The Google Podium: transcript

August 9, 2006
Search Engine Strategies Conference

Q&A

Eric Schmidt

David Krane (Director of Corporate Communications, Google):		Welcome, thanks for coming. We're going to spend about the next 45 minutes with Eric. Very informal, the agenda is your agenda, not ours. On the record questions on any range of topic that you'd like and Eric [will respond]. So I'll clarify anything Eric said [in his earlier remarks] or answer anything else that's on your mind.
Q:		I'm with Washington Internet Daily I cover public policy issues. I have two questions, please. How does Senator [Ted] Stevens' description of the Internet reflect on his qualifications as one of the most important policymakers in the world on communications issues? Number two, as specifically as possible – what do you see as the top legislative and regulatory challenges or problems that you face the next 12 to 18 months and why and what's the outlook? And to the extent it might be different, what are your top priorities in terms of what you'd like to accomplish from the local level to the international level?
Eric:		I'd rather not comment on what a specific politician says, or how he characterizes the Internet. From our perspective, governments are grappling with the implications of having their citizenry online, and in most countries there are appropriate laws and cultural restrictions and so forth to handle all of that. I think that it's reasonable to say that we at Google will have, perpetually, people working with government leaders all around the world to try to make sure they understand what the Internet is, the impact of the Internet, its benefits, etc., etc. Our highest priority right now is the net neutrality issue, which I'm sure you're well familiar with. The reason net neutrality is so important is that it's one of the fundamental principles about the way the Internet actually works. The funny thing about net neutrality is that as a company, we have enough money that if there were a hybridization of the Internet, as has been proposed, we would be able to fund it – we would be able to handle those expenses. However, one of the key aspects of the Internet has been that the standardized way in which everyone has been treated enables the next generation of innovators, the next two very smart graduate students out of Stanford or Berkeley, or whatever. And so that's why we've taken such a strong position there.
Q:		So I was wondering if you could comment, please, on your keyword policies, specifically how they relate to small businesses and your mission. When you test Google print ads or Google ads to radio, you're touting that [they're] going to give access to the small business, to the local restaurant. But now under your philosophy if AdWords [ads] get low clicks, you claim that that's irrelevant to users. So small businesses are kind of blocked out. Just to put in perspective – let's say there was a disease that affected five people in the world, and one scientist that created a resolution for that. I believe under your current system, if only five people were interested in clicking on that ad, you would deem it not relevant to users. But if close to five people whose life depended on it – if [small business] livelihood depends on supplying that product, it would also be quite relevant.
Eric:		I'm not sure I agree with any of the premise of your question, so we'll have to explore that. The way our system works is that we accept ads and bids from essentially anywhere. And if the ads are not clicked on, or if the prices are below some very, very low threshold, they are deactivated. The scenario that you described where there's an incredibly important cure for five people – would, in fact, be ads that were clicked; they would, in fact, create a market; and we would, in fact perpetuate those. In the case of radio and the other things, the same principle applies. Most of those markets are established and mature and, by the way, quite successful advertising markets, sales professionals and so forth, but they don't reach everyone. One of the things about Google AdWords and AdSense is that they have democratized access to advertising for the little guy, if you will. And that little company – this is sort of the long tail argument that so many people have made – now has access on an equivalent basis to an auction, and access to the same customers.
Q:		People read the article today about AOL [inadvertently revealing search query data for its users] and say, “Oh my God, Google or whatever is going to give my searches away and everything about me is going to be known.” Can what happened at AOL happen at Google, and what do you say to the searchers out there?
Eric:		Well, our number one priority is the trust that our users have and that would be a violation of trust, so the answer is, it won't happen.
Q:		About the software that you're releasing – oh, by the way, I'm the number one “Chris” on your search engine, so thank you.
Eric:		For the moment! Enjoy your fame. [Laughter.]
Q:		You have a range of software products and unfortunately, they all look completely different because they're obviously developed by different teams. Is there a chance that at some point in the future, to unify the UI and the experience that people have with those various products?
