 I'm going to read from the official notes here, and in my extremely inadequate Spanish. I'll skip the long titles for things. Fernando Vermejo, was I correctly pronounced OK? He's an associate professor of communications at a university in Madrid. And he's also taught there, and got his PhD from another university in Barcelona. And it is MA from the Attenberg School at Penn, as we will say here, and is the author of the Internet Audience Constitution and Measurement in 2007 and Editor of Uncommunicating Otherness, Meaning and Information. And before joining his current university, he was Senior Analyst at the Internet Research Group and Research Analyst at CDNOW, Inc. So he's a fellow with us this year. And without delaying him any further, I'll let him proceed. Well, thank you. Thank you, Doc. First of all, let me say it's an honor to start the 2010 Lancian Series here today, especially with such a nice crowd. Before I actually get into the content of the talk, I would like to say a few things about the topic and the way I'm going to frame it and the title of the talk. The idea of talking about advertising, it always strikes me when I tell anybody that I'm doing research on advertising, they always assume that I'm actually working for the industry or that I'm trying to improve the way advertising works. That's not my goal. I understand that most people do research in advertising. That's their goal. I just try to understand something which I think is a powerful shaper of the internet. And so I just want to understand how it works and what the consequences of its functioning is. The second thing is about the actual frame of how we talk about advertising. I thought that before we make any kind of assessment about what the consequences of advertising are. We need to know how it works and we need to know how it's structured. So that's why I thought about talking about mapping online advertising, which is not easy. It's a much more complex phenomenon that it was offline. So there were two options when I tried to decide how to pursue this talk. One option was to talk about actually this structure and see who are the stakeholders and the different organizations who are having an influence on the way online advertising works and who are the different pledgers and what kind of new things are happening there. My concern with that approach was that it was actually going to get old very soon because things move so fast that it probably six months from now, whatever I said, would be irrelevant. So I thought that it would make more sense to actually talk about the dynamics and the logic of online advertising, which is also probably more difficult to grasp, but also I think it might endure a little longer. So that's why I'm not going to be talking about the different actors and the different companies or structures that are being created, but more about how it functioned as a logic, the dynamics of online advertising. The third thing is about the title. Actually, that idea of anxiety and method comes from a 1967 book by a French anthropologist called George de Vero, who wrote a book called From Anxiety to Method in the Behavioral Sciences. And his whole idea, which is kind of funny, he says in the introduction to the book, is when you look at a researcher doing an experiment with a rat in the lab, that experiment says much more about the researcher than it says about the rat. And so he says that we need to study social sciences and behavioral sciences from the point of view of researchers who have some kind of anxiety when they're confronted with a subject that they need to study. And so the methods, or the pseudo methods as he calls it, they use are actually what it's meaningful about those sciences. So I thought we could basically take that same idea and apply it to advertising and say, well, maybe there's some anxiety going on and there's a method as a result of that anxiety. So having said that, let me see if we can actually grasp the anxiety. I don't know how many of you would be able to actually end that sentence. It's probably the most famous sentence in the history of advertising. Have the money spent on advertising is wasted, the travel is, I don't know, which half. You don't want to make it probably credit as one of the creators of modern advertising. And it's probably the most repeated sentence in advertising and in advertising industry. I think that's a manifestation of anxiety. It's actually the things like we don't know how this thing works. We're actually wasting our money. The funny thing is I've never been able to find the exact date in which that sentence was said, which makes me to, leads me to think that it might actually be apocryphal and it's just a reflection of a general mood of the industry and that's why it's become so popular. Then we move to the method. There's a successful advertising person called Claude Hopkins who said in 1923 that advertising was actually a science. He actually published a book called Scientific Advertising and the book basically starts saying that advertising is a science. We've tried everything. We've figured out how it works and we know how to make it efficient. There is no problem. We know how to do this. So that's the method. Then you go from some guy who says we don't know how much money we're actually wasting to somebody says don't worry. This is all done is control. That's 1923. We go to 2006 to use just one of many examples. That's the marketing pitch of a book called What Sticks, Why Most Advertising Fails and How to Guarantee Your Succeeds. Basically again, we're back to the wanna make your problem. It's like we're wasting our money. We're not being good at what do we do. Again, that's anxiety. So the method, that same year the economy says well this is finally getting better. We're out of the wanna make your era. Now we know how these things work. So there's this constant tension in the advertising industry between this idea that something is going wrong, that we're not doing it properly and the idea that oh no, everything is controlled. Don't worry, we know how to do this. There's actually a very interesting book probably a few years ago by John Durham Peters called Speaking Into the Air in which the subtitles a history of the idea of communication. And he actually claims that the idea of communication can be explained in between the tension of conception that says we're isolated islands, communication is impossible. And another opposite view that says oh perfect communion of the souls and on the minds is possible, perfect communication exists. So he traces a history of the idea of communication by the tension between those two views. I think we could actually trace the logic of advertising by the tension between the anxiety and the methods. So quick question to make sure I'm following is there more to anxiety other than uncertainty? I'm just trying to map it into some of those. Is there more to anxiety than uncertainty, producer? Or maybe you're gonna get into flushing out what you mean by anxiety and what you don't mean by anxiety. I think anxiety, well, you get into like a whole psychological thing about what anxiety is. It's some kind of fear without any specific event that triggers our fears, a more general fear that has to do with future events. And so I think, and I hope as I develop this that what I'm talking about is more of a general, I'd say it's not a specific fear, but there's like a general sense of anxiety. And I would like to explain where I think that anxiety comes from. So that takes me exactly to that they should like, where do I think this anxiety comes from? I think it has to do with the different ways in which economic exchanges take place over the internet in terms of content. We have this model of the gift of the peer production, which has been widely studied here, in which basically people are producing things and they're giving them away. There's no expectation when in a concrete or immediate return on what they're doing is just, so there's no need of being anxious about the end result. You just produce something, you give it away, you hope somebody is gonna actually find it useful, but that's about it. Then you have another model, which is a model in which people assert certain types of rights