companies and the people they serve. With cutting edge data solutions, AI and technology, they empower brands to engage their audiences in smarter, more personal ways. Learn more@dataaxle.com.
- [00:56]
Eris Levin
Welcome to the Outcomes era. Or is it the Quality era? The industry is selling a vision of driving outcomes using sophisticated database targeting and attribution. But our guest today argues that the industry has long been obsessed with outcomes, much to its detriment. And if you take a look at what's really going on in digital advertising, we're actually entering a new era where high quality media, rather than cheap impressions that exploit advertiser KPI's will win the day. I'm Anthony Vargas, Senior editor for Ad Exchanger, filling in for our usual host of Ad Exchanger Talks, Allison Schiff. And I'm joined today by Eris Levin, principal at marketing consultancy EMET Advisory. Eris's experience in digital marketing stretches back nearly 20 years, including a decade plus as a product and strategy specialist at Google Aris. Thanks so much for joining us today.
- [01:46]
Anthony Vargas
Thanks for having me, Anthony.
- [01:48]
Eris Levin
Before we get started, we have some quick business to take care of. We're about a month away from ad Exchanger's programmatic I.O. las Vegas. The event is taking place May 19th to the 21st at the Park MGM on the Vegas Strip. We'll be kicking things off with our brand new Innovators Lab and we'll be discussing all the hot ad tech topics from curation and retail media to alternative IDs and AI. Book your pass today if you haven't already, because time's running out. All right, with that out of the way, let's get on to it. Erez. We usually like to start these episodes of Ad Exchanger Talks with more of a fun kind of light hearted question. And as we'll kind of discuss as we go forward, you had a 13 plus year stint as a Googler and everyone knows that search is the window to the soul. So what have you been searching for lately that tells us something about you that we might not already know?
- [02:41]
Anthony Vargas
Oh, searching lately is probably a lot of vacation planning. So spring break, heading with my family down to Florida and then this summer planning a trip to Portugal. So I guess the, the, the deeper insight there is that my wife and I both Love to sort of like travel off the grid. We were big backpackers before we met and since we met and now we're trying to get our young kids sort of into that lifestyle. So excited for Portugal and sort of taking a little bit of adventure travel with our younger children.
- [03:13]
Eris Levin
How, how easy is it to kind of bring the kids into the backpacker lifestyle?
- [03:18]
Anthony Vargas
We're, we're easing into it. We went to Belize in December and that was our first sort of trial of, you know, sort of like, you know, not as developed and comfortable living. But they loved it. They had so much fun. So we're getting there.
- [03:33]
Eris Levin
Yeah, it feels like in some ways it could be difficult, but maybe kids are more inclined to embrace that kind of off the grid living.
- [03:42]
Anthony Vargas
As long as it comes with a promise of ice cream or a popsicle after a hike up a hill, then they're generally cool to keep going.
- [03:50]
Andy Frawley
There you go.
- [03:52]
Eris Levin
Okay, before we get into anything else, I have to get your reaction to the verdict in the Google Ad tech antitrust case. Were you surprised by the verdict and what kind of fallout do you expect from it?
- [04:04]
Anthony Vargas
Don't. No, I definitely was not surprised. I don't think anyone was really surprised. It seemed like kind of the, the obvious or most likely outcome of sorts and not going to opine like on the specifics of it. But you know, are they a monopolist and you know, have they committed some anti competitive practices? I think everyone agreed that the answer was yes. The fallout I think is honestly, I think it's really hard to predict. So no one seems to agree. And I don't think I have any sort of Magic 8 Ball on this. I think there's too many moving parts and pieces and negotiations and it's not entirely clear to me. I think I do know what Google wants, but I see some signals that indicate otherwise. So I don't know if that's misdirection or maybe I'm just not miss, you know, interpreting everything correctly. One thing that I do feel confident about, which I think is missing from the discourse broadly because a lot of folks are sort of opining on what's going to happen, but also what that means, what that means for quote unquote, the open web. And, and, and there's even different views there. Some people think this is going to, quote unquote, save the open web. Some people think this is the end of the open web, like Google is propping it up in some way. What I think is missing is sort of defining what the open web means and what I do believe is going to happen is we're going to see the splintering of this open web. We're going to see the long tail, I'd say on the lower quality range, mostly undifferentiated inventory and media, that's kind of going to go away and maybe it mostly should. Whereas quote unquote, the open web that is more differentiated, that's unique, that's valuable both to advertisers. Advertisers and consumers sometimes. You know, we call that the hedged gardens. I call it sort of the, the kind of open web because it is accessible to multiple sort of DSPs. It's a little bit more open. It's not, you know, fully in this walled garden world. I think that is going to grow and I think that's, that's sort of going to continue this sort of like going away, the erosion of this fully open web, the long tail stuff and then the growth in this hedged garden world that's going to happen no matter what sort of the conclusion of this result is, whatever the remedies are, whatever happens with Google's parts. And so that to me is the most important thing. And I don't think that changes with or without the monopoly ruling.
