Below, we have provided the full transcript of the panel discussion, Back to the Future: In Need of a New Antitrust Framework from the second episode of our series, Digital Advertising & Antitrust.
Pallavi GUNIGANTI:
Hi, I’m Pallavi Guniganti, an attorney at the US Federal Trade Commission, and I’ll be moderating this discussion today of digital advertising and antitrust. Just to apologize, we had some technical difficulties that are why we are beginning a little late, but we hope to still have our full hour of discussion with our very knowledgeable experts.
I am here in my personal capacity, and anything I say or do as moderator does not necessarily reflect the views of the FTC or any commissioner. Of the people who can go ahead and speak freely, the first person we have on our panel is Leigh Freund, who is the President and CEO of Network Advertising Initiative, which is an industry trade group that develops self-regulatory standards for online advertising. Leigh joined in NAI in 2015, after 11 years at AOL, where she was Vice President and Chief Counsel for Global Public Policy.
Next up, we have Ramsi Woodcock, who is an Assistant Professor of Law at the University of Kentucky, and also has an appointment as Assistant Professor of Management in the Gatton School of Business and Economics.
Third, we have Christopher Yoo, who is the John H. Chestnut Professor of Law, Communication, and Computer & Information Science, as well as Director of the Center for Technology, Innovation and Competition at the University of Pennsylvania. Thank you all for joining us today. And we’re looking forward to an interesting discussion about digital advertising and antitrust.
Just to begin, I think it’d be helpful to talk a little bit about why we care about advertising, where it came from, and why it matters for content. So, advertising is a long time financial basis for free or below cost pricing of content. Newspapers, magazines, radio, broadcast TV, you can say these are some of the original two-sided markets, because antitrust historically has treated both advertisers and the content consumers as relevant when analyzing media mergers and other matters in this area.
It’s also still continuing to be a concern, it was not long ago that the Department of Justice challenged a newspaper merger in Southern California, saying that the acquisition would harm both the buyers of the daily newspapers in the area and would also harm the buyers of local advertising in those newspapers. So just to begin, I’d be interested in hearing the panelists’ thoughts on why they see this as a matter for antitrust to be concerned about, and where they see the direction going, as we move from traditional media into digital.
And Christopher, if you could start us off.
Christopher YOO:
Sure. Thank you just everyone at CPI for arranging this. And again, please I extend my apologies as well for the late start, we hope to make it worth your while. It’s fascinating, because a lot of the law generated for how advertising is treated by antitrust is quite old.
And one of the things I find fascinating is, to show how long I’ve been doing this, the amount of gray hair I have. I started off, my very first article was on broadcast television regulation and what really seems quite quaint right now in some ways, and the media ownership proceedings, which were a very big deal for a long time.
But what’s interesting is there is a lost economics of advertising, which is really focused around broadcast television, which many people have forgotten. And this has become very important. And many of the key precedents came out of a time when newspapers, particularly one, print newspaper towns were critical.
What’s the most interesting development here… In those markets, clearly, you could have market power, you could have consumer harm, and there’s a lot of traditional analyses. The most important change that’s happened is the case of Ohio v. American Express, which is really when the Supreme Court made a decision by five to four margin to analyze two-sided markets by looking at both sides of the market.
But what’s fascinating about that decision is the discussion of a precedent, which is Times-Picayune Publishing, which is an advertising precedent. And they made it a point to bracket off in that opinion, whether we should treat advertising markets as two-sided. They say, no, Times-Picayune says we treat this as one-sided. Even though as Pallavi you say, every economist would regard it as a two-sided market, or at least in large part.
And what we see now is, in my mind, once the Supreme Court says something like that, lower courts don’t get to deviate from that. And so this rule was made very clear, lower courts don’t get to overrule us, we overrule ourselves, even if it’s wrong. And you have to stick with it until it’s right, and it’ll be interesting to see, it’s clearly dicta, because there was not an advertising component to the Amex case.
And it’ll be interesting to see how it gyrates around, but whether it reaches the court again with an advertising case, it’ll come up differently because it was a five to four decision. But as of right now, I believe the lower courts will continue to adhere to these newspaper era precedents, because they have really a compulsion to follow what the Supreme Court has said they should.
GUNIGANTI:
And Ramsi, what are your thoughts on this? How do you see the antitrust aspect working with regard to traditional media now moving into digital media?
Ramsi WOODCOCK:
My sense is that ultimately, the two-sided market approach to advertising is too sophisticated for its own good, that we should think about advertising as traditional marketing, in which you’ve got an input, in which the input is consumer retention. And that’s being supplied downstream to distributors of advertising, which are selling their services to advertisers.
And I think that’s part of the motivation behind the bracketing that goes on in the Amex case. I think eventually, the courts are going to come to see it that way. You see in the Google complaint that was filed earlier this week, it goes out of its way to make the point that this is not a two-sided market, that we don’t have a situation in which both the consumers of the content get more out of the network, the more people advertise. That’s not the case. It’s certainly true that advertisers get more out of the network, the more consumers are on it, but it doesn’t go in the other direction.
I think it’s consistent with this view that traditionally, courts are viewed there as being both the content consumer market and also an advertising market. We have to think about the content consumer market as really being, trying to answer a question whether the input, the consumer retention input has been monopolized.