Eric:		So Larry, Sergey and I talked a lot about this. My background had been to build these integrated solutions, and they said, “Eric, you're wrong. You're better off having each of the teams build a user base and if they develop a user base, then it's possible to integrate them.” And I fought with them over this issue for a long time. And as usual, they were right and I was wrong. Because their argument is actually very simple. If you build an integrated solution that nobody uses, it doesn't matter if it was integrated or not. So we've taken the position that we want to release products quickly. We want to get them out there, we want to see if they work. And I don't mean work technically, I mean – do they solve a real problem? If they [do], it's relatively easy to produce an integrated solution. And so, indeed, we're doing that now. So the answer to your question is, yes, and it's a big priority, but it's in the context of a strategy which is to get the stuff out so it has users, we can get feedback, find out what we did right. And we're not perfect, we iterate, we learn from things.
Q:		[Re the possibility of profile-based ad targeting given deal between Google and MySpace and other Fox Interactive properties]
Eric:		We do not link search results to the advertising system. They're two separate systems. I hope that's clear. That was a clear separation. We operate from the standpoint of user permission, so virtually all of the things that we're going to do that are personally targetable will be done with user permission. You're using very precise words. “Search history” has a precise meaning at Google. Let's just say that we're not going to use demographic information and specific personal information without the user's explicit OK.
Q:		[Follow up question on ad targeting]
Eric:		Again, you're using a very specific example and I don't want to comment on partly because that's a MySpace decision, not a Google decision. In our agreement [with Fox Interactive], we have agreed to follow our principles, but they also have to decide. So they might, for example, not allow it for other reasons. I'd rather not parse the outer layers of the deal, because again, the deal itself is confidential. We model very carefully in MySpace a combination of things. It was text ads, other kinds of ad formats – and remember, there's also a [Google] Toolbar distribution. MySpace users will have an integrated search experience and will, therefore, search on Google more. It is the sum of all of those that make the economics work. And that's why we're so excited about it. Plus, wherever we are today with respect to our targeting improves. So when we improve our targeting, our revenue improves. So it's the sum of all of that. Plus, the fact that the MySpace and IGN teams have committed to working with us very closely on a technological basis to provide good end-user benefits. That's the reason.
Q:		[Question on extent of click fraud]
Eric:		As part of a litigation, which was settled, we produced a technical report which actually analyzed this, which is now public information. We also have chosen to disclose what we estimate is the bad click rate on a per advertiser basis. We're doing these [things] because they're the right things for advertisers. I've also been misquoted by a number of people as somehow saying that this was not an important issue. It is an important issue that is under control. We have very good technical people. We have very good computers. We're monitoring it. It's not material to the company. It is important, and it's not going to go away. I would not characterize the problem as having gotten worse – I think it's gotten more attention. And I think the attention is good, by the way. Because we want people to know about this stuff. It's part of how that system works. It's hard to know [the precise extent of click fraud] because we only know what we detect, but based on our estimates, the problem is manageable. And we have a lot of ways of detecting it, but I can't make you an absolute guarantee, because you never really know.
Q:		[Re mashups using Google products]
Eric:		Well, the most entertaining mashups are all the ones involving real estate. Because basically we knew when we bought Google Earth, that whole team, that people would use that information in interesting ways. But it did not occur to me that so many people were looking for apartments, because you never see it in a visual form. So to me that was a really positive surprise. There is a very funny mashup where you can play golf on the golf course from the satellite [view]. So there's fun stuff like that. And what it has taught us, I think, is that there is tremendous creativity on top of this mashup platform that's yet to be unleashed. And I think we should encourage it.
Q:		[Re the possibility of Ajax “obliterating” traditional ad metrics such as page view – and whether Google is developing new metrics]
Eric:		That's a very, very interesting point. I don't think I fully understood it until you asked the question the way you did. It is true that Ajax is a change in the page view model. Ajax is so much more powerful – for those of you who don't know, Ajax stands for asynchronous Java Script with XML. It's quite extensible. I think in many ways you see the old model falls away, and a newer and more sophisticated model replaces it and I think that is the story of technology. There are always new ways of measuring, and I think we'll have to find some ways to accommodate that. It's an example of something we need to anticipate from advertisers.
