Danny:

I'd like to welcome all of you out here today. We're doing a little bit of a late start, but I figure after yesterday's Google Dance you could probably use a little bit of extra time. We are joined here today by Eric Schmidt, Chairman and CEO of Google. Thank you very much for coming out and being with us here today, Eric, we appreciate it.

Eric:

Thank you very much.

Danny:

I'd like to say hello to everybody who is listening to us live on the Internet. We're going out through Webmaster Radio FM, so if you're tuning in out there, hello to all of you as well. And now we're going to dive in with in the questions.

So in preparing for this, I was talking with a bunch of people at Google about things that would be interesting to address, things I was interested in, things that you're looking at doing. And they talked to me about how you're developing this idea of trying to help people not "bet against the Internet." And I almost threw up my hands and said, who's betting against the Internet? Does somebody think we're going to shove the genie in the bottle or change it? So tell us more about this.

Eric:

Well, in the first place, thank you for having me. Congratulations on an amazing conference. This is really very much the center of the Web and of Web advertising today. I really appreciate you all coming and spending a couple of days, and I hope you enjoyed the party. And we certainly did as well.

It may be helpful to put architecture and the Internet in something of a historic context to start with, and say that 20 years ago everybody was in love with a concept called client/server computing. Many of the companies that are here in the [Silicon] Valley and in the industry as a whole were built up around a model where you had a PC client and then a set of servers, Unix, OS2, Windows, etc., and a lot of proprietary protocols between those. And many of those protocols and architectures are still present today.

What's even more interesting (and technical people never really think about this) is that there was also a business model built to fund and hire all those engineers to build all that. And the business model was largely invented by Oracle. It was a direct sales force that would go in and sell complicated software to enterprises that they would integrate and do important business functions.

What's interesting [now] is that there is an emergent new model, and you all are here because you are part of that new model. I don't think people have really understood how big this opportunity really is. It starts with the premise that the data services and architecture should be on servers. We call it cloud computing – they should be in a "cloud" somewhere. And that if you have the right kind of browser or the right kind of access, it doesn't matter whether you have a PC or a Mac or a mobile phone or a BlackBerry or what have you – or new devices still to be developed – you can get access to the cloud. There are a number of companies that have benefited from that. Obviously, Google, Yahoo!, eBay, Amazon come to mind. The computation and the data and so forth are in the servers.

This is the same talk that I gave in this room 10 years ago about something they called the network computer – which, I can assure you, none of you are using, because it didn't work. But in the last 10 years, technology and architecture, things called Ajax and something called LAMP, which is Linux, Apache, MySQL, PHP, PERL, Python, and all of these new platforms have come out. In looking at it, the analogous thing that happened to make this possible that I certainly didn't see 10 years ago was the development of advertising in this new forum.

And so what's interesting is that the two – "cloud computing and advertising – go hand-in-hand. There is a new business model that's funding all of the software innovation to allow people to have platform choice, client choice, data architectures that are interesting, solutions that are new – and that's being driven by advertising. The reason that I said "don't bet against the Internet" is an awful lot of people are still trying to do stuff the old way. They're still trying to build proprietary protocols, they're still trying to not build standardized protocols. They're still not trying to solve problems in a simple and extensible way. But when somebody does it right – let me give you the example of mashups, which are taking over the world by storm. It happens very fast. And that's the power of the Internet.

Danny:

One of the other things you mentioned, the advertising [that] helps [Google] go out and fund all these things – search has been this huge driver. We continue to turn to search. We confess our innermost desires to it. And this week, on Monday, AOL had released a large amount of search data. And the intentions were honorable, they released the data so that researchers could see how people were searching. And they thought that they had protected the privacy of the people because they replaced all the user names with tokens or anonymous numbers. But people quickly discovered that because I could find the profile of a particular person that we could track it back. And The New York Times [published a front-page story about a woman whose identity was easily found this way] – this woman has become the poster child of search privacy. So she's kind of probably the most innocent person of some of this data that they could pick out. But it was very revealing. But it raised, this whole issue has raised all sorts of issues. Now we've had queries where you could see someone doing searches that looked like they were planning to murder their wife. You've got searches where you think people might be doing illegal stuff, and just searches that are embarrassing. And it comes back to what do you do to protect this [data], to ensure that it's not going to happen accidentally out of, say, Google, or on purpose, in terms of we are trying to get research data, but we want to help, or to prevent the government from coming in and saying "well, this is great. We need to go in here more and see these records, because look at what people are planning."

Eric:

Well, [this sort of privacy breach is] obviously a terrible thing. And the data as released was obviously not anonymized enough, and maybe it wasn't such a good idea to