And that’s always going to be relevant in any traditional case, whether the input is under the control of the defendant is going to be key in deciding whether downstream, it’s exercising its power over that input to exclude competitors downstream. So although analytically, it’s never been posed exactly that way, I think we can see it that way. I think that’s where the courts are going to converge.
GUNIGANTI:
Leigh, can you bring us your private sector perspective? How are digital advertisers seeing both the Amex case and also traditional antitrust law in this area?
Leigh FREUND:
Thanks, Pallavi, and thank you to the CPI for arranging this. I’ll caveat that my area of expertise is not antitrust law. So I’m the solo non-antitrust lawyer on this panel. But I do bring practical experience, in terms of the digital advertising industry and what we’re seeing, in terms of competitive effects on the marketplace.
I think, for me, the way I think about antitrust and competition policy affecting our industry is data access. Ramsi spoke about consumers and consumer input and the way consumers interact with content online, is the main source, it’s the cornerstone of digital advertising. And so the access to understanding consumers and the access to consumers online is really either the choke point or the release valve and how bigger players in the market gate access to that data, really depends, really affects downstream the rest of the industry. And so when we think about it, we think about it in terms of access to data.
GUNIGANTI:
Leigh, I’d like to stick with you about this. So what do you see as the implications for advertising, of seeing data as an asset itself? And you mentioned that you’re not antitrust lawyer, but I imagine it’s significant to the companies that you’re working with. Because data has been seen as a useful asset for targeted advertising, even before we had the digital context, there was a infamous New York Times story about eight years ago about how the retailer Target already knew that a teenager was pregnant before her father knew, because it could tell based on purchases she’d been making… and they started sending what would seem like useful advertising and coupons to her family, who were perhaps not aware of what was happening.
So given that the data has been useful, but also, I think, sometimes concerning for consumers in the public, just how much is it being collected, what are the companies at NAI thinking about this? And how do they see the right way to handle both the antitrust and the privacy aspect of data?
FREUND:
I think the Target example is a good example of, sometimes we have too much data for our own good. I think the implications of the wide availability of data, and then the algorithms and the information that we can analyze to find out different things about different people, are really the differentiating point of digital advertising versus other advertising.
As you rightly point out, Pallavi, data has been used to inform advertising for decades or centuries, even. And it’s only this recent proliferation of data online, and the specific uses of data, the specific types of data that can be used to inform relevant advertising.
So, targeted advertising uses real-time data, usually its browsing history, but to gauge relevance of somebody’s viewing history and can serve relevant ads across a broad spectrum of sites in real time. So it’s a broad and instantaneous departure from the traditional norms of how you gather purchase data or other data offline and attach it to advertising.
I think when we talk about data as an asset, it really is the cornerstone of what we’re doing in digital advertising. So who has access to that data and who can cut off access to that data, really, really affects all of the downstream players in the industry.
So, today, we rely on something called the cookie. I know everybody probably knows what that is. It’s a little piece of code that goes on to your computer or onto your browser, rather, that allows folks to gain information.
The third-party cookie has come under fire for privacy reasons. I think that’s where the intersection of antitrust and privacy is really interesting, because we find that some dominant players in the industry are acting in the name of privacy, but are also receiving economic benefits of doing so.
We can see this in Apple, for example. Apple does not get a cut of advertising revenue from ads that are served on apps within a user’s experience, but they do take a 30% cut of any paid apps. And so the concept of taking away an advertising ID that folks could use to monetize free content online and make that available to consumers, in lieu of an arrangement where they get a 30% cut, is a really interesting argument. Are we gating access to data from downstream market players in a way that impedes competition?
GUNIGANTI:
Ramsi, what do you think about this aspect, that the privacy of data, but also the necessity of data, as Leigh has said, and who’s the gatekeeper on it?
WOODCOCK:
My view, digital advertising and targeted advertising, it’s better, right? It’s more effective, because of the data. Unless everything that the industry has been telling us turns out to be a massive fraud and we can’t actually learn things about people from their data that create conversions, which I don’t think is the case, it makes advertising much, much, much more effective.
And that poses a real problem, I think for antitrust, because there’s a historical strain in antitrust, which has been about protecting advertising markets. And treating advertisers as consumers who need antitrust protection. But there’s also a long strain in antitrust, particularly in the mid 20th century, that viewed advertising as fundamentally anti-competitive.
There were cases brought in the ’50s and ’60s, there’s a famous one, challenging Procter & Gamble’s purchase of Clorox. And the idea there was they’re going to be able to bring their advantage in advertising to bear, in order to get consumers to buy Clorox over other forms of bleach, even though all bleach is chemically identical. And the implication was, advertising has a manipulative, irrational quality to it. It’s not just about providing consumers with useful product information.
And today, in the information age in which we’ve got free availability of information on any product you want, you just go on the maker’s website or you go on Amazon and read specs. It’s hard to see advertising today providing any useful product information we can’t get elsewhere.
And so if we look at it that way, we have to ask is advertising today, one, more effective because of all the data it has about us, but two, also no longer socially useful, just purely manipulative? And I think if that’s the case, if it’s just manipulative, it’s bad for competition, because it means that the better product doesn’t win, the better advertised product wins, thanks to this manipulative characteristic.
I think this is a serious problem. Do we want to promote more competition in advertising, in targeted advertising, in an era in which all advertising may well be manipulative? Maybe instead, this alter ego in antitrust that in the mid 20th century viewed advertising as anti-competitive, needs to kick in today, and start treating advertising more as a problem for competition, rather than something that should be promoted through competition.