Q:		Hi. I just want to follow-up on the AOL [privacy breach. Did you speak to AOL about this?]
Eric:		I did not contact them on this because I was busy doing these other things. I've been deal mode, unfortunately. So the answer is I did not.
Q:		[Follow up question asking his opinion of the eventual outcome of this AOL breach]
Eric:		In many ways, it may be positive because we want people to know things. We don't want people to wandering around saying everything is secure, everything is protected. So I think that the awareness is positive. The specific is obviously bad. So please don't get my message wrong here. It's clearly a bad thing. No one disagrees with that, but I think more awareness is good. From a Google perspective, again, this would not happen. I don't want to criticize AOL, they're a good partner of ours, and they should answer the specifics. I don't know enough of the details. It is awareness, awareness that people know what goes on online, that is a good thing. The fact that people now understand that credit card fraud – again, please don't take this out of context. This is important. I don't want to be quoted as saying “this is a good thing.” It's not a good thing. Awareness of what can happen, that bad things can happen, is a good thing. I hope that's clear.
Q:		Awareness of what can happen anywhere? At Google?
Eric:		No, to be precise – online, especially with things like identity fraud and things like that.
Q:		[Question about Google's growing role in various media and content partnerships after less successful earlier efforts at this]
Eric:		First, a year ago we had a few very smart people working on this area of content. Now we've actually built a management team, we've hired some great leaders who have experience in it that I, for example, did not. So we have a more sophisticated business operation. So in that sense we're able to approach people who are very sophisticated in their own business, but don't know much about ours. And we can approach them in a way that we can actually do business together and not screw things up. So in that sense, you're seeing the maturation of a management team understanding the strategy, working with people. We've always wanted to expand our advertising reach and our advertising network and monetize other forms of content. So for example, we've always wanted to monetize video and online communities. You have to wait until you have both the partner – because we don't do those things ourselves – and also the technology. We're now at the point as a company where we have both better targeting technology, and people who understand how to target media. We have scale, we have the systems and the solutions, as well as the business negotiations. So I would say the answer to your question is yes. We are, in fact, talking and have been talking for a while to many of the online communities, many of the content companies – as we should. And we're open to exactly how we're going to work with the folks as these businesses evolve. In fairness to them, their models are also changing, and we're learning from them. So what happens is, one of our executives would meet with some of the senior executives in those firms, and they will explain that their existing model is being affected by Internet, and how can we help them. They don't perceive us as hurting them. And that's a good thing.
Q:		[So is Google actively pursuing more of these deals?]
Eric:		The answer is yes, we're pursuing them. I think the highest priority right now is not more deals, but implementing the deals we've announced. I hope that's helpful. There will be more deals, but it's easy to get into deal mode and forget you actually have to deliver. So internally we always say, “Well, great job, now get back to work.”