- [06:36]
Eris Levin
And then I guess this is a more like immediate in terms of its impact change that Google recently announced. Last week Google made the decision to abandon its plans for a new prompt in Chrome that would allow users to opt out of third party cookies. It looks like we're sticking with the cookie status quo on Chrome after years of deprecation talk. What are the implications here?
- [07:01]
Anthony Vargas
I think kind of similar to the previous point, I'm still kind of getting used to this. It's crazy. I think everyone is both shocked and not shocked at the same time. I do think the end of what the cookie has propped up for so long, it's within sight. I think enough buyers have moved on from attribution. They know that it's, you know, highly inaccurate and likely misleading. They know it only helps them on a very small subset of the Internet. And even that part, if you misinterpret the cues and the signals and the models that it sort of builds on, then you could potentially be misdirecting a lot of your media spend. And so I think especially for the enterprise marketers, it's not going to make a huge difference. I'm hopeful it doesn't. I'm hopeful that we've sort of moved beyond it and we've sort of been talking for so many years about how the cookie was flawed anyway, especially once there was, you know, as much of it missing, as much of that coverage for addressability missing as we have today. And so I don't think it's going to make a gigantic difference there. Where it could play more interestingly is in the like SMB PMAX type of world where advertisers may be a little bit less educated, less discerning, more willing to just give Google, you know, their money into a black box solution and let it run. And let it run not just on Google's O and O, but also on any third party inventory that can now be sort of somewhat bolstered by third party cookie coverage for targeting and for measurement. That's assuming that there isn't sort of, this isn't part of the bigger spin out of, you know, the sell side or some other programmatic business or even of Chrome, but assuming sort of all stays equal on that part. I think this doesn't have a huge impact on the enterprise or it shouldn't rather. And, and I don't think it will. It might delay a little bit longer. More buyers that have been resistant to move off of cookies can stay on it for a little bit longer. But I think there's enough of a movement of buyers that have moved on aren't looking back and that's going to force the market to adapt to better ways to target measure by media.
- [09:18]
Eris Levin
And with that kind of cookie decision. Is the writing on the wall for the Chrome privacy sandbox or do you think it could still be a viable tool going forward?
- [09:30]
Anthony Vargas
I personally don't see it. I mean, I haven't been close to it, but it seems so bogged down with trying to solve everything. And I think it's just fundamentally really hard to do any of that. There might be some use cases that work, but to do it as sort of this broad solution I think is really, really tough. Just marketing doesn't work that way. Attribution doesn't work that way. There might be parts of it that are valid and valuable if you sort of, you know, acknowledge that they're imperfect models. But I don't think that's how it was presented. It was presented as something that could almost lift and shift all of our systems, attribution, buying, et cetera, from cookies to these cohorts. And I don't think that was really viable. So it'll be interesting to see if some parts of that go on and sort of live either branded as the privacy sandbox or maybe somebody else takes them and runs with them. But overall I don't see Google investing much more in the privacy sandbox moving forward.
- [10:35]
Eris Levin
All right, so that's, that's the news out of the way. Let's dive into some other topics now. So let's kind of, you know, we'll shift gears here and kind of get down to the nitty gritty. In a guest article for Ad Exchanger recently, you really kind of pushed back against this framing that digital advertising is entering its outcomes era. What's going on with all this outcomes era talk and why do you disagree with it?
- [10:59]
Anthony Vargas
Well, I think outcomes is, it's tough because it's a kind of a moving target. I think some folks can be referring to outcomes and maybe most when they talk about the outcomes era, they are talking about real business outcomes, sales and incremental impact. But I think too often for the last decade plus, a lot of folks have translated or just interpreted outcomes to mean sort of the media metrics. And those could be really, really the vanity metrics that we see too often playing out in our space. And so I think that's my concern is that even while some people might be referring to outcomes in this aspirational, really sort of beneficial way, too often it's going to be translated into, even if it is real outcomes, not necessarily incremental ones, just ways that platforms and such will take credit for sales and business objectives rather than actually driving them.
- [11:58]
Eris Levin
Yeah, and like you kind of alluded to at the end of your response there, like, you know, in some ways like this, like obsession with outcomes has really kind of like led the industry down the wrong path and led it to focus on the wrong things and like, you know, enabled, you know, maybe bad actors or just like people who are like self interested to kind of like, you know, claim credit for things that maybe they shouldn't be. So like, you know, in what ways would you kind of like highlight as far as like how the industry's focus on outcomes has led it down the wrong path?