GUNIGANTI:
Christopher, what do you think about that idea, that we should go back to looking at advertising as a kind of consumer harm? And one thing I’d be interested in is also, was advertising only ever useful as a source of information? Or is it also useful as a creation of identity, right? The reason why you would do the very beautiful, full page ads and vote for perfume is not to inform people about the chemical makeup of the perfume, it’s to create a sense of identity around it.
YOO:
I think laws pretty much recognize that there is an informational component to advertising. It’s not just directly about products, it’s about pricing as well and what we learn about specific opportunities. We now have a commercial speech doctrine where we protect that with the First Amendment. It’s a fairly strong line. And in addition, what we understand, it’s also about branding, about investing in quality, signaling there’s a much more dynamic way we think about how information markets work.
But what I really love is Ramsi’s focused on ad effectiveness, and this is where the old literature on broadcast television comes in. You have essentially two choices of how you want to fund an industry, a content industry, which is you can get direct payments for it, or you can do advertising.
Now, the problem with advertising is the revenue that you generate through an ad supported content space is not determined by how much the consumers like the content. It’s determined by how responsive they are to the ads in the content. And that could have any relationship that you could imagine, it could be more, it could be less.
In broadcast television, it’s seven times less than people actually like the content that the ads are embedded in. And so if you rely on advertising, the broadcast television world says you’re going to underfund content, you’re going to get too little content, the breakeven number goes up, you’ll have less diverse content, and you have all these problems.
And there’s a number of great studies of when the paid television industry came up, first in terms of subscription over the air, but later cable television, where you have some great, really interesting studies on this. Now, what’s interesting is the great… and there’s another, if you really want to go long here on the economics about this, advertising responsiveness is flat. So it makes it what economists would call a voting regime instead of a pricing regime.
Which is, suppose that I like some obscure sport that’s only popular abroad. And there’s 1,000 of us, or 100 of us in Philadelphia who like this. Well, on an advertising basis, 100 people, we’re never going to see this content, it’s just not enough of us.
But if we’re so nuts about this and we’re willing to spend $10,000 a year for it, the small 100 of us could actually pay off this business model, because we can signal the intensity of our preferences by paying more, which you can’t do in advertising. Unless you buy a lot more of the product and that’s just not a real credible way to do this.
And so there’s a long literature realizing that, in fact, advertising is a pretty bad way to fund an industry. But if we have better targeted ads, you close that wedge. And in fact, it’s an interesting sort of question here, whether that’s part of the solution.
Our normal response in newspapers is, there’s some that are purely subscription supported. There’s something that are purely ad supported and some new hybrid models in between. And left a free choice of how they’re going to generate it, the normal ways let people decide for themselves.
What’s interesting, to touch on two things that were mentioned earlier is, we started off talking about advertising and then Ramsi shifted to attention… We moved away from straight advertising and Ramsi characterizes them as attention markets. There have been a number of high profile proposals along those lines.
They haven’t really gone anywhere, in my opinion, because the big problem we have there is if it’s about attention, then everything competes with everything. Because attention is movies and games and walking outside and spending time with your family. And it becomes really hard to think about how we’re going to make that tractable.
Leigh shifts it to data, which is actually a different analysis, which is, there’s two ways you can think about it. When data is sold as a separate product, it’s a conventional analysis when we do that. When it’s sold as an input to another product that’s not separately sold in and of itself, we have a whole different… I have an article about our attempts to do this in the past, particularly in innovation heavy markets, when we look at… there have been attempts to look at innovation markets.
And it’s really hard, because you have to then tie the impacts on data to the impact on consumers. And as we were chatting before this began, even the Google complaint acknowledges there’s actually some benefit to the use of data for consumers, in terms of improving search. And so we have a very complicated world.
Last thought, I love Leigh’s example of the Minnesota case. First, it’s creepy, and that’s the standard reaction a lot of us have. Second, it’s not clear it’s an antitrust violation, because it’s not clear that’s going to hurt markets as a whole. And third, to me, it’s an example of failed data science, because any conclusions you make about whether a person is pregnant, or has any other demographic characteristic, is only probabilistic.
And actually, you’re going to make errors. And if you send out enough of these, you’re going to make errors. If it’s one in 100,000, you can count up the number. And what they’ve learned to do is actually, the retailers now send ads that are clusters of different products, including things that have nothing to do with pregnancy, like lawnmowers. But they include a higher number of pregnancy focused ones.
And in fact, the redemption rate on these coupons is huge. And for those of us who’ve had kids, they include diaper coupons, those things get redeemed 100% of the time because kids go through diapers like crazy. And so there is some consumer benefits.
So trying to find a good data science way to take what is an interesting example, and play it out in the real world in the way we’re going to benefit consumers is actually much richer, and much more interesting.
GUNIGANTI:
And on this topic of the traditional media, I’m interested in how does digital advertising affect our long standing market definitions and regulations of traditional media? So for example, historically, there have been limits on whether TV and radio stations can cooperate on advertising sales.
Whereas on the other hand, there’s a long standing, decades old law that allows newspapers to do this, at the point that Congress saw that newspapers were already struggling. So I’d be interested in hearing, I guess, Christopher, if we could start with you.
YOO:
Sure. To me, this is an empirical question, and I see a lot of generalizations and statements. And really what you want to see is, do these forms of advertising substitute for one of these? I don’t know.