Q:		[Further questioning about content partnership deals]
Eric:		Well, let's talk about it as an advertising business as opposed to a partnership message. We started with the premium ads business which was text-based, and we added our AdWords auction, then we added AdSense for content (which is now a very, very successful business for us). And along the way we learned how to monetize the web very, very well. We set out a strategy 18 months, 24 months ago, talking about working on these other aspects of advertising. We analyzed: how big is the advertising industry, where is the low-hanging fruit, that kind of stuff. There are some obvious ones – working with magazines and newspapers. Those are folks who could clearly benefit from more targeted advertising if our technology would allow. We are in the process of building that business. We try things, we do a new approach, we buy inventory, we do targeting, those sorts of things. In the radio business, of course, we bought a company called dMarc, which looks like it's going to do very, very well. Those [audio] products come out in the next few months as integrated ad products. We've already now mentioned two new ones that I think are quite notable. With Viacom MTV, even before [CEO] Michael Wolf had joined them – it's a wonderful story. He suggested to us that it would be possible to take video that was content – owned by content people, and precede or append video ads systematically, and then do targeted advertising for video online. This was Michael Wolf's idea. So here he is, president of MTV, so he calls me and says, “Why don't we try to build this product?” Which I think is just wonderful. So that's an example where the partnership really paid off. Now we announced that deal on Monday morning. If that deal works, and of course we're just beginning the trials of that, then it will offer a model that's at the scale of AdSense for Content, AdSense for Search, for video and multimedia, because it will be possible to have people make video. Look at all the cameras here in the room. People are constantly using videos to film one thing or another, to find a way to monetize them besides just distributing them for free, or selling them with a credit card. The other big area is online communities. I think people here in the audience know how large these communities are, and how quickly they are growing. Our evidence and our measurement indicated that the MySpace folks were, in fact, growing fastest and were the largest, although I'll let them talk about that. And so that's why [we made] that deal. We've been talking to them for a long time. So I think that's it helpful to say those are three or four areas at various stages of maturation. In all of the areas we need partners. We need partners for the content – because we're not a content company. We need partners to educate us about how the business works. What do we know? We're just building the platform.
Q:		[Followup: what content platforms and ads targeting approaches are working?]
Eric:		So first place, we're doing very well in search and content. We've been doing the print stuff during this calendar year. It seems to be doing well. It's not growing very quickly yet; we're trying to figure out how to get it to grow quickly. In the case of radio, that's supposed to come out in the next couple months. Looks like it's going to be very successful for us. The other two that I mentioned were just announced and they're just beginning. Was that a crisper answer?
Q:		[Question re existence of “parallel Internet” – e.g. RFID, etc.]
Eric:		I'm not sure I agree with your characterization that it's a parallel Internet. As far as I know, all of these are still part of the Internet, and they're just new devices that are showing up online. So I don't think it's different, I think it's just part of the Internet. To the degree that those devices are web searchable – they use our site maps product or they're HTML indexable, they'll show up naturally inside of Google.
Q:		[Follow up on whether sites/devices use search engine friendly descriptors]
Eric:		Well, it depends – but that's a question for the manufacturer. That's the question for the site owner. We would obviously encourage that information as much as we can. We are not ourselves doing specialized deals for them, if that's your question.
Q:		[Question re content targeting technology]
Eric:		We have developed some very sophisticated targeting technology, which is derived from the targeting technology that we use for AdSense for Content. I think those of you who have followed us know that the AdSense for Content pages are actually read by an AI engine and to use the targeting for the ads. A similar technology can be used in these online communities. So we do, in fact, use some of the content for targeting.
Eric:		One [example] is MySpace, the other one is Viacom. In the case of Viacom, we can do the targeting to the closed captioning. We have not yet, and I've not yet seen, technology that can see the frames of the video and figures out it's you in the video. But if we have closed captioning or other metadata about the video, then we can do targeting. And that's our strategy.
Q:		In the other content deals, what would Google bring to the table to content owners? Will it be appending and targeted AdWords, would it be distribution, would it be searching the content?
Eric:		Yes. It's all of that. Remember, our contribution is minor. The majority of the work is done by the content producer. Our job is to get that content to the viewer and all we are essentially a distribution engine with this specialized targeting that can product revenue. Now it's minor, but it's central. The majority of the revenue, the majority of the work, is still done by the person who makes the content and produces it.
Q:		[Question about advertising minimums or guarantees in partner deals]
Eric:		The numbers are much, much larger [today]. I mean I had a cow when in the original AOL deal we put in a guarantee of a small number, of millions of dollars. So today, of course, the economic structure – and the forward commitments are much, much larger. But the business is so much larger, the scale, the global operation. Larry and Sergey and I talked about this, and we said, in many ways the company has more assets. We have a lot of cash, we can take bigger risks. We can grow into what we believe are going to be very large new businesses and there are strategic benefits to being first, especially if you can move quickly. That's part of our leadership culture.