over the content or whatever they're producing, and they get into some kind of exchange. So I give you this in exchange of something else. You pay me for the use of this. And there's a whole legal structure surrounding these copyrights and all these issues and contracts and I don't know what. So there is expected return, but there's like a whole structure that kind of preserve that return is gonna happen. The problem with advertising is I think that doesn't happen. The model that I've pictured there at the bottom, it's kind of the traditional way in which people talk about advertising. It's kind of there are two transactions. There's like the media have a transaction with audiences and then the media have another transaction with advertisers. Actually, I think that really the way of the victim that doesn't get to the core of the problem with the anxiety of advertising, that's why I think we should do a little more complicated version of that, which is this, basically what's happening there is advertisers are giving something to the media so that the media introduces ads with the content. The media produces the content, make it available to audiences, which pay attention to that content. And the tricky part is that kind of orange arrow that goes from right to left because the issue is you have to get people to go from the ad to the actual purchase of some product or send service. Advertisers are not happy with just having people seeing ads. They just want something else. They expect a return and the return is actually, audiences are gonna be actually purchasing something and they're gonna pay advertisers for that because they produce their products and services and advertisers are gonna give them the product. If you think about it, actually, what advertisers are hoping is that that return they're getting from audiences is larger than the two other arrows that depart from that actually, that what they pay to the media and that what they pay to manufacture the products is actually less than what they're getting from the audience that pays for the product. That's kind of the trick, that's the logic. And so that process, you expect that it's gonna happen but you actually have no guarantee, there is nothing that says that audiences are gonna actually do that. They might actually consume your content and your ad and just don't pay attention to it and do something else. Therefore, you're wasting your money and you have a wanna make a problem where am I wasting my money? Hold on a second. Interesting thing about this from a media's perspective is it very nicely shows how the customers and media and their users are separate populations and there's a fundamental split there and that's been a problem for some time. Exactly. Yeah. So from that anxiety generated by this particular modeling which audiences have to be moved from the content to the actual purchase, there's a several general themes that appear when we try to understand how the advertising industry tries to reduce the anxiety generated by this model. And all of those things that I put there actually exist in the advertising industry of previous media, but I think they take a particular new shape when we talk about online advertising. I'm talking about collecting information. Every process of kind of controlling the flow of the audience or the users requires that we collect information. The second one is metrics. Advertisers need some kind of tangible number that allows them to track what's going on. That's a reduction of complexity actually. From the collection of information which is you collect as much information as possible then you need to reduce that to something which is concrete and those are the metrics. Actually I very much like in that sense the work of James Beniger when he talks about the control revolution, the control revolution when he says when we think about the information society which we think about the information society as something with a lot of information, the abundance of information, but he says in order to understand the information society the key is to understand the reduction of the complexity of the information, how we reduce the information to something that is usable. So the third theme that appears is the theme of targeting actually advertisers. They don't wanna give their money or their product or whatever to people who are actually not gonna respond in the appropriate way. So they need to figure out who are the right people to show their ads to. The fourth is having the right context. They wanna show their ads in the right context and that refers both to a general sense of the environment in which the ads are placed but also about the specific content that is attached to the ad. There's an issue of pricing. There's some kind of transaction here and advertisers need to figure out how much they have to pay for what. And finally there is an issue of funneling that is pushing people from the moment in which they see the ad to the moment in which they actually purchase something. And that's the most complex thing that's probably the largest source of anxiety in the whole process. So let me go one by one through this. Collecting information before with traditional media you basically had to go through a different route. You had to actually put in place some kind of mechanism for doing research based on surveys or panels or some other specific thing so you could collect information on audiences. Online you actually can collect everything automatically. You actually can collect too many things I guess. But that's a big difference. That's a big difference that actually shapes a lot the logic of the advertising process. Another thing that we need to take into account is that when we say that every page, every ad, every video, everything that we do actually can be recorded. That is fine, that is true but it's also true that those are discrete events. So we need to figure out or advertisers need to figure out and other people too how you move from those specific discrete events to actually some kind of identity or some kind of profile. And in order to do that there are certain technologies that we are aware of such as cookies, first party cookies, third party cookies, web bugs and flash cookies which is something that I have to say I'm not a very sophisticated computer person and I only found out about flash cookies just like a few months ago, I had never heard of them. So- How many people here have heard of flash cookies? I used to say it's the first time for me. All right. So that's interesting because if within this room only few people know about this I wonder what happens outside this room. So what's up with flash cookies? Well, they have a hundred KBs of information as opposed to four KBs which is a regular HTML cookie. They have no expiry days, they're storing separate location, they're not controlled by browsers. There's a study published last August by Berkeley by a group of researchers at Berkeley that they took a hundred most visited sites according to Quantcast and they look into the use of flash cookies. They realized that 54 out of 100 sites use flash cookie and only four of them mentioned it in their privacy policy. And even when they went to a network advertising initiative and decided and took an opt-out cookie that says we don't wanna be tracked, actually flash cookies were still employed in the sites that they were visiting. 