I have a confession, which is, before I went to law school, I went to business school, and I worked in the advertising department for Procter & Gamble. At the risk of falling into… we held an event yesterday where Maureen Ohlhausen called this anecdata, appropriately.
I will say, we… turning the knobs on everything, everything was just… and so that’s a terrible way to begin this. But there is a very small literature and empirical literature that looks on whether offline traditional advertising substitutes for online, and whether different forms of online substitute for each other. I can go into a chapter and verse, but the answer is the preliminary data suggests yes.
Now, these are unnatural experiments, and they’re very limited, in terms of, before we get a case, we’d actually have to do a lot more. But that’s the kind of question that you’d like to see is to actually look at what the switching behavior is, and look at how major advertisers think about this and getting testimony.
Because if they are in the same market, then this looks very, very different than it would otherwise. And so it does change it when we look at market definition and it may change over time. And in fact, what we see is a growth of online advertising where it’s now over 50%.
And so what we’re going to have to look is individual behavior and understanding that and that’s the natural fact base why antitrust should do this.
GUNIGANTI:
To follow up on that, when you’re saying digital advertising, is this lumping together all forms of it as being substitutable or is it looking specifically at targeted advertising? And because there’s EU law on that where this was a 2014 merger decision. Viacom, Channel 5 broadcasting by the European Commission, where the commission said the specificity of online advertising sets it apart from less targeted forms of offline advertising.
YOO:
I find conclusions like that without evidence behind them a little bit not helpful. To me, it’s all about price. We always looked at fairly untargeted blunt instruments that had a certain cost per impression and a certain responsiveness to the medium. We had other ones where it was relationship advertising, where you’re actually really getting deep into that, which cost a lot more on the data science, which has gotten a lot cheaper, frankly.
And so the idea of making categorical statements like that, to me, I would not want to be the government lawyer with a burden of proof trying to make that case in a court of law. You need a little bit more than that, in my opinion.
But the actual studies I’m talking about, there’s four of them. They’re not a lot. They follow different methodologies. We’d have to get in the guts of some of them about how these statements rely on a targeted… an answer is some do, some don’t. And so unpacking how that works is the hard work of antitrust, in the case by case analysis we do.
GUNIGANTI:
Leigh, what do you see as the perception within industry? Do companies see online and offline advertising as substitutes, as competitors? And then how do they perceive targeted advertising in particular?
FREUND:
Yeah, I think, it’s really interesting, because I have in my notes of preparation is digital advertising its own market, right? And then how do we define what that is? Is social media a sub market of that? Because there are lots of different types of digital advertising.
I think what we’ve seen in the industry is certainly digital advertising, and by that I mean all forms of digital advertising is taking share from the more traditional means of advertising. The pricing models are completely different. The pricing models depend on data.
So targeted advertising is a much more expensive form of advertising, that wields much higher results, because consumers are, depending on the study you look at, three to five times more likely to click on an ad or interact with an ad, if it is relevant to their interests, as defined by the targeting. And so the question of, digital advertising was a $125 billion market in 2019, up year over year, taking share away from more traditional media senses.
And now we’re seeing especially with the pandemic, we’re seeing digital video ads actually just eating up share from traditional broadcast television, as well as other forms of digital advertising. I think it’s a really interesting question when you talk about defining the market, even advertising in general, but digital advertising and then the sectors within digital advertising.
So we see mobile advertising, web advertising, video advertising, completely sold separately and interacted with separately, with different groups within agencies, handling all of those different components, almost as silos of an overall bigger campaign for an advertiser. So defining the market, I think, is a really, really difficult and tricky subject.
GUNIGANTI:
Ramsi, what do you think about this, and including your description about advertising as a potential consumer harm? Is that intensified by having it be digital, by having it be targeted? And how do you perceive the interplay of competition there?
WOODCOCK:
I think it is, to the extent that digital advertising is more effective in a world in which consumers don’t need the information content of advertising, I think we can conclude that it means it’s more manipulative. And so certainly, I think one of the… that is the classic, the central problem posed by digital advertising.
It’s obvious, it seems to me, that over time, targeted advertising and digital advertising will come to dominate. A lot of the reason why we don’t have it dominating yet is simply because we have these legacy technologies that still have user bases and attention basis.
But as those disappear, it will become an all digital world And certainly the growth of digital has been coming at the expense of legacy forms of advertising for the most part. And it tells you right there that they’re competitive with each other.
Although obviously when we go out to define a market, we have to look at the nitty-gritty and so we may not be able to define markets quite as broadly as I’m speaking, in general terms, it’s pretty obvious that digital advertising is a substitute for traditional, and will become more so as time goes by. And its power, what makes it not just a substitute, but a superior product is precisely what should make us very worried, I think in antitrust about what its competitive effects are.
FREUND:
And so if I could, sorry, Pallavi, if I could just jump in on, I think we’ve been using digital advertising and targeted advertising almost interchangeably. I think it’s really important to recognize that they’re completely different. Digital advertising mean a much bigger umbrella over top of what targeted advertising is.
But I think the other concept that is interesting, I started by talking about data access. And that is this, that in digital advertising, especially as led by demands from advertisers, data is not just used for targeting.