- [12:25]
Anthony Vargas
Yeah, I mean, I think programmatic, even more so than digital, but digital as well, especially in the search era. It just gave the ability to what, the promise? I look back at all of these articles from 2010 before after about solving Wanamaker's conundrum and sort of being able to eliminate all the waste and knowing exactly how our ads were working. And there was probably some truth to that at some point, but I think once we started to scale, it was easier to actually scale by taking credit because you served ads to certain people, you knew kind of who they were, which devices they had, and then if they made a purchase later and you could take credit for it because it showed up on the spreadsheet in a dashboard, then somebody was willing to sort of accept that as effectiveness. And slowly but surely budgets shifted towards that media that could take credit, not necessarily was more effective, but could take credit for those outcomes.
- [13:26]
Eris Levin
And you know, your column brought up this interesting idea that like, you know, that obsession with outcomes produced all these kind of harms and like, you know, negative kind of like behaviors, even when we had wide scale availability of signals like third party cookies and device IDs to aid targeting and attribution. So, you know, those signals, it's not going to come as news to anybody listening to this podcast. Those signals are vanishingly available. And now we're seeing kind of like a bit of a shift to probabilistic modeling instead, although everybody's still using those deterministic signals where they can as well. But when we think about this kind of shift to probabilistic modeling, does that shift kind of make it harder to address these attribution challenges that you're highlighting?
- [14:12]
Anthony Vargas
It doesn't necessarily make it harder to. It makes it harder to address those using our existing systems that assumed that everything was deterministic and accurate. I have sort of long advocated for leaning into the probabilistic. If you think about all marketing sort of pre the digital era, it was all based on models, it was all based on probabilistic, looking at large sort of populations, large cohorts of audiences that were reached with, you know, millions of ad impressions, we don't call them impressions then, and what happened in those sort of different population groups as a result, and then deciding what were some of those variables, those inputs that helped determine the likelihood of getting the outcomes that we wanted. And so I think we just need to lean back into that. I think looking at the sort of current era, my biggest concern about this shift of probabilistic is that it's not usually declared, it's not usually transparent when something is deterministic or probabilistic on the targeting side or on the sort of measurement of outcome side. And so I think, you know, I think that's the number one concern I have with the topic of ID bridging. I'm okay with ID bridging. We just need to make it really transparent so buyers aren't checking a box to say, you know, target these 18 to 34 year olds that like motorcycles, thinking that those are all deterministic signals, knowing which ones are probabilistic and assigning some sort of probability, right, some sort of discount to the value that they would ascribe to that because they don't know for sure that that person matches the audience that they're looking for.
- [15:53]
Eris Levin
And when you think about kind of like, you know, moving more towards that model, is that something that market forces are best positioned to kind of effectuate that change? Or do you think there's room for there to be some kind of, you know, standard or like, you know, industry kind of like, you know, guidance or mandate around, you know, doing that instead?
- [16:16]
Anthony Vargas
Generally, I would, I would probably vote for both. I think we should always look for the market forces to push them. And I think we see that. But there's really strong market forces or incentives to ignore those market forces, right? To ignore those proper solutions. I think there have been really good solutions for years around quality that have not been completely ignored, but certainly haven't caught traction as much as I think they should given their proven effectiveness and sort of the, the alternatives not working very well. But I think the, you know, let's say top down, more centralized, you know, decisions and guidance and standards, they should be somewhat, they shouldn't be too rigid. And so for the exact purpose like going back to ID bridging, just make it really clear that the bridged IDs should go into a different field. Just understanding the incentives of the system are going to be to pick something simpler, to pick something that spends the money, make it so buyers have to intentionally choose to say no, I do want the probabilistic. And they would see that showing up separately somewhere. It's sort of more of an explicit choice rather than, you know, in a way where no one's really accountable. Everyone can sort of raise their hand and say, well, we told, you know, we put it in there, it's in the standard somewhere and buyers check the box and buyers can always, you know, make an excuse that they didn't know. And so I think just inserting more accountability and that can be from the top down while there's market forces that show a better way to, to, for everyone to get their money's worth.
- [18:00]
Eris Levin
And then like you kind of, you mentioned that, you know, there are some like, solutions that are maybe like, you know, underappreciated or underutilized out there. Like what are some like, things that you would point to as like, you know, things that maybe have gone underappreciated, that could, you know, introduce more transparency into the marketplace or like, you know, more of a priority around like quality or like, you know, more control around like whether buyers are buying against like probabilistic modeling or like deterministic signals. Like what are some things you would point to?