31 sites of those 100 had an overlap between an HTTP and a flash cookie which means, and most of them came from advertising networks, that means that you can actually, if you delete your cookies, your HTML cookies, they can actually be recreated through the flash cookie. And that happened in several sites including Hulu or about.com. So you deleted your cookies but because you had a flash cookie, next time you went back, the flash cookie would recreate the cookie that you deleted. So that to give us an idea of what's going on in terms of collected information. Also, there's another cool study. Actually, Tony shared it with me, so thank you. From people at ATT, Krithamurti and Willis in which they look about over a thousand websites, some of the most visited websites in the US and look at how many of those websites actually had some kind of mechanism for a third party to collect information from them. As we can see that first graph is kind of difficult to read because the lines are really thin and I just actually copied it directly from their paper. But just to give us a sense, almost 80% of those websites have some kind of third party system that is collecting information of whoever is using that thing. And the one to the right actually shows the extent to which the Google family is actually collecting information from those websites. According to their study, as of September 2008, Google was able to collect information from around 60% of the top 1,000 visited websites. So the results they say in their study show increasing aggregation of user-related data by steadily decreasing number of entities. It's been growing over the years. They've been doing this since 2005. The issue with metrics, as I said, metrics, yeah. Information, I'm just trying to understand what that word information means. Is it just literally mean that you went to that site and that's all they're really getting at that point or what's the bundle of information? Well, it depends a lot on what kind of website we're talking about but the most basic information that these people are collecting are your IP address, your old browser, whether you have any other cookies, that kind of thing. So they're basically trying to create, they're trying to identify users by using some proxy information which is never perfect but it kind of serves a purpose. Would it be like profiling? Yeah, actually, I mean, we're gonna get into targeting in a little bit and that's part of what is going on. Okay. So the issue with metrics, I said once you have all this information you actually need to do something with it and you need to come up with some kind of metrics. The problem with the online world is that because you collect so much information there's so many things that you can actually be measuring and it's not always clear what those things mean. So just to give us a sense of different metrics that have been appearing over the years with different types of uses, some of them have become obsolete, some of them are not used very much anymore but just to get a sense, people talk about heats, sessions, pages, views, visits, clicks, time, users, unique users, unduplicated unique users, reach, engagement. There's so many metrics that are being used and actually it's difficult for everybody, advertisers, media companies, users, to understand what all these metrics mean. One of the most interesting to me and I think I say something about it in the paper that is at the bottom of the announcement of this talk. By the way, these slides are also now at the bottom of that announcement. One of the things is the idea of the visit, I always found that visits are some kind of weird creature. What is a visit? It's something very strange. You take something which is a page, a visit is defined as a subsequent request of pages to the same server or the same site by a specific user followed by a predetermined period of inactivity, which is kind of like a very convoluted version, a definition of a metric. But basically it's just taking something that comes from the press, which is the idea of a page and something that comes from broadcasting, which is the idea of time and trying to mix them both together and say that a visit is a period of time in which somebody do a subsequent request for pages. But I've never understood what the point of that whole metric is and actually the way it's measured, it's always been kind of shaky because you kind of know usually when the visit ends, really. If charge parts were cut. Yeah, actually, if you look at web 1010, that was kind of like one of the visits it was like, you would always be hearing about visits. But it just doesn't make much sense to me. But anyway, it's just a way of trying to make sense of this new environment and adapt the ways of measuring to a new, completely new environment. The next one that the one that is being talked about and discussed a lot recently is the one about engagement. Why? Because when we get into websites that allow a lot of interactivity, there are lots of things going on, people are creating content, are sharing things, are doing things, are talking to other people. How do you measure that? How do you reduce the complexity of that activity to something that actually can be traded? And so there's a huge effort to try to measure engagement. This is the best formula that I found, which I find it great. That's engagement. That's a click index plus recency index by depth index plus loyalty index, brand index. F, I don't even remember what it is. Interactivity index and... We can imagine what it is. ...suscraven. With those scaling factors. There's no scaling factor. Each of those indexes then has a formula behind that you can actually calculate. So if this is what it's been offered to advertisers and marketers, good luck anyway. So that, the issue about metrics. Targeting, what's happening with targeting? It's becoming increasingly granular. It's just going, we would buy audiences and we would advertise to buy audiences and they would buy segments of audiences. Now they're trying to buy individual profiles and smaller and smaller segments of the audience. These are here some criteria through which targeting has been done. There are other, there are many actually criteria because it's just like what kind of rationale you use to decide who you're gonna show your ads to so that you can create as many targeting criteria as you want, but these are the most common ones. The traditional one that has been used in previous media for a long time is demographics. You just look at people's age and gender and things like that and then you just target according to that. The other thing that has been used is content consumed. You try to put an ad that matches some kind of content because you assume if you're gonna try to sell some sports cards, you assume that people who read car magazines are gonna be interested in that. So you just match the ad with the content, not necessarily with the demographics of the audience. The thing with that is that online that can become much more granular because that you can actually go into specific pieces of content. You don't need to talk about a medium. You don't need to talk about a magazine. You can talk about a specific text and you can target the ad according to that. By the way, one of the problems with YouTube is that that's more easily done when you're using text than when you're using images. So that's one of the struggles that YouTube has. The next level, which is very much peculiar to the internet is that you can target according to the content that people produce. So if people go to a social network and they start typing things and saying things then you can just try to target them according to that thing that they're typing. That's also kind of complicated so far but the version of that that it's been working very well as we all know is the search keywords. Actually, when we do a search we're actually creating some kind of content. We're doing something. We're typing something. So that's been used very successfully by Google and by other search engines to target advertising and that's working really well. And the next level, that line there points where it's what is peculiar actually to the internet. The next level, the one that is being talked about and discussed a lot recently is the issue of behavioral advertising which actually the ads are targeted based on all kinds of behaviors on the part of users and that's why we need to collect as much information as possible from users because the more we have the better we're gonna be able to target because the more we know about their behavior. In order to show that, well this is something that I'm gonna skip, the targeting segments is really funny how they're creating more and more targeting segments on actually you as an advertiser can go to an ad network and tell them, no I wanna create my own profile. So you wanna, you can define a segment that you wanna target by giving them a certain number of criteria. Just to see that on the general graph that we put before, we can target according to the audience demographics. We can target according to the