So data access, all of digital advertising is dependent upon data access. There’s measurement and accreditation and all of the analytics that folks do on the back end to determine if somebody has seen an ad and interacted with it, or went offline to make a purchase. I hear a lot of, especially in the privacy world, a lot of critics of digital advertising or targeted advertising say, well, just don’t use any data and that will be fine. Leave people alone, don’t use any data.
But even contextually, the very most simple form of contextual advertising, which is serving a Nike golf ad on the PGA site or a Nike soccer ad on a ESPN site, requires data. And the market is demanding more and more uses of that data to ensure that their advertising dollars are spent the right way. They’re no longer satisfied with, well, the New York Times has a circulation of print media of 7 million people. I know that 7 million people will see my Tiffany’s ad in section A. They’re no longer satisfied with that.
And so the demand in the marketplace is for more and more data to be used, even if we’re not talking about targeted advertising. Sorry, I just wanted to make that distinction.
GUNIGANTI:
No, I think that’s very helpful. And it brings up a question that we got from the audience, which is, can some digital ad assets be framed under a central facility doctrine? So that you would need consent as a privacy requirement for doing the advertising targeting.
But there’s the question of then, you’ve referred several times to gatekeepers of data, how would that affect sharing of information in the digital advertising market, as well as consumer privacy?
FREUND:
Yeah, so it’s a big conversation in the privacy world, under GDPR, many would argue that there are very few avenues for advertising to go forward without consent, the express affirmative consent of a user. Here in the United States, we’ve typically changed the consent requirement, based on the sensitivity of the data that’s used. So there are many types of data that require opt-in consent already, in order to be used. Sensitive health information, location data.
But other more benign uses, which is, I’d rather see a Nordstrom shoe ad than a Nike golf ad, my husband may be the exact opposite. And that makes for efficiency within the marketplace. But I think it’s a really challenging environment to think about dominant companies gaining access to that.
So without the underlying information, it is difficult for advertisers, brands, and even publishers to understand their audiences and to have an efficient marketplace altogether.
GUNIGANTI:
And it’s interesting that you bring up that point about what you and your husband would be interested in seeing as advertising, because with the baseball playoffs, I was watching games a little bit, but I’m not normally a big sports watcher. And so it’s always a little startling to me after being so accustomed to the targeting of online advertising, to see so many ads that are so clearly not meant for me. There’s so much being sold during the World Series, it’s like, oh, yeah, it’s beer. It’s a lot of stuff that I’m not interested in.
I think something I’ve been seeing a lot is some kind of very specialized healthcare service for men. I’s a big shock to see old style advertising in that way that’s meant for a broad market that is assumed to be, in the case of something like the World Series, apparently, mostly men.
So, Christopher, I’d be interested in your thoughts about that as well.
YOO:
So it’s funny when you say a essential facilities doctrine, from the US perspective, one answer is what essential facilities doctrine? This is a court of appeals doctrine. The Supreme Court has reserved deciding it several times on cases where the lower court clearly based it on the essential facilities doctrine. Reading the tea leaves of the Supreme Court is always hazardous, but they’ve been offered the chance multiple times, teed up and declined that at some point saying no means not reaching it, eventually starts to feel more like no, than it does as an open issue.
What I would say is, in addition, there is, to the extent to which it exists in the lower courts, there is a concern about the administerability of it, which is interesting. There’s a more dynamic doctrine in the EU, which just comes off the case, IMS Health, and they really limit it to exceptional circumstances, even there. And that’s the most correctly reporting to the data and related aspects we see here.
And that’s requiring it has to be indispensable to compete, it has cause for a complete foreclosure of a market. And that the refusal to prevent access prevents the emergence of a new product, for which there’s a substantial demand. And so they have a fairly… even jurisdictions, lower courts in the US and Europe that have allowed that to move forward, have fairly restrictive ideas.
And so a lot of this determines the idea of how many other sources are there for the data? And it’s interesting, Leigh is adding some complexity, I think, helpfully. To me, there’s different kinds of data. A lot of it for the actual doing the search stuff, apparently, if you look at the statistics, that sort of really structured data, you acquire very small amounts. They only sample a fraction of a percent of what they do.
The advertising side is very different. To develop those profiles is a different animal. So us just talking even about data, I don’t think is particularly helpful in the sense that it invites us to gloss over some very important differences.
For the kinds of information that you need for advertising, there’s a lot more sources of that, because that’s more demographic information. We forget that the people who use Google also use Facebook, also use lots and lots of other platforms. So there’s multiple points of contact. It’s not as if, just because you have one search engine that you rely on the most doesn’t mean that other people don’t have opportunities to gather the same data.
I did want to say one thing because Ramsi has mentioned it a couple times where he’s asked, with the improved quality of access to information, do we even need advertising anymore? This very much reminds me of a 1990s set of writing that talked about the internet and the death of intermediation altogether. Death of wholesalers, why do wholesaling now?
Because you can now buy direct, or the death of marketing, or the death of record companies is another example. We can see all the content, we can pick and choose. I guess what we realized is there’s a second wave of writing that says, wholesalers, record companies haven’t disappeared. And the question is, either we’re all making a mistake, or there’s functions that they’re serving that are more important than we thought.
So one of the things we learned with wholesaling is, if you’re a hospital, you’ll buy your x-ray machines direct, but you don’t want to buy gowns and silk and masks, all these things, and the scale economies and quality verification, where you don’t want every hospital researching all this by itself. It’s just not a terribly efficient way to do that. If it’s a big ticket item, differently, it may apply differently.