- [18:28]
Anthony Vargas
Yeah, you know, I definitely have a bias more towards the media quality side and especially things like when I think about media quality, the way that I've generally defined it is all of the signals outside of those tied directly to an identifier, to a specific audience, like to an ID of sorts. And so, you know, I think where we first saw this innovation coming up and where I certainly sort of entered the media quality space was on attention measurement. And so that's just considering the relative prominence of an ad on a page, the likelihood that someone is going to pay attention to it given, you know, its various dimensions. And so I think that's probably still the number one that's underutilized given how much buyers still rely on viewability and completed views as sort of what they consider quality. To me, those aren't quality, they are perhaps hygiene. Even though I don't think, you know, there's plenty of non completed views that have value. But I do think that thinking about the relative value and not on a binary is, is super important. And so I think the attention category is one that's very much undervalued. There are other ways to value media quality besides attention. You can think of the time of day, you can think about the geography somebody is in for those. I think custom bidding algorithms are probably the still most underutilized solution to help buyers assign that relative value. And so they're not necessarily bidding the same amount and paying the same amount for an impression at 2am as they would at 2pm as they would at 8pm or in different cities where they probably have different market share, different priorities, depending on their objectives, depending on the creative.
- [20:22]
Eris Levin
And then when you kind of think about like, you know, your framing of like, you know, the quality era and like, you know, your argument that like actually like we're like, you know, we're entering the quality era of digital advertising. What's like, what does that mean, like for people who are kind of unfamiliar, familiar with this concept of the quality era, how would you kind of describe it?
- [20:41]
Anthony Vargas
Yeah, so I think we're moving away from where the primary currency of media buying is not. It's not certainly a user, it's not audience data necessarily and it's not the predicted outcome from that individual impression. It is thinking about what are the dimensions, the qualitative dimensions of media. And it's not just media. There's also audience data, there's also creative. And I'll get back to sort of how I frame this up. But when you consider all of those things together, what is the likelihood that an outcome will come later even if you can't tie it back to that individual impression? It's something that outcomes are still measured, they're still very important. But it's one of those things that you measure more likely at the end of a campaign to validate or maybe within an experiment to validate. What are some of those qualitative dimensions that help you determine the relative value of each, potentially each impression. But even sort of big buckets of media. Now, the one way that I framed the quality era that's been helpful for me, there's generally speaking three sort of pillars of media buying and measurement. So you've got media, creative and audience. And so I think about quality for each one of those three pillars. So you've got media quality, creative quality and audience data quality. And the bets that I'm making, I think what's going to sort of we're going to see success in for the quality era is the companies that either have really high media quality, they have really high quality, or higher than average, let's say audience and data quality, or can create higher quality creative or the companies that can differentiate between high, medium and low quality in any one of those three pillars. So the attention vendors, for example, they don't own any media, but what they do is sort of help differentiate so that you pay the right amount. I think there's a lot of value in low quality media. Like you should still spend there, but just make sure you're paying the right amount. Same thing can apply to deterministic versus probabilistic. There's nothing wrong with deterministic, with deterministic, sorry, with probabilistic data. But make sure you're probably going to want to pay less for it. And then the same thing with creative, I would love to have the highest quality creative all the time, but there's still some value in deciding that's maybe not the perfect creative for this audience and for this media, but, but it still has some value. And so just being able to price those things I think is really important.
- [23:29]
Eris Levin
And like, you know, it feels like nailing down media quality is like a particularly like tricky target to hit. So like, what do we really mean by like media quality? Is it like about avoiding, you know, the stuff that buyers seem to want to avoid like MFA and you know, outstream video, invalid traffic? Or is there more to it than that.
- [23:48]
Anthony Vargas
I, I think there's a lot more to it. I mean that, that to me going back like that's the hygiene buyers should have the ability and they should say there's some type of inventory that we never want our ad to show up next to either because we're concerned sort of with the implications in appearing next to it, or we just think it has no value to us as a brand or it's just not worth sort of like spreading the peanut butter to, to, you know, too broadly on that. But generally speaking, where I think we've completely ignored quality within the media context is just assigning that relative value according to these various dimensions. The reason MFA has done well is because most buyers, the currency, the thing that fuels their buying decisions are various vanity metrics, which is either reaching a cookie no matter where they are, or reaching a cookie that also is likely to. That person is going to make a purchase later that I can attribute it back to an impression that was served onto their computer at some point. And so it didn't matter if you're on an MFA or the New York Times, you're probably going to get it for a little bit cheaper on the mfa. I think the media quality doesn't say don't buy the mfa. It just says the likelihood of that being the, you know, the driver of incremental effectiveness is relatively low because it's a really small ad, it's one of 20 on a page, or maybe it was served at 2am or in some zip code that you know, you've never heard of or, or something else or had some other signals that made it seem like this is probably less reliable of a driver of the incremental impact. And so to me it's, it's just about that relative value. Again, some buyers can say this is worth zero to me. And that's a little bit more on the hygiene side. When you get into quality beyond. It's thinking beyond binaries and assigning that relative value for potentially each impression. But even, you know, you can think about it as buckets of, of different impressions too.
- [25:48]
Eris Levin
And what, what signs would you kind of point to in the market right now that show that we've entered the quality era?