contents. We can target according to what the audiences, the content that the audiences are producing or we can target in general through all the information that we're gathering from users and that's behavioral advertising. The next level, having the right context that refers both to the general environment in which we're placing your ads and a clear distinction here would be it's very different to put an ad in a search engine but people are searching for something that will put an ad in a social network where people are probably not looking for anything in particular, they're just trying to communicate with their friends. That's, those two environments are very different and therefore advertising will prefer one or the other according to their interest. Obviously the search seems to be much more appropriate if you wanna try to sell something. The other thing is the issue of specific content. Advertising is very careful as of where what kind of content they're associated with. So in previous media, you could have some kind of, if you were gonna advertise in a newspaper or in a TV station, you could, I mean, you could assume that they're not gonna do something where your ad is not gonna appear next to something very strange because they're all these norms, professional norms and organizational norms, et cetera, they're gonna prevent that from happening but when you're talking about online, you have to be careful because there's lots of, for instance, user-generated content and then you can end up showing your ad next to something that you don't wanna be associated with and that's why there's all these efforts to more or less clearly or subtly control the content, sorry, the content that is being associated with particular ads. So, for instance, Technorati and the new Incarnation, it's much more careful about what kind of things they're aggregating and what kind of things they're showing and what kind of things they're ranking because they wanna create a more friendly environment for advertisers or YouTube is very careful as to which kinds of certain type of content being too high and too prominent in their website. That's part of the deal. And the other side of this is the creation of content specifically for advertising purposes, which is, of course, something that we say, well, that also happened in previous media. We could just, let's create a sports magazine because we know that all these advertisers are interested in getting in touch with these audiences. That's always happened. The thing that is happening now, it's becoming much more granular, it's becoming much more specific and it's becoming much more sophisticated and probably the best example is Demand Media. Demand Media is a company that is aiming to produce, if I'm not mistaken, a million pieces of information like every month or something like that is absolutely crazy. They do, what they do is they link, they search for what kind of keywords people are searching through what's the price of those keywords and what kind of content with those keywords exist out there. And this is all automated. So they have this whole system in which they're going around, checking the prices of keywords, checking which keywords people are using when they search and checking what kind of content is out there with those keywords. And they decide whether it's worth for them to actually build a piece of information, be it text or video, that matches those keywords and therefore can get advertising. And they're doing it very successfully. There is a wonderful piece on them in the wire magazine in October, I think it was. The next issue is pricing. How do we price advertising? In previous media, the kind of standard way was CPM is cost per thousand cost. So you pay for a thousand members of the audience and that's what they give you. Online, there are new ways of doing that. CPM still is 38%, this is data from the interactive advertising bureau for 2009. So CPM, in which you're basically buying audience, is 38% of online advertising. However, there's a model that has been growing every year online, which is the one based on performance. That is, you don't pay just to get an audience. You pay only when the audience does something else. The most common and obvious version of that is a click. So in Google, people, advertisers only pay when actually somebody clicks on their ads. So that's a cost per click model. That model can actually be expanded. It's not easy in certain instances because it's kind of difficult to sometimes do more than click on an ad, but you can go all the way to a purchase. So you can say, okay, I'm gonna put this ad here and I'm only gonna pay if actually I get somebody to buy my product through this ad. Of course, besides the idea that that's difficult in certain contexts, there is an obvious resistance on the part of the media, on the part of content producers because when you're selling audiences, that's a product that you can manufacture. Actually, you can control, you can produce audiences and then you just sell the audience. And then it's up to the advertiser to do some kind of appealing ad and to sell some kind of appealing and interesting product and convert those people into customers. The problem with this is that more and more advertisers are saying, well, we're not willing to pay for audiences, period. We want to pay for something more. So we want some guarantee that actually something else is happening. And part of it is obviously been triggered by the success of Google and their cost per click scheme. But if we look at it again on that general picture, we can see that pricing used to be just in terms of how many members of the audience I'm getting. That's what I'm paying for. And now it's moving more and more towards the actual sale. We wanna only pay if we get what we want, which is to get the audience to go and buy a product. So that's what is happening. Funneling, so after you've shown the ad and you show the ad to the right people and you have information and everything, you need to make sure that actually people are gonna go and purchase your product. And of course that's the most uncontrollable part. That's really complicated because people can do whatever they want, supposedly. So there are several ways in which there are times that funneling people to that purchase. One of them, of course, is the issue of branding. If you spend money in advertising and you promote your product, your worst tragedy is that somebody is gonna go and say, oh, I actually need that product, but it's gonna buy it from a competitor. That's the worst thing that can happen to you. So that's why branding is so important. You just have to make sure that they're not only gonna buy the product, but they're gonna buy your product with your brand on it. So branding is a very important part of this funneling thing. You just wanna make sure that people are not gonna end up going to a different brand. And that's why there are all these issues about preserving the value of brands and making sure that your brand is not used in the meta tags of somebody else's website or somebody else's advertising because then that's gonna actually divert people from your brand to other brands. So the whole idea of branding is part of this kind of funneling process towards purchases. There's another phenomenon that is taking place which is the issue of tracking and retargeting. The first time I heard about these retargeting themes, what is this? So it's just an idea. These are quotes directly taken from the websites of Yahoo UK and AOL US. Yahoo UK says, we know where users have been and we can tell where they're going. That's the first thing they show on their advertising page for advertisers. We know what's going to happen. And then AOL says, well, don't worry. If they saw your ad and they didn't purchase your product, we'll follow them and we'll retarget them again. We'll just remind them that actually your product is there and then they should buy it. So that's another thing that is happening. Finally, another movement that is happening is in previous media, the whole finally process was almost impossible. You just watch an ad on TV and then whatever you did after that, the advertisers had no control over it. They had no idea what was going