And so what you discover is there are some very important informational functions played by these intermediaries. I would say advertisers are among them, because they’re able to give you ideas and information. I have an article that was really a free speech one that talks about the myth of the internet as an intermediate experience.
And what I said there is, I don’t wake up in the morning and crawl the entire internet to see what’s new. I don’t have that capability. I rely on email exploders, or content venues that I go to, or even search engines to do customized queries to assist me to do this. But in the end, I am relying on tools, someone else’s tools or someone else’s editorial judgment, to help me cull the internet, the fire hose of content.
And oddly enough, the bigger amount of content makes that more important, not less important. And in fact, I depend even more on people to separate the wheat from the chaff, because the quality is more varied, and the volume is more varied.
I think that, I understand what Ramsi is saying, and for certain areas of the market, it might work out that we don’t need them as much. I think in some areas, it actually makes it somewhat more dependent than ever.
GUNIGANTI:
Ramsi, I’d be interested in what you think about this. I’m so sorry about the background noise. Is advertising, can it play this useful intermediating role? And to the extent you spoke earlier about, for example, people just going to user reviews as their way of determining product quality, which I think could be considered a form of intermediation, but it’s not typically considered a form of advertising. It’s considered something else.
So, what is the role that advertising can play in actually being useful to consumers, still, to help them digest the amount of stuff out there?
WOODCOCK:
I wonder whether, if consumers, if we were to go to an opt-in regime for advertising, how many consumers would actually opt-in and avail themselves of this alleged information, continuing information advantage. There might be some, there might be some. But I’d be interested to see whether consumers would do that, or whether they would just do what you many consumers already do, which is when they’re interested in learning more about products, they go to independent news sources about products and learn about them there.
Or run a search or choose to educate themselves in the same way that consumers choose to educate themselves about all things in life. By going to school, and so on. When you want to get a particular product, or you want to know what’s out there, you run a search, you read a newspaper, and you try to learn. And when you’re not interested in learning, you don’t.
The issue with advertising is that we don’t have that option. It’s intrusive and we’re asked to learn even at times when we didn’t choose to learn. Once upon a time, there was a rationale for that, which was, how else are you going to find out about products? You can’t just run a Google search.
But that’s just not true today. We could have a regime in which, when consumers want to educate themselves about our products, they can flip on the ad button, and get all of that, or use some other form. I certainly agree that all this depth of stuff that we’ve been hearing about with respect to the internet, since the ’90s, is overblown.
But I’m very, very skeptical that the information and function of advertising continues to exist in this day and age.
FREUND:
Pallavi, if I could just respond to that just a little bit. I think the interesting concept is that all of the things that Ramsi just mentioned, in terms of the ability for people to go out and investigate products, or things that they might be interested in buying, all cost money to produce. So how are we going to get access to that? And you raised the question, how many consumers would opt into an ad, to digital advertising?
I think, first of all, it depends on the consent and what you say. And second of all, it’s… So sorry about that. Second of all, it depends on what the value exchange is that they’re getting, in exchange for seeing that advertising.
I think sometimes there are days, and as a digital advertising lawyer, I’ll have to admit that there are days when I go on to social media or on the internet, and I don’t even notice the ads. But then there are days when I look and I say, wow, that was really great. My son and I laugh over some of the Progressive insurance commercials.
So sorry. I’m so sorry. This is the disadvantage of being in an office, where my phone is.
So what are we willing to pay for that? Oh, the Progressive insurance commercial advertises it as an art form. It was entertaining. I don’t even know what insurance I actually have, if I have Progressive or if I have Allstate, but the 60 seconds of that commercial will certainly bring a bright note to a day.
So advertising as a freedom of expression, as an art form, as other things, I think we can’t underestimate the value of that. Plus, $125 billion. I think it’s an important industry to our economy, and certainly employs a lot of folks.
I’m here to defend digital advertising and its usefulness even beyond the necessity of obtaining information about it.
YOO:
Pallavi, if I may, can I strike a middle ground, which is, I think sometimes advertising is the right way, sometimes it’s not. And I think focusing on the content is a great point from Leigh. Because advertising, you got to follow the money.
And just to give one very concrete example, I always think about Angie’s List. Angie’s List serves the purpose of Yelp, it’s supposed to vet consumer reviews for different things. And Yelp has a lot of competitors, too. But Angie’s List was once a paid service.
Apparently, I assume they didn’t generate enough revenue, they’ve now shifted to a free service supported by advertising. And if we want these services, they have some latitude to decide which route they’re going to take. I think a world in which we see that, then the question is, in a world where people can choose either model, do we have any reason to believe markets fail? That’s the real interesting question.
And Yelp has a lot of solidarity. There may be some functions there about that. But there are more targeted recommendations that we use, and we see how this all flexes around.
The one last thought is about opt-ins versus opt-outs. We used to think that those defaults were really sticky. What I’m discovering in the anecdata part of this is GDPR now requires opt-in for everything, which has not really changed my behavior. I just click accept a lot more often than I used to. And it doesn’t seem to actually protect many people.
But my colleague Herb Hovenkamp gave the keynote address at the international competition network of plenary last month, opening it up. And he cited a remarkable statistic. He said that Microsoft Windows 10 is shipped with Bing as the default search engine. 93% of recipients change that default to Google.