- [25:55]
Anthony Vargas
I don't know that we have entered the quality era, but I think we are on the precipice. So one, we're seeing a lot more discussion about quality, like the number of brands and agencies and everyone else talking about quality, talking about attention, showing success there, certainly the attention vendors have gotten a lot more Traction but I think other quality vendors, whether on the data side, creative side, custom bidding, have all sort of for me and you know, I have a bit of a probably bias here because of what I'm exposed to, but it seems like they're getting more and more traction and as well as seeing like the folks that were not leaned into quality before coming in, you know, different vendors, measurement vendors, also seeing the writing on the wall, I mean when DoubleVerify bought Cybids, I thought that was not necessarily a watershed moment but it to me started to indicate that there was, you know, sort of more seriousness beyond that. It's not just about quality, it's kind of just like getting away from vanity metrics, getting away from attribution specifically. And I think we're seeing that over and over. Not just the number of buyers leaning into mmm, the number of providers from mmm, we're also seeing, you know, Google and meta creating their own open source MMM tools. Like to me that is one of the best indicators that the writing is on the wall. Even the big guys that have benefited the most from these legacy systems know that they're not long for this world and so they're trying to build alternatives to be able maybe not own the measurement as much as they did before, but still be able to influence and give their version of the events that help buyers probably spend more within their platforms. And maybe it's less Google search and more YouTube, maybe it's less Meta, you know, Facebook in feed, you know, 300 by 250s and more Instagram stories. Which I do think sort of thinking about where things have probably been a little bit overvalued due to attribution and sort of last touch to where things have been undervalued that were actually relatively high attention. But if you're only looking at viewability or view through conversions or sorry, completed views, they also probably undervalued.
- [28:18]
Eris Levin
And then when you think about how publishers can kind of lean into this new focus on quality, what are maybe your top three pieces of advice for publishers?
- [28:28]
Anthony Vargas
Yeah, I think this has always been a tough one. I'll just like back up a little bit on this answer. I've sort of spoken to you know, hundreds of companies over the the years and this brands and agencies and ad tech and every intermediary from the sell side and the buy side. Generally speaking when I speak to a brand directly I tell them like they should incorporate quality immediately. This is something that will like automatically help them stop wasting money on low quality supply, that they're probably overpaying for and help them find high quality media that probably is available at somewhat of a discount now because not everyone is valuing it properly. And so brands should make the move to quality asap. They have all the financial incentives to do so. Almost everyone else in the industry is not necessarily going to make more money on day one if they shift to quality. If a publisher decides to take out some of, you know, instead of having 10 ads on the page, decides to have five ads on the page, slowly but surely those ads will increase in value because buyers are starting to say, hey, I want less ad density per page. But on day one, day two, they might see a decrease in their total sort of monetization on a page. And so I do think it's a matter of one, being really thoughtful and sort of probably being a little bit gradual in this. But two, it's not just ensuring you have higher quality media. So fewer ads on the page or whatever the, whatever you can do to sort of make those higher quality ads, it's making sure you're telling that story. Because the buy side, for the most part, a lot of the demand continues to favor quantity, continues to favor vanity metrics and at the cheapest price. And so that's not what you're offering if you move from quantity to quality as a publisher. So I think it's really important to go and tell that story. And that means probably it's sales and marketing. This is just like I think Bloomberg and the FT and sort of other publishers that have rejected the open web, the open auction of Programmatic and said we're pulling out. This has not been beneficial for us. They also needed to supplement that with, you know, their efforts to go and tell the market and to tell the buyers, hey, we are offering something that's higher quality. One, you can't buy it in this open auction environment where you probably want to, and two, it's going to cost you more. The CPMs are higher and this is why this is a unique audience, whether it's with first party data, audience data that we have, or without it, or it's just a really valuable audience that you want to reach that you probably can't reach as easily anywhere else. This is why we're still worth it and this is why we still need to be on your plan. That's not something that every publisher can do. That's not something that every publisher has probably that unique either audience or something else that buyers can't or think that that is interchangeable. And so it's hard to give really broad recommendations, but I think for the publishers that have higher quality media audiences, they need to lean into this. They need to ensure that buyers understand why they deserve a premium over the cheap stuff that they can get anywhere else that someone is telling them ties to a third party cookie, which it probably doesn't.
- [32:04]
Eris Levin
All right, that feels like a good spot to take a break. When we come back, Erez and I will be discussing principle based buying and of course, Google. Stay with us.
- [32:26]
Allison Schiff
I'm Alison Schiff, managing editor of Ad Exchanger and with me I have Andy Frawley, the CEO of Data Axle. And I have a few questions for him about some burning issues. Hi Andy.
- [32:39]
Andy Frawley
Hello, Alison. Thank you for having me.