on. They didn't know whether you actually bought the product because you watched the ad. There was nothing. The thing with online is that at least in certain instances allow you more accountability and more ability to control and to track that process. So a way of improving the finally process is by trying to move more and more activities to environments in which you can actually do that tracking. And I think next year what we're gonna see is a big push in this whole online, I mean, mobile advertising. I think that's gonna be the next level. If you look at the recent movements of the big companies such as Google or Apple, they bought companies in the mobile intelligence area, AdMob and Quattro. And Google, the same day that they announced their new smartphone, they reminded all advertisers that actually they have something called a paper call, which actually if you advertise on their smartphone and then people just, you can advertise for free and then you'll only be charged when people do place a call through the phone to you to purchase what you're advertising. So basically it's the same idea that they're using for the click, but just apply to the smartphone environment. So that's an interest. Now, the problem is with this. So what are the consequences of all this logic? And that's I think the tough part of advertising. When we're talking about other kinds of control online, the consequences are usually very obvious. If somebody is censoring a website, that's very obvious. You just like not allowing people to use that website. If somebody is stealing your products and you know somebody's stealing my product. But when advertising is involved, the same way that the control process is very diffused, the consequences are also very diffused. So it's always hard to figure out actually what's going on. I've tried here to kind of summarize what I think are some of the consequences of the different elements of the method that I've been explaining. In terms of information, the more information, the better. If we are doing especially behavioral advertising, it's even better because behavioral advertising is based on knowing as much as possible from people. So let's collect people. In terms of metrics, advertising is not comfortable with things that kind of be easily measured. And that's why it prefers certain environments instead of others. It prefers certain engines instead of social networks, for instance. Targetings, it tries to get more and more personal. So it's trying to get more as a direct relationship between the advertiser and the actual users rather than going through the media. And so the media are losing part of the role. It prefers again certain environments and it shapes content because there's certain content which is more prone to advertising and certain content which is not. Pricing, it tries to reduce the risk by asking other people to run the risk. Before, there was no advertiser would pay for audiences but they run the risk that the audience would actually end up purchasing anything. Now, they wanna make sure that actually they only pay when somebody's doing something. And in terms of funneling, they try to be more intensive and more extended, meaning they try to retarget and follow you and target you again until you purchase and also try to expand to new arenas such as mobile phones. Some general consequences that I see which of course are with a big question mark because this is, as I said, the part that is more complicated. Privacy issues, it seems that the advertising industry is playing this kind of cat and mouse game in which all we do this to collect information, all they found out about this, let's figure out something else. Oh, let's move here. Oh, be careful that they're not gonna come and regulate this whole issue. So it's kind of playing this constant game. In terms of content, the de-intermediation which is a horrible word, but I think what is happening here is actually that what we usually consider to be the media is on the losing side. If you look at, for instance, the two forms of advertising which are being really successful nowadays, one of them is search engine advertising. Content producers are not getting their share. It's just the search engines that are making the money. So they're collecting all the information, they're making it, they're ranking it, they're making available, but they're actually getting the cut and getting the money. So the media is on the losing side on that end. On the other side, the other part that is being successful is behavioral advertising. Why? Because the cost, the CPMs are much higher if you attach behavioral information to people. Again, if you are a newspaper online, what kind of information do you have about your users? It's very limited. You have information about what they do in your site, but you don't have a behavioral profile of those people. In order to create a behavioral profile of those people, you actually have to sell that information to some entity which is above you which collects information from many different websites and therefore is able to create a more deep and detailed profile of people. So again, you're on the losing side because you don't have anymore like that product that you could directly sell which is, I have this much audience, which these demographics, you need to have behavioral information which can only be gathered from many different sources. Finally, the issue of structure, advertising as asteroids, I think that we should, one of the things we should think about is how when some company or some environment is able to figure out how to make money with advertising, it actually becomes very powerful. Everybody can think of Google, of course. They figure it out, they know how to do it, and they are probably the most powerful force shaping the internet these days. So the moment you get that advertising asteroid, then you become like a really powerful figure. By the way, it's very ironic, I should say maybe, that the only, I come from the field of communications. So years ago, I read this paper by Bremen Page in which they present what the Google search engine is like. It's the first academic presentation. It's called like Anatomy of a Search Engine or something like that. The only communications book that is mentioned in that particular paper is Ben Bagdikian's Media Monopoly, who actually says advertising is just a force towards monopolies. They use it in a critical way, but I wonder whether they think the same nowadays. And finally, just as a way of conclusion, I think the online content layer, you know how Yo-Hai and other people talk about the physical layer and the logical layer and the content layer. Well, I think the content layer can only be explained through these three different logics. There's the logic of the gift, there's the logic of the pay, and there's the logic of advertising. And we need to pull each of those three threads and start untangling this whole mess. See, we want to understand what's going on with online content. And that should be it. Thank you. Thank you. Thank you. Thank you. I want to get questions from the crowd, but I want to ask one quickly to begin with. Sure. Do you think advertising is a bubble and are you at all concerned about the level of dependency that the world has on Google, which has a 100% level of dependency on advertising alone? Let me see, the other day I heard of a Spanish writer who says that he refuses to go out to any talk show because he always comes up with the right answer like two minutes later and he feels terrible. So I'm afraid that's going to happen with your question. I'm probably, maybe that would be the best solution. But... Pretend I asked this two minutes ago. No, let me see. The whole issue of the bubble, I know there is this whole idea that actually, again, advertising doesn't work and this is just like, this is going to disappear because it doesn't make sense and things are different. But still, I think it was Seneke, one of the founders of American sociology who said when human being is considered and something is real, the thing is real in its consequences. So if we think that advertising has actually an effect, then that has consequences and the consequences are probably that we still see an increase in the