And so what’s interesting is even if we have these different opt-in, opt-out ideas, that’s really suggesting that there’s real value in differentiation where people have enough to take the time, they’re interested to take the time to change that. And so what I find is, what I like is Leigh’s point, which is there areas where people, consumers vary in terms of their preferences about sensitivity of information. All the data suggests that.
I remember under certain regimes, they even make a big deal of phone numbers. I’m listed in the phone book, a lot of us are, it’s not considered personally sensitive information. I have the option of making it that way. But if it’s not personally sensitive, I would have the defaults follow what consumers want. If they don’t regard it as sensitive, let it be opt-out, if they do regard it as sensitive, let it be opt-in, and we can create this context sensitive, rational way that would actually suit consumer interests.
Because if people aren’t regarding it as sensitive, making them jump through more hoops to get it isn’t actually helping them.
FREUND:
That’s exactly what the NAI code does by the way is, it has that slope of sensitivity. I think it’s a really interesting way of doing it. The issue, of course, is that we have maybe 90% of the industry following the self-regulatory guidelines on that, and the other 10% can do what they want, which is another different aspect of competition policy that’s interesting, is if you have certain market players trying to do the responsible thing, or the right thing in terms of something like privacy, and then the other 10% of the market goes and takes the business opportunities, takes those risks, that other companies are prevented from taking, what does that do to the industry and the market players?
GUNIGANTI:
I think that brings up, specifically about phone calls, brings up an interesting point, because while it’s something that historically, we wouldn’t have thought of as, oh, it’s private, as you said, it’s in the phone book, I think we feel like there’s a lot more exposure at risk nowadays. It once was, especially if you lived in a small town, there would be maybe 20,000 people who would actually ever see your phone number, would look it up in a book, people far away from you wouldn’t do that. Commercial advertisers and so on, wouldn’t necessarily be able to have that easy access to it.
But of course, now with the internet, it’s very easy to get massive lists of phone numbers and to auto call them. This is pretty clearly imposing some burden on consumers. They don’t like it. It’s certainly one of the things people complain about the most to the government, especially to the FTC.
The Do Not Call Registry is extremely popular, but also not entirely effective, I guess partly for the reasons that Leigh has noted. You might have 90% of businesses playing by the rules, but another 10% are very difficult in that respect.
From that privacy perspective, how well do a sliding scale of just the nature of the data versus the level of exposure, how effective do you think that is?
YOO:
So what’s interesting is, there’s a concept that goes back to a Supreme Court case called Lawyers Committee for Human Rights that talked about practical security. And in fact, one of the realities of the data world is our information that was fully public can be aggregated in ways that we now make them more sensitive. I would say the Supreme Court’s recent decision in Carpenter in the Fourth Amendment context is a reflection of this. And there’s another case called Riley v. California, which is a reflection of this.
Which is, you now have your whole life on your phone, which wasn’t the case before. So its social importance can change. I think that there’s something to this. We have to think about what that means.
What I want to anchor us back to make sure we turn those corners though, and this is particularly appropriate in your position, Pallavi, because you’re in an organization that spans two aspects of consumer protection, one is privacy, and the other is antitrust. They sound in very different ways.
So the main focus we’re talking about, about this events teed up to talk about antitrust. And there’s a big case in Germany where they’re fighting about this, and the answer is we’re they’re about protecting markets. And it is possible that a systematic privacy violation policy could affect markets. But if it’s just a one-off violation about a particular conduct, we have different rules to get at that.
So to understand before we bring antitrust to bear, we have to make sure are we just talking about a garden variety privacy violation, where we have enforcement regimes to deal with that? Or is this an antitrust violation where it’s a systematic behavior that represents exclusionary conduct that actually affects markets?
When we look at the consumer protection side, we have to look at the two H store that Ramsi brought up in our conversation before, which is there are good uses of data and bad uses of data. And one of the problems that we have in GDPR and other regimes is by treating all forms of personally identifiable information as suspect, you lose the ability to do certain forms of targeted advertising and certain uses of data.
We saw this in COVID, it actually interfered with some of the COVID responders, had to get a bunch of special waivers. In the context we’re talking about here, it’s going to make advertising supported businesses less viable. And consumers we’re discovering love free, and we could ask them to start paying for search and paying for all these services we’re getting now through advertising support.
Should law have a preference for which business model these companies have to follow? Or is it a harm big enough to justify putting a thumb on a scale one way or the other? Or should we follow the normal course of letting people explore different ways and look at different things?
Because particularly when you look at the free aspects, proving consumer harm on that side of the market gets very hard. And it’s interesting, I think one of the reasons I think we keep moving off that to other things and other aspects, my question is, we do have to remember, as Leigh pointed out, there is a content industry being supported by this, partially supported. Some of it’s by direct payments, and we see with the growth of Netflix and certain video, we’re seeing different pay models show up. And that’s I think great, too.
But to me, I love the diversity of models where different people can follow different things. Because if you force everyone into the same business model, my guess is you’re going to make market concentration worse, not better. Because if you’re going to have the big incumbents on one business model, the natural thing is to try to approach consumers to appeal to them in a different way, instead of putting out a me too product, which has to compete with a much larger, better established player.
GUNIGANTI:
Ramsi, I’d be interested in hearing your thoughts with regard to any current matters or cases, the role of advertising there, and also the perception of free. Because one of the things I think it’s interesting, where Christopher was saying about diversity of business models is that subscription models can leave out some people who can’t afford them.