- [32:41]
Allison Schiff
So it's no secret that many businesses struggle with fragmented data spread across multiple platforms. This is something Ad Exchanger covers all the time and this leads to inconsistencies and inefficiencies. So how can organizations break down these silos and create a unified data foundation for better marketing and better business performance?
- [33:01]
Andy Frawley
It's a great question, Alison. This is one of the burning questions that the data acts what we see within our customer base. And while a lot of brands have made progress on linking legacy data together from their operational systems, what's happened over the last really five years is we've seen this massive new set of data that's being generated which is the exhaust of all the digital advertising platforms. And so linking all that data together with third party data with the first party data really requires an identity spine. Historically, brands have relied on third party identity graphs to do that work. The trend we're seeing is that brands would like to own that identity spine and so help build that identity out with third party data, with first party data and have a spine that links the known to the unknown. Obviously, you know, in a highly compliant fashion.
- [33:56]
Allison Schiff
Yes, always in a highly compliant fashion. That's very important. Well, marketers are, are shifting from vanity metrics to outcome driven strategies. How can brands use cross channel analytics and AI to ensure real business impact in a changing digital landscape?
- [34:16]
Andy Frawley
We, we certainly recommend to our clients to, to focus on business outcomes, whether that's, you know, new customers, more customers, more profitable customers, customers that stay longer and, and really move away from the vanity metrics of opens and clicks. So there's two important concepts when we, when we think about this. One is incrementality. You know, is the marketing effort creating incremental outcomes, I.e. sales and causality. Is it the, the marketing treatment that's actually causing the, the consumer to buy or act in A certain way and the crush analytic tools need to embrace both of those concepts. The complication or challenge is somewhat back to the first question. First, you have to be able to link a lot of data together and second, there'll be places where you have sparse data, where you don't have complete data sets. And so we're also seeing people using Gen AI to help generate those customer journeys and have a analytically based view of what the media exposure is across multiple marketing treatments.
- [35:21]
Allison Schiff
Down with data silos and down with vanity metrics.
- [35:25]
Andy Frawley
Absolutely.
- [35:26]
Allison Schiff
Thanks for the insights, Andy.
- [35:45]
Eris Levin
Okay, and we're back. So Eris, for those of us who've never worked for Google, it might be easy for us to think of it as kind of like the prototypical big tech company with all the baggage that comes with that. What's a story you can tell us about working there that will either prove or disprove the stereotypes about big tech?
- [36:05]
Anthony Vargas
I don't know that it's prototypical there. There's a couple of things that make it quite unique and I would say probably two specifically that I would point to like my diagnosis for kind of what's gone wrong within Google and across the industry. But I think the conflict of interest is a big one. And there's other companies that have a buy side and the sell side platform or sort of these sort of competing platforms internally, but none of them have the power, the market power that Google has. And I think that has created just really bad incentives and dynamics internally and that bleed out externally. And I think that's made it really hard for an increasingly harder over time for Google to honestly serve, you know, the various parts of Google. Right. Whether it's DV360 to ensure they are always serving their advertisers best, or for the sell side of Google to always serve their, their publishers best, or is it sort of the best compromise that they're able to find that satisfies a lot of them. And I think smaller companies that don't necessarily have that monopolistic power can't, can't do that, right? They just had to pick one or the other. They had to be a little bit more honest. And, and I think Google has been able to get away with more than others to, to sort of satisfy neither, right? To find this compromise that's probably in Google's best interest and probably not necessarily in the best interest of their, their publishers and buyers. And it could be a compromise like that's not to say that you could rationalize like actually this is a good compromise. Position, but pretty much 99% of the time that's also the best position for Google too. And so you never know if you're really getting the best thing. And so I think that's the, the one that I would pick out as, as what makes it different from all other big tech.
- [38:17]
Eris Levin
You posted on LinkedIn a little while ago that you still believe that a majority of Google's employees and leaders believe in the company's old don't be evil motto. But you added that they're stuck in a system that doesn't allow them to do the right thing. So how would you kind of point to how that system has contributed to, I guess, kind of the current state of the Internet? And is the current state of the Internet kind of like a place where the bad guys are empowered to kind of win online?