Q:		[Question regarding press criticism or negative reviews of products]
Eric:		Well, again, we get our feedback from our end-users on every search, on every use. So with all due respect to all of us, my opinions, and what you all think – we collectively are not the judge – the user is the judge. So when we bring out a product, or a new service, we look at how end-users respond to it. So the question about AOL, the real answer is, it's an end-user trust argument. So we will do things that are specific to maintaining or improving end-user trust. And that's how we're going to sort all this out. There will always be people who criticize Google and that's fine. We think the criticism is healthy, we learn from the criticism. It's all perfectly fine.
Q:		[Regarding the Kinderstart lawsuit on biased search results]
Eric:		It's probably really important that I not comment on a specific legal question, since we get sued to death around here. Maybe I could answer your question more generically. We work very hard to have algorithms that are independent of bias and we try particularly to not have manual intervention. So I hope that's an answer.
Q:		[Regarding the AOL privacy breach …] On one side you said it's a good thing that people understand that's potential fraud out there. On the other side, it won't happen here. Now why won't it happen at Google?
Eric:		Hang on. Since we're all on tape, I retract my previous statement because it was obviously confusing and I apologize. “[Privacy breach is] a bad thing. It's a bad thing.” How am I doing? “It's a bad thing.”
Q:		Why won't it happen at Google?
Eric:		Because we have systems in place that won't allow it to happen. Again, I'd rather not reveal the internals of how Google operates in these areas, but we have very sophisticated security plans. Let me do it in as a general answer. We have very sophisticated security plans for attacks on information, and the reason we have those plans are twofold: one because it's the right thing. And then the second is because of Sarbanes-Oxley. And in Sarbanes-Oxley, you're required to have corresponding manual controls to avoid the bad actor scenario in many of the systems inside the company. So we know it for two reasons. And of course we're 404 compliant.
Q:		Do you think that nationwide Wi-Max will have any effect on the network neutrality debate and the follow-on? What kind of pressure is Google going to continue to put on Congress as network neutrality sort of flounders?
Eric:		We, along with many other companies, are working hard to make sure that Congress does not do the wrong thing. And historically the Congress has done relatively little on the Internet, and it's been the right thing, and that's fine. In this particular case, by virtue of all of our work, I think we educated people into some of the bad things they were doing and those things have already stopped. I think at the moment everything is sort of on hold. So the first question had to do with nationwide Wi-Max. If you look at broadband deployment, more broadband is good for end-users. More broadband is good for Google. So anything that we, Google, can do to promote the adoption of broadband, broadband choices, whether it's DSL, cable, Wi-Max, any other – it's all positive. For example, narrowband users search less than broadband users. Narrowband users click on the ads less than broadband users. Now correlation does not prove causation. But we know that broadband is good for Google, we know it's good for the world, and so forth. So that's a generic answer. With respect to the specific on net neutrality, we have not yet seen a technical proposal of how to bifurcate the Internet. And so it's hard for us to judge technically whether the other side's proposals have any chance. But if I could just pile on for a second to make a point, historically, these multi-tiered solutions have not worked because the technology improves so quickly that when you establish a differential pricing tier, all of a sudden the cheap tier has the same performance as the expensive tier. Now I'm not in charge of those folks' business, but it's a bad idea. And we've taken this position and we feel very strongly about it.
Q:		[Question re Internet access, broadband and net neutrality]
Eric:		Well, you'd only do those sorts of things in a defensive move, and we've not had to do that. The reason is that it's very capital intensive. The last-mile problem. The folks that are busy fighting that out have huge capital needs. It's a very, very expensive business. Even the Wi-Max stuff is reasonably expensive. So from our perspective, it's much better to have other people build those businesses and then we can be the attractant. One of the things missed in this discussion – if you look at the economic structure in the telecommunications industry, the vast majority of the growth is in people adopting broadband. And we're assisting that in the sense that people want to use Google on their broadband connections. So we do have alignment of interest here. People make more money on broadband, it's new, it's exciting, it's good for the telecommunications industry, it's good for us. So we should be able to come to an agreement on this.