So the person who can’t afford to pay a monthly Netflix subscription is not completely shut out of video content, because we do still have what is still somehow technically free over the air broadcast TV, although I think it’s getting increasingly confusing for a lot of especially younger people to figure out how to access that. But Ramsi, I’d be interested in your thoughts on this.
WOODCOCK:
When you talk about the information obsolescence of advertising, there’s obviously a concern that there’s so much that we value that’s funded by advertising, in which we worry it would disappear. And that immediately makes me think, well, if these are things, newspapers and so on, which we believe to be socially valuable, but we think they disappear, if they weren’t funded by advertising, then we’re really talking about public goods, right? We’re talking about things that we believe that we need as a society, but which a system, a market system in which people actually pay for them won’t work to support them.
And the natural solution to dealing with public goods is to have some sort of government subsidization or government funding scheme. So in a cultural and a market dynamic such as ours, in which we worry about government subsidization, we turn to advertising as this kind of surrogate or proxy for dealing with the public good. But it’s the least democratic form of proxy you can imagine, because it’s a funding system in which people are paying through the extent to which they’re manipulated by these advertisements that they’re shown without really even being aware that they’re paying, often.
And so to my mind, that’s highly problematic. We’re talking about essentially non-consensual market transactions that end up occurring because of the influence that advertising has on us. I certainly agree that a lot of these institutions, think about the media, need to be funded. But this moment where the information saturation has undermined, I think, advertising’s reason for existing should be a moment when we sit down and reevaluate and ask whether it ever made sense to fund these important public goods, in this peculiar undemocratic way.
But I certainly agree that when you start, if the model, that approach was simply to go just to subscription, and not have government subsidization, you’d have price effects that would have deadweight losses and so on, and price some people out of the market. What needs to be balanced against the price effects associated with advertising and what that does to the prices people pay for alternate products.
Because the $125 billion that are spent on advertising are being taken, recouped, and more, through higher prices charged for the products that are being advertised. So it’s also important to keep that in mind.
FREUND:
I think it’s interesting, when you talk about consumers paying either with their data or in another way, some of the research that we’ve done indicates that consumers feel like they’re already paying for the internet, and don’t necessarily have a deeper understanding of the fact that the content on the internet also requires funding. I think how the market dynamics might be changed, if we move to more walled gardens, other than the elitism, as Pallavi pointed out, that some folks can’t afford access to those things, is that some of the bigger market players still have access to that data. And they will use it in a way that harms the rest, who don’t have access to the data, in terms of the current business model and the current industry.
And so I think who controls access to that data? And under what circumstances that data is used is really interesting. We have the whole data leakage fear among consumers that I think is driving some of this conversation, at least on the privacy side. Data goes, there’s a whole bunch of data out there, and I don’t know what it’s being used for. I think there’s a whole lot that we can do and investigate in our marketplace of digital advertising to make that clearer.
But I think that it’s important that that data access flow be considered when we think about defining those markets.
YOO:
I actually think Leigh is making a really interesting point. There’s a myth about people complaining about having to pay twice. I always think about this in terms of Netflix, which is really changing… We subscribe to Netflix, and all these things, we pay for them. And the question is, we could make it so that we didn’t have to pay Netflix, we just pay, I’m a Verizon customer, we pay Verizon and have them make a side payment for it. I don’t see any reason why we should structure it that way. It seems like a very inefficient way to do it.
But not only that, it means that a bunch of people who aren’t subscribers to Netflix have to essentially participate in that, and targeting people who are actually paying, especially because video is what’s killing the internet, that’s 70% of the traffic right now. Make the video people pay for it. it’s a very rational way to allocate this.
Two other thoughts. One is, Ramsi mentioned public goods. And in fact, all content is a classic public good. Going back to Paul Samuelson, it’s non-rivaled content, you make the content, doesn’t matter if one person watches it, a million people watch it, it’s the same price.
Interestingly, the economics actually say individualized pricing is a necessary condition for the provision of public goods, short of public financing. And even then you got to figure out how much to finance it, but it goes hand in hand. If you’re going to do this by markets, actually, individualized pricing for public goods is actually a good thing.
He raises the possible… Well, maybe you don’t do this by markets. I actually think banning ads and having government subsidies for search and social media, I think is a non-starter. I don’t think politically that would be very palatable, because one, how we pay for it, we’re in a time where we just, our deficit is now 3.1 trillion, well over GDP and probably going to get worse in the next six to 12 months.
Second is a huge one. First Amendment concerns about this, having the government determine what these are like. But the biggest thing for me from a consumer standpoint is innovation, which is government subsidized stuff are not renowned for being very nimble or meeting consumer demand, in ways. And between those two solutions, I think it’s a creative idea and Ramsi is right, it’s a logical one in many ways, but it carries with it some baggage, that I don’t think is going to carry the day.
GUNIGANTI:
Thanks so much for everyone participating in this discussion. There’s so much more still to talk about. I had all these thoughts in my mind about the fact that I just bought a new e-reader. I got a discount on it by agreeing to accept ads on it, and does that indicate exactly how… can we now put an exact dollar amount on my attention, now that I’ve done that?
But I hope that we can all meet again, and hopefully soon meet in person to have these conversations. Thank you so much to CPI for sponsoring this and making it available. As I said, I hope we all get to meet again. Thank you.
FREUND:
Thank you very much.
YOO:
Thank you.
WOODCOCK:
Thank you.