- [38:46]
Anthony Vargas
I think this is a little bit harder because I really think Google is emblematic in a lot of ways, but it's not unique here. And I'm critical, but really try to be thoughtful and empathetic. I think this industry as a whole, right, there's a lack of accountability. We sort of created these systems that no one person or entity is fully accountable and sort of, that distribution allows a lot of the bad stuff to thrive. And we just sort of like, you know, put our hands up when that happens. Going back to Google and to my previous point, I think it is their market power that should, in my opinion, makes them more responsible. Right. They do have that power. They have that singular power if they say, you know, they dictate something, as they've done before, Right. And it wasn't necessarily a dictate, but they can set up a signal and be this sort of moral voice and, and also move the market. And so adstxt was always the example that I pointed to. I think Google did a, a tremendous job there at the time. And they moved first. They told the market they were adopt this and hold everyone accountable. And that got everyone to follow along that success. Unfortunately, that was probably one of the last times that Google stepped up to, to kind of set the bar high. And I think that failure to continue doing so is what allowed this sort of unaccountability across the industry to continue to thrive because everyone could always point to, well, like, well, this is the system and they're not doing it and if I go first, then I'm going to lose. And so that to me is probably the bigger, the bigger issue and why I do, you know, with as much sort of like respect and empathy as possible, but Put the onus more on Google than others to in some cases not in all, but to step up and sort of do the right thing because of this ethos of, you know, do no evil. And, and if they don't do it, then the market doesn't move.
- [40:54]
Eris Levin
And yeah, to, to your kind of point about like, you know, Google's like ability to really like, you know, throw its weight behind moving the market. You and I have talked a bunch about how Google is still mixing the old and new standards for classifying in stream video ad inventory and its platform. You know, given, given stuff like that, you know, like, like is there, is there like, you know, do you feel like Google is like sufficiently kind of like pushing forward that like Quality era mindset or is this kind of like, you know, perpetuate is like what specifically what they're doing with like in stream video, is that kind of perpetuating some of the old problems that are preventing, you know, us from really entering into that Quality era?
- [41:36]
Anthony Vargas
Yeah, the answer is yes. There's a bit of a caveat. I would say like three, six months ago I would have very much said that Google is not really preparing us for the quality era. There's a little bit of a, there's a resistance. Right. There's money to be lost and certainly some disruption. I've always made the case. I still believe it to be true. Google has parts of its business like YouTube that are very undervalued and parts that are really overvalued. Net Net. Will they be better off after we sort of shift into the quality era? I don't know. But at the very least there will be disruption and I can't blame them for not wanting that disruption. So I think that's been part of the resistance for a number of different sort of shifts to the quality era. The in stream outstream is, you know, one of the most acute examples that, that I certainly am aware of where that resistance comes in because some parts of the business will benefit, some parts will probably be neutral and some parts will lose. And I think there's no one at Google, is no individual at Google really feels accountable. If, if they don't do anything, then just the system happens and then if somebody does step up and try to change it, they're going to face a lot of resistance internally. And so I understand sort of the incentives there aren't to act. What gives me a little bit of hope and I think we're not going to see the impact of it in this programmatic world as quickly is and I've pointed this out to some other folks like Google is has been signaling for a couple of months now that they are shifting not necessarily towards quality. They don't talk about that word as much, but away from attribution at the very least. And attribution is one of those things that props up sort of the fake outcomes and this, you know, the vanity metrics and such. And so I do think we are seeing Google with their bets on their MMM product, with their emphasis on how YouTube is amazing and great at brand building and the biggest thing on ctv, they are signaling that they're sort of preparing for that shift as buyers are going to move from sort of an obsession. Right. It's not, they're not going to give up on performance and sort of short term effectiveness, but they know that there's probably going to be less of that because buyers aren't going to be able to prove it out as much without cookies, without attribution. And they want, Google wants to at least be able to pick up some of those dollars that look at performance over sort of a medium and longer term time horizon.
- [44:20]
Eris Levin
Google also recently reversed its policy and now allows its tech to support fingerprinting, which seems like it'll boost probabilistic targeting and attribution. So is fingerprinting a good thing or a bad thing in the context of the quality discussion?
- [44:36]
Anthony Vargas
Well, no one has really defined fingerprinting so I won't sort of go too deep on that term specifically. Generally speaking, I think the IP targeting is a good thing, you know, notwithstanding privacy concerns, which are real. But I'm not going to get into those. Generally speaking, I like deterministic targeting, I like probabilistic targeting and I think it's necessary IP address. I think we need to sort of understand that it is mostly probabilistic. We don't, you know, there's probably some instances where an IP address is, there's only one person in that household and we know exactly who they are and we can track them. But even there that person is leaving their house, they're getting served ads in different places and there's going to be some probabilistic matching that's done not to mention the ads that they see not in digital channels. And so there's always going to be some modeling applied. I'm as I mentioned earlier, like let's lean into that. Let's say when things are probabilistic and allow buyers, allow the ecosystem to validate how accurate those probabilistic models are, there's no Perfect here, but I do think that it's going to give us more ground truth, it's going to help validate where things are working, it's going to give more targeting ability. But we can't get to perfect here. And I think we just need to accept and acknowledge and really just sort of lean into this fact, fact that we're living in two worlds, one where we know exactly who the user is and one where we have some sort of predicted match to it and not even necessarily to a specific user. Maybe we don't even know, but we know amongst these 10,000 people that we're targeting, 5,000 of them are going to match pretty closely to roughly who we want to reach and then just using those various signals accordingly.