amount of money being put in advertising online regardless of whether we think that actually, things should be different or that actually that that's not the way that we want this thing to be. And as for the whole Google, see, I don't know, because I think that has to do with the whole idea of the asteroid. It's like, okay, once you get some system that it actually works very well, it kind of creates this kind of virtuous cycle thing in which it keeps on growing and growing and growing and I'm not the one here to talk about regulations or monopolies or things like that because I'm sure there's lots of people in this room that know much more about that than I do. So I would not pursue that route. Interesting, the title of your presentation because it seems that we started with anxiety and we touched the matter, but then we are back to insight. Just to be sure that that was your conclusion. Yeah. Because it's true that the market needs to operate and I saw many millionaire agreements being closed with IACPs exactly on this basis and from visitors to purchase to pick through because they couldn't track the purchase effectively. So that happens and actually this is what is keeping alive many of the IACPs we see nowadays. So how much at the end we always gonna have this anxiety because there are not perfect methods and there is still no perfect method from TV either. Yeah, actually one of the slides that I hear before until I think last night was something that said I'm back to anxiety. After all the method, let's go back to anxiety because I mean, there's always anxiety in terms of like, all right, we're collecting all this information from people but people might be deleting their cookies and so what do we do? Or actually one of the funny things that there's some reports out there is like, oh actually a cookie is not a person. So how many occasions you can be sure that actually a cookie identifies an individual person and there are many instances in which it's impossible because there are different, is the issue that they call device per person and persons per device. If several people use the same computer or the same people use the several computers then you kind of track them with this kind of system. So there's an inside, there's like a new source of, or you can have issues with fraud. It's like, oh, but you know, actually media are not giving us the impressions that we think they're giving us because the whole technical structure is so complicated now. Before if you put an ad in the newspaper, the first thing you did in the morning was go buy the newspaper. Just to make sure that your ad was there. And then also you would have some kind of independent source measure in the audience of that newspaper. So you could actually kind of be sure that you were getting what you paid for. But now because this is so complicated that you get these reports from ad net worth saying, oh yeah, we showed it and this is what happened, but you don't see it. So there's always a source of anxiety in, and I think partly it's because the final goal would be perfect control and that's not possible. That is, there will be no anxiety if I as an advertiser or as a producer of something knew, okay, this is gonna be sold this way to this many people and that's it. So you can actually have a concrete and very clear explanation. If I show this ad to this person, this is what's gonna happen. But you kind of control that. And those feedbacks mechanisms, you bought these, I don't know, X size and then there is always that question, where did you see that? Do you see some efficiency on that type of feedback or no, do you know anything on that? Meaning like? The questionnaires that the seller send you after you close it down online. So give us feedback on your purchase experience. And there is always a question, say, where did you see that product? Well, yeah, that's part of trying to figure out exactly like what is the process from the moment in which somebody sees an ad, to the moment in which somebody actually purchases. So they're kind of trying to trace back your route to try to figure out actually where you're coming from and what is more efficient as a way of advertising. I actually thought that Tyler's done some work on the networks of ads and how in fact, there are many, many different hops between advertising firm and website. Yeah, well I haven't done much work directly in that but I know where people have, is there a relationship between ad networks and area of more fits that are changing over time? I was just trying to say that in fact, this relationship is hard because you don't have direct contracts with the person who is actually putting bits into a bit stream. I had a question if I could jump in. It seems that for 100 years, 60 years I'm watching Mad Men now, people are very happy with the anxiety that they have and there was large marketing budgets that drove the classic area of American commerce. Why do we demand less anxiety from an online context and why do we not put a message out there whether it's subways or whatever you're doing. My message is out there, my brand is out there talking about why do we demand greater accountability and method in online space and why aren't we happy with the idea of oh, just flash my logo and I'll be happy. I think because there is the possibility of having more accountability. Who's that is? So, but I guess that doesn't make sense in terms of if I'm managing investment at a rational level, just because there's the possibility all I need to do is make sure I'm getting as much accountability out of this money as I'm out of that money. Yeah, I think the problem is probably with the word rational. I mean, it's not that, for instance, I've done lots of research on the structure of the audience measurement industry. And I always find this thing, when I try to explain this to people, some people would say, oh, but they make it up. Some other people would say, oh, but this is like a exact scientific procedure. Well, actually it's always in between. They never make it up and they never measure it exactly. And the industry, there's this Leo Bogger who said there's a need in the media and audience industry to believe that the numbers are actually real and there are no shadows at the end of the cave. So it's this kind of need that actually there's something tangible and that actually we can run with it and I think in the whole business world including advertising, oftentimes the evidence, the data, the whatever is just a way of calming the anxiety but it's not actually getting to where you wanna be. So if you have, if you're showing ads on TV and you don't have any guarantee that actually or any proof that actually that is working, then you shouldn't be demanding the same kind of, I mean, you shouldn't be demanding a different kind of guarantees to a new medium. But the problem is that it seems that the new medium might be able to offer some of that. So everybody wants to take as much as possible from it. And I think that's what it's happening. So it's an actual path dependent story that just the people doing TV buys have always been doing TV buys and so no one's gonna touch their budget but the new guys have to actually prove something that the old guys don't. I would think that that's partly what is happening. Just a quick lady question. I'm assuming the old way of measuring success was just, okay, I put my ad on TV, how much does sales increase? Does that no longer work or does it still work? How has that been showing success of online advertising? Yeah, see the problem with that is that there's always this kind of like conflicting evidence about the effectiveness of advertising. If you get Michael Shatzen's book, Advertising the Uneasy Persuasion, he basically says, well, there's no real proof that advertising works. All that might happen is actually you're shifting customers from a brand to another brand but actually you're not increasing any kind of consumption of any kind of products. There is no real evidence about the effects of advertising. But then if that is the case, then one has to wonder, well, why is there so much money being put here? They must have some kind of evidence that is actually working. But again, the evidence in this whole area at