Q:		[Question re perceived UI differences between paid and objective results]
Eric:		It may be helpful to just review how Google does this. Google has two sets of results. There is what people call it algorithmic results, which are not affected by advertising. You can't buy your way in and so forth. Your question is, should we do more to promote that? I think most people know it, but I'm sure we could do more. It's certainly very disclosed. The earlier [question] was about a lawsuit where somebody is claiming that's not true and, in fact, what I just said is true. With respect to the ads, you can absolutely pay more to get your ad in a higher position. We run an auction, which is our business. The two [search results and ad placement] are separate. I hope that's a clear answer.
Q:		[Would Google consider having an ombudsman role?]
Eric:		The ombudsman idea is a good idea. It's interesting that in this audience these questions are coming up. I don't hear these every day – is this a specialized problem, or is this really a systematic problem? Most of the advertisers I talk with are very, very happy with the way we operate. On the fringes, there are situations where people are trying to game the system. We have very clear terms of service, and we do police it, and they do get mad and send us hate mail. There are probably cases – and I'm personally familiar with a couple because I had to talk to the customer, the advertiser – where they had a third party do something, and the third party convinced them – that this would improve their advertising performance, their ranking, but in fact it was a violation. We then stopped working with the company and the company said, “God, these guys at Google are really, really pains in the ass because they won‘t work with us,” when in fact it was a third party's fault. So I'm familiar with those situations. That's an example where it's very, very important that third parties fully disclose our terms of service. There's more clarification we could do on why we're doing it, I'm certainly open to specific feedback on that. We benefit by being incredibly clear of what is an acceptable advertiser behavior and an acceptable website. And if anything, we are tightening the criteria because of the concerns that were expressed earlier about overall quality. We don't want people just doing aggregations of ads to make arbitrary disintermediation of money and their ads and stuff. We don't think that's a healthy thing. And we do reserve the right to make these decisions because this is our business.
Q:		[Google reaction to Microsoft IE 7 search preferences?]
Eric:		So far we have not seen a clear and crisp answer to our concerns. The Microsoft platform products are in beta, in various stages, they're in beta for awhile, and Microsoft does make changes to them. So we would hope that they would address the concerns that we have. The specific concerns that we have had have to do with what appears to be favoritism with respect to changing [search] defaults and so forth. Their general counsel recently gave a speech where he made some claims about what they're going to be doing in the future and so we'll judge their actions based on what he said, versus what they actually do, and we'll see the answer. We will look very carefully at what he said and we'll see.
Q:	[Why do you say you don't understand how page rank works?]
Eric:	There is a small group in the company that understands the very specifics of how ranking works. I chose to not be a member of that group, because I think it's so important to the company. It's sort of like the formula inside of Coke. Only a small number of people hold it…If you're in my position is that you can make mistakes. It's just better not to know. Since I'm not actively programming in that part of the code, I don't really need to know. What I can tell you is that the specific combination of factors, which include much more than link farms, is a huge competitive advantage. And it's very sophisticated. It's not three things. It's a very complicated formula. And we're very proud of it.
Q:	[Question re trademark issues concerning the inclusion of Google as verb in OED]
Eric:	We sent them a very humorous letter [not to use it as a verb], which you can read. Use Google [to find it]. Or should I say, please Google it? Sorry. Delete that. [Laughter]
Q:	[Re progress of Google Talk … Are you pleased with the uptake?]
Eric:	Google Talk was a new entry into an established market. And so we shouldn't apply the same standards as it applied to them as they are to Google Talk. Our hope with Google Talk is to fully integrate it into the online experience and search experience at Google and all of the things that you're seeing now with Gmail integration and with the other applications is part of that. It's not the same thing as the other products. It does different things. I should also mention that Google Talk uses XMPP and other standards in the Internet, whereas most of the other competitors are proprietary. And if you go back to my ideas about “don't bet against the Internet,” we have the, hopefully, correct belief that eventually open Internet standards win over a long enough period of time. That is our view.