- [46:31]
Eris Levin
And then Google, like just about every other big tech company, is heavily investing in generative AI. Do you see Genai as a threat to the quality error or could it actually be a benefit?
- [46:47]
Anthony Vargas
It could be a benefit, absolutely. One of the sort of framings that I look at is when you know a lot about a user. So let's say like in closed ecosystems, whether it's like Google or Meta, when they know precisely who a person is and they know a lot of things about them and maybe they can even do some closed loop attribution, the ability to modify an ad and customize it to that specific person can be really powerful. Outside of that, I think when you don't know exactly who a person is and you don't know exactly how they're going to respond, you don't want. I mean, people talk about like thousands of variants of creatives. I think that's silly. Outside of these sort of walled gardens, if you're going to serve every single person, let's just say me and you and a thousand other people are watching some CTV ads or watching CTV and we're getting served some ads. If every single one of us gets a different variant because let's say the car on the screen is a different car being driven, that seems interesting. Like I'm not opposed to that, but how are they going to prove that that actually had an impact on us? Because none of us are clicking, none of us are scanning a QR code. Most of us, 95% of us are not in the market for a car at any given moment. And so they're not going to know for six months plus, you know, if that had any effect on us at the very least. And so I think the idea of Gen AI with like a ton of variants is not really the solution outside of These walled gardens where you know a lot of other things about the user and can sort of validate that different variants worked well. So there's going to be use cases where I think it's really powerful and it doesn't necessarily contradict the quality story. But I do think that for quality to work, I'm a big believer in sort of humans having a really kind of like oversight over quality decisions because these are human things, these are human insights that you get that you understand your business, your needs, your consumers. So the AI without a ton of signal about what's effective, which I just don't think is going to exist outside of, you know, these walled gardens, it's going to be more of a human deciding what's higher quality that might be informed from AI analysis and less from gen AI.
- [49:24]
Eris Levin
And now to shift to another topic that's very relevant to the quality era, principle based buying. You recently wrote another Ad Exchanger guest column in which you argued principle based buying has potential as long as media quality is protected. So what's kind of like all the hubbub around principal buying about and what are some things buyers need to keep in mind when buying principal media?
- [49:46]
Anthony Vargas
Yeah, I mean, there's a lot of forces here from what I understand. You know, agencies getting squeezed on their rates and then they have to make their margin elsewhere. Buyers sort of being happy to look the other way because they feel like they got a win here, you know, and getting reduced rates from the, from their agencies and then also getting sort of reduced media rates within the principal media. The problem is that there's no sort of quality controls. And so most of that media is being measured with vanity metrics like viewability or completed views or reach. And the problem is that agencies can just buy really, you know, sort of cheap stuff and sell it back to the buyer at a cheap price. Probably not cheap enough. And so I don't think there's anything sort of fundamentally wrong with the concept of principal media. This is something, you know, a practice that is done in so many industries, right. An agency, a manufacturer, they're buying in bulk, they're selling it for a different price. There is a, I think a fair argument that we want, you know, buyers should expect their agency to act as their agent and not as a principal, like really operating in their best interest. But I think we're sort of beyond that. There's not many industries where that actually works as intended unless they're regulated industries. And even there they're probably everyone's pushing the boundaries as much as possible. And so let's just operate in a little bit more of a trustless environment where the buyers just need to understand what are they actually getting and what are they actually getting. Has to be measured with quality. It can't be measured with vanity metrics because those are easily gamed. And then you get into this sort of lemon market. You want to create a lemon market of cars? I will sell you a vehicle with the steering wheel on four wheels. It doesn't, I'm not guaranteeing that it actually drives. And if I start to do that at mass scale, it just, the prices go down, it just gets worse and worse. And so the quality controls and that signal to buyers, I think could allow principal media as a practice to work. It'll probably be much smaller of a practice. It probably won't be anywhere near as lucrative for the agencies, but they can still make a margin from that. There's still value for them to buy a lot of this media up front, do the quality control on their side, working with different partners and then making sure that their buyers are still getting their money's worth, even if the agencies are making a little bit of a margin there, which I don't think is necessarily a problem.
- [52:20]
Eris Levin
Yeah, so it sounds like, you know, prioritizing quality is the best way to turn that lemon market into some lemonade. Thanks so much for an awesome discussion here. This is really great. And thanks, thanks a bunch to our listeners for joining us. Really appreciate you guys and we'll be back next time.
- [52:38]
Anthony Vargas
Till then, thanks Anthony.
- [52:52]
Allison Schiff
That's a wrap. Thanks for listening and a special thanks to our sponsor, Data Axle. Helping businesses create deeper, data driven connections with customers. Whether you're reaching new prospects or strengthening existing relationships, Data Axle delivers the insights and solutions to make every interaction more impactful. Explore more@dataaxle.com.