times is not what we think is evidence or it's not used in the same way that we would think it's used in science or anything like that. I don't know. Just a reformulation of the same problems except the precision with which we can reformulate the questions is proportional to the new amounts of anxiety, something like that. As in we can measure more and there's, which goes back to what you were just asking, because we can measure more, there's that much more anxiety. Yeah, I mean, I think the anxiety doesn't disappear because you have a more efficient method of what seems to be a more efficient method of tracking and following and measuring. There's always new sources, there's always new things and actually the pure complexity, I think the pure complexity of the way advertising works online, I think just that in itself generates a lot of anxiety. Everybody can understand, every advertiser can understand what happens when you put an ad in a newspaper or in a TV show. I mean, it's very straightforward. The problem is just understanding what's the life of an ad online and how it's placed, what it's placed and when and to whom is so complicated that that in itself probably creates anxiety. Advertising a good result, excuse me, always involves uncertainty, there's some level of uncertainty, it's always gonna have anxiety. Yeah, and that's what I was trying also to show by comparing with other models of providing content. If you're just giving something as a gift away, then it doesn't generate any anxiety. But you could be anxious about it. Well, yeah, you could always be anxious about everything. But it's just like a different mindset and if you're just doing an exchange of things, I mean, if you just sign in a contract and say, okay, I'm giving you this in exchange of that and you give it to me and then we have all these laws regulated then there's no anxiety there. But the problem with advertising is that actually expect a return, but there is no specific way of guaranteeing the return. Can I suggest an issue about this in terms of risk and investments? It takes something like affiliate marketing. Take a publicly traded company and they do affiliate marketing, which really means that they're offloading the risk on the affiliates. And it is risk-free. It's anxiety-free for the publicly traded company to do so because they're only paying for the actual sale in the end, okay? So they're getting rid of the anxiety and they're parting it on, or they're getting rid of the risk, they're parting it on to the affiliate. And the affiliate may also take the risk, but the affiliate may be happy to take the risk. Now, why would a publicly traded company want to pass on the risk because they have to produce the numbers on a regular basis for their investors and so on? So they want to get rid of the risks, so they are anxious, but there are plenty of people or organizations out there who embrace risk, they're not anxious. So just like you can have people buying bonds or they can go in and do venture capital activities, you can do the same in the advertising world. And so there is a space for anxiety in some places and there's a space for risk taking in other places. Absolutely. I mean, I think the issues of control and risk are always very close tied together. And I think it's a perfectly reasonable way of putting this. It's just talking about risk and sharing risk. Description of the content kind of generating a context that is appropriate for a particular piece of advertising content. If you look at the privacy policies underneath these online service providers like Google and Yonah, you'll see that they explicitly make clear that the information they collect is under their privacy policy and all of the advertisements you see, they're not served directly by them, are under a completely different privacy policy that you have to go seek out on your own. So you've now got this kind of, kind of got this convergence towards a particular, what looks like a nice homogenous content of you on the surface, but underneath you actually have possible divergence and privacy implications on what you thought of that kind of. That is, I mean, I think it's happening at many different levels. For instance, one of the things that study that I showed with the graphs, I showed how much, in how many websites, third parties are actually collecting information. For instance, one of the things that Paper discusses is how more and more what it's happening is that these ways of tracking users through third parties is actually being disguised as first party information, which is kind of probably the opposite of what you were saying, but it's in the same line. It's just, well, if you decide you don't want any third party to track you, there's no problem because the first party is getting you a tracking system that is actually gonna end up in the third party. So there are all these different ways in which you can actually, that's why I said that the whole issue of playing kind of mouse, it seems like there is this constant game of, well, if you try to stop this way of doing it, we'll just figure out a different way of doing it. Yes. Here's a hypothesis for you to react to, which is that, in fact, we are out of the age of anxiety that what you've described is the end of advertising anxiety. So we scope that a little bit. There's still anxiety that your product may suck and nobody ever wants to buy it and you're advertising will suck and we'll feel get it right and you'll fail. That's eternal. But when it comes to advertising, well, you're paying per click for sale, per transaction, per lead, and so you know that you're advertising, that you're paying for something as opposed to, as Ethan Zipperman, our beloved friend and colleague, pointed out a few months ago, it may be that newspapers are cratering now because the web has revealed that the old print form of advertising where you couldn't see what the effect was, it was, in fact, entirely built out of paper and horribly overpriced, and this new model, the web one, is in fact much closer to the real value of advertising. So you have ads that you only pay for because you're getting some value from and there's a market mechanism, an auction mechanism that means that the price that you're paying for those ads are also, in some sense, market rational or reasonable, and so your anxiety is not about the advertising model, it's only about your business, which is where your anxiety should be. Therefore, this hypothesis says we're having back on from anxiety the method and not back into anxiety, but out of anxiety. So, let me try, let's imagine that you actually control perfectly the whole advertising process and you're absolutely successful with your ads. You could think like, why be anxious? Well, I think there might be the anxiety of saying like, well, maybe actually these people would have bought my product anyways without me advertising, maybe what I'm doing is actually wasting my money. And I think that's one of the things that when advertising becomes very, very efficient, it runs the risk of actually preaching to the choir and just trying to convince whoever is already convinced. So there's always that thing, it's like, okay, if I actually show my ad to a thousand people and those thousand people end up buying my product, which is something which is probably never gonna happen, but let's say that happens, then the next thing you're gonna say, your boss is gonna come and say, you wasted our budget, because those people were probably gonna buy it anyways without us advertising. So there's always some kind of anxiety that could come up, I think. Following on the same point, I remember quite vividly talked by the Christian, one of the people who works on designs and these sorts of ad options and the way he framed the problem from the beginning, says, okay, if we take as a given that there's a fixed budget that an advertiser has to spend and fix what they wanna optimize, one of the metrics, a number of clicks, then everything is a science in the sense that they can analyze what is rational for what keywords to bid on and so on, and also design the options in an appropriate way, but there's two choices there that I think, as I understand it, are the source of your anxiety. One, which is what you were just talking about, is that still doesn't answer the question of how much of a company's budget should be spent on advertising and maybe it should be zero if people were gonna buy the product anyway. And the second is how does that final metric, number of clicks or whatever, translate into the actual increase in revenue? Yeah, absolutely. I mean, I think I made a parallel before with John Doherm-Peter's book about the idea of communication and it's the same tension. It's the idea that you can actually put yourself into somebody's head and figure out how they're gonna go about their behavior and how they're gonna end up purchasing something. It's probably, I would say, impossible. But so there's always some kind of anxiety about that. So if people click on the ad, does it mean that they're actually more interested that if they hadn't clicked or what does it mean actually that they're clicking on your ad? Does it mean that you're gonna get more purchases? How is that thing? And then the other part is like, okay, then if everything runs smoothly and efficiently and perfectly, then why are we doing this? Why are we actually spending money in advertising? It should be like a fixed cost. Long-tail advertisers should be more confident at less anxiety because there's less possibility that somebody's gonna buy their stuff through other means because they don't have other means because they're long-tailed. So this may, the anxiety curve may be different for long-tail or head of the tail, head of the curve. I don't know if, I'm not sure I understood that. If you're one first company making decorated bongo drums, the chance of anybody knowing about you is vanishingly small, except if you advertise and furthermore, the difference in sales would be far more dramatic there if the ad succeeds because you go from zero to something. And so you should have less anxiety to further down the tail you are. Whereas if your GM or whoever, Ford Focus is getting pretty good coverage just through natural PR and media stuff, much harder to track. I see, I see. Yeah, that might actually make sense. So similarly, there are short-term kind of spontaneous purchases as another long-time thing. If I'm driving my crappy car down the highway and I see a BMW poster on the side of the road and I buy a BMW 10 years from now, it's gonna be very hard to extract whether that ad had any value or not. And that's still lingering out there. And I guess the question is, are there things that are only bought in the immediate term? And if so, then you have a much, much better of the chance of measuring that. Absolutely, the kind of product that you advertise, it makes a big difference. There are certain things that, as you said, I mean, you probably, the cycle from the moment in which you are aware or you see an ad or you think about it to the moment in which actually you do the purchase and where you do the purchase and how you do the purchase, it can be very, very difficult to track and follow. While there are other things that are much more immediate and therefore the usefulness of this whole system might be much higher, absolutely. You think that the value of brands is gonna decline as this kind of old world advertising where you wanna get yourself into people's heads in anticipation that at some time they're gonna buy a new television or TV as opposed to saying I'm only gonna spend money when there's a direct connection from point to purchase. It seemed like in that world, it doesn't matter if you're Coca Pepsi, it's whoever's gonna offer you the best cola at the best price at the best time. So is corporate America maybe making a mistake by getting so obsessed with getting every penny's worth out of their advertising dollars? Yeah, one of the interesting things about, for instance, what the type of advertising that Google does search-related or advertising based on display of banner ads is that actually the banner ads have more of that thing that television spots used to have, which is that kind of branding and creating some kind of value and kind of at times convincing you that you actually need something that you might probably not need. While the other type of advertising, like Google advertising, is much more in like, oh, I'm already in the mood of looking for something, so I'm just gonna, oh, I happen to see this ad, I'm just gonna go for it there. The value of brands, of these kind of like suggestive advertising values and content is much slower. So what I think is probably, there's gonna be a balance between those two things. I mean, I don't think brand advertising is gonna disappear. I don't think branding is gonna be less important or it's not gonna be important in the future. I think it is. But also, this whole idea of getting a brand a whole idea of getting people the right ad at the right moment when they're already kind of in the mood for something, that might actually increase also a lot and create a lot of efficiency in the consumption and production process. The idea of efficiency, it's great for our consumer or environment, it's like we'll just buy what we need to buy and not do tremendous amounts of impulse purchasing. That's a little bit of the Google side ones, but it seems like a huge amount of the advertising industry in general is really meant to drive consumption and to create desires that you didn't have to begin with and to do a lot of this branding and positioning and everything and that the money behind it is much less in the large sense about the, I'm looking for a pair of garden shares and okay, oh here's the nine places I can buy those garden shares. It's very useful and it gets this big feedback, but it doesn't, I mean I think this is, personally I think it's a good thing if this would just disappear entirely. But I think there's a funny side too that we are, is there a way of giving people some measure of how free the web is not? That if everything you're doing is being supported through having your desire for consumption raised, how much are we paying for our content through unnecessary consumption? Yeah, I mean I think you're both are touching some of the core issues around advertising. There's like, there's a whole debate forever about whether advertising is just an issue of information. Actually we're just informing people, you know there's this product available, these are the characteristics of the product and this is the prize. Yeah, so just like, and so that's a view of, oh advertising is like that and there's the other view, no advertising is not about that, advertising is about making you want things that you don't want. So those two things can actually coexist and at times one is more powerful than the other. The issue is in that part of like more like persuasive advertising in which they're trying to convince you of something. You actually might end up paying a prize. The thing is like in that graph that I showed before, the whole end result is that what you as a consumer pay for whatever you purchase has to compensate for whatever the advertiser paid to the media and for whatever the advertiser spent in building or producing that thing that you're purchasing. And so the more you actually do this kind of persuasive advertising, the more you're gonna get from that process as an advertiser and the less we're gonna get us. Doesn't seem kind of perverse, given all the problems that different media on the web have had with getting people to pay for content is that you can get people to pay for content by making them buy all kinds of things, almost anything else in the world they will buy but not the content they wanna be reading. One of the things that I thought of before preparing this talk is like I'm gonna try to stay away from the kind of like big cultural side of advertising that is like what kind of effect has advertising in our culture and in our economy as a whole and what kind of like perverse results might come out of this whole advertising process. I think it's a very important topic. I think that will call for at least another talk or two. So we might come back. But yeah, I mean it's a very relevant issue. It's now one minute past the end so unless it's a pressing question somebody has to ask, we'll thank you very much. Thank you. Thank you.