Q:	So major advertisers are working at eBay to build what they're calling the eMarket, eMedia Exchange Auction System to buy traditional media. Do you see that as competitive? It seems like that could really compete with Google to facilitate auctions with traditional media. Are you talking to them?
Eric:	I'd rather not talk about specifics with our conversations with eBay because there are lots of them. I don't think there's going to be any single one way in which media, online media, Hollywood media and so forth, come and get monetized. I think there will be many opportunities. You can decide to pitch these things as competitors, or you can decide to pitch them as I prefer – as coexisting explorations. In other words, there's a sort of a new tack here and there's a new tack there. It's not obvious that the success in one prevents the success in another. Many of you continue to focus – this is going to sound negative – I don't mean it – to focus on this pitched battle between competitors. And I think that analysis is just not correct, if I may. This is a very large market, and the entry of new competitors seems to be good for the existing competitors. I'll give you an example. Microsoft recently announced that they're entering the [ad] space. They're in the process of doing that. Is this good for Yahoo! and Google or is it bad for Yahoo! and Google? Now in traditional economics, that would be bad, because there's a fixed supply of things and there would be more competition. But in fact, advertisers don't advertise on any one solution. So Microsoft will bring more advertisers in, who will then choose to advertise on Google and our other competitors. Clearly good for Google. Furthermore, even if a new entrant underprices us, it will still be economically of interest for that new entrant to come to the other sources, because the way the auction works. So the auction produces a competitive value and as long as the new model is a more effective way of reaching advertisers, people will shift to it regardless of the number of competitors, if you follow the reasoning. So it's a very different model. And everyone is obsessed with this, but in fact, the building of this ad – when I talked about this, don't bet against the Internet model, the cloud-based model with the new advertising, new monetization, is a historic event. It's of the size of the building of the PC industry or the mainframe industry, which I'm obviously familiar with. It's of a size of the media industry just in terms of its scope, because it has such reach worldwide. And there'll be many players, many approaches. Google will not be the only one, as much as we might like to. We have no expectation of being the only one of any of these things nor should we, because it's not a zero-sum game. It's not a platform where there's only one winner, which in some other businesses it is.
Q:	What's Google's strategy for classifieds?
Eric:	Well we have partnerships with a number of partners – some of which are announced, some of which have not been announced – where people who are in traditional classifieds are putting their ads into Google, the ad network – essentially into Google Base. And that seems to be working well. So that's roughly where we're starting – we're going to see how that goes.
Q:	[What has been the most financially successful initiative you've introduced since AdSense?]
Eric:	Well, the thing that drives our revenue the most has been product and quality improvements in ad targeting. Since you've covered us as an analyst, you know that our revenue has outpaced that, and that outpacing is due to the fact that our computers are getting smarter with respect to ad targeting. It is that intelligence that has created the high gross margins, the cash flow and so forth, which enables us to then invest in these new areas. So I think that's the secret, if you will, and it's a beautiful thing.
Q:	[How far along is Google in developing these ads targeting improvements?]
Eric:	We don't know. We're clearly nowhere near an end. The question is, are we 5% into it or 25% in it? But we're not toward the end. If you look today as a user of Google and you look at the advertising and the advertising product, you can see there are clearly improvements that can be made for targeting. And every time we make an improvement in targeting, the value of the ad goes up, overall revenue goes up. That's one argument. To give you a different argument, which is also worth stating for the record – regardless of the pricing in this industry, in advertising industry, remember we're competing against traditional advertising. So if you think of yourself as a vice president of sales or marketing, you have an allocation problem. Do you allocate to the online advertising industry? Do you advertise it to traditional media? Do you allocate it to your sales or so forth? As long as it's on a cost per revenue dollar, literally on a cost per acquired customer, our industry is lower – the money will continue to flow with us. So a lot of people keep asking us, what about the upcoming recession (always an interesting question in and of itself, how somebody can predict a recession). But if you assume there will be a recession in the future, it's not obvious that an upcoming recession would materially affect us, because remember, we're benefiting from a secular shift from lower performing to higher performing sales spending dollars. So the business of our sales force – and I suspect of our competitors as well – is to convince people to both grow the market, but also to shift dollars to it.
Q:	Do you think there'll ever come a day that those – that the advertisers will realize that those gross margins are due because they're bidding their own prices up and maybe want to go back to the traditional model of negotiating rate cards down?
Eric:	No, I do not. I hope that's also a clear answer. And here's why. The model that we and others, but in particular Google, have built, is fundamentally more efficient. There's less overhead, there's less manipulation, there's less opportunity for error. It's much more scientific way to advertise to your customers. As a result, we get scale and we get value. As long as the scale and the value over performs relative to the other choices, which indeed it does today, we will continue to get more and more of a share. Now I'm not suggesting we're going to get 100%, so please don't conclude that. But at the current rate – you all would know better the size of the current online advertising numbers as a percentage of the world's advertising market. Do you have a number, Safa?
A:	About $12 billion [online advertising].
Eric:	So we'll use Safa's numbers, $12 billion out of an industry that's $500 billion, plus or minus. There's a lot of room for growth, of shifting. Again, I want to be clear, television advertising is not going to go away, radio advertising – traditional forms are not going to go away, newspaper advertising – but there is a shift as well as aggregate growth. The studies that I've seen indicate that the majority of the new investment is going into online advertising, precisely because of this ROI argument.
Q:	[Is government intrusion forcing Google to give up data – internationally as well as in U.S.?]
Eric:	Well, I'm offering a general comment that the information is of value, and there is always a balance between the interest of the individual and the interest of the state. That's why we have a Constitution. We , Google – in the United States. We look at it on a country-by-country basis, and the reason I mentioned the government is that we already had such an example occur with us where we were successful in making our argument in a Federal Court. So if you think of it, that's evidence that it might occur again in some other context.
Q:	[When will Google's ad auction model extend to video ads?]
Eric:	We're going to use our auction system to power that. So that's what the deal is and, again, we have to build it, we have to deliver it and so forth. But we think ultimately it will produce more accurate pricing for the value of these currencies, because auctions tend to produce more accurate pricing.
Q:	When is that coming out?
Eric:	As soon as we get it delivered, but very soon. I mean literally we're working on it right now. Very soon.
Q:	[What are the biggest cultural impacts Google and Internet have had on society?]
Eric:	Well, I think the Internet is changing everything. The Internet is changing everything and I'm proud to have been associated for a long time, I'm proud to be at Google, which is one of the companies that's involved in doing that. First, there's obviously a tremendous improvement when people have the access to more information, especially when it involves, for example, people who are not telling the truth – like governments or people – [where they are offering] misstatements of one kind or another. I think it empowers the press, it empowers individuals and so forth and so on. So the overall message of that information is very strong. The development that to me is most interesting is the development of these new social networks as online lifestyles. I highlighted MySpace before. The evidence is that especially young people, the demographic is sort of college, in their 20s, [are spending more time online]. And it's true not just in the U.S. Google has a product called Orkut which is successful in some countries doing something similar. And that's a really new phenomenon. And our evidence is that it's a very big phenomenon. It's certainly of a scale of the phenomenon of instant messaging five or ten years ago. Remember when instant messaging just sort of took off and became this hundreds of millions of people kind of a phenomenon. The social networking broadly defined is such a phenomenon. And obviously we want to help empower and make that even more. People are choosing this and they're choosing this for some really interesting reasons. For those of you that are not part of these social networks, I would encourage you to spend some time really understanding it, because it's a big deal. It's not a transitional issue, it's not something that's just going to go away. So with that, I am really done. Thank you, guys.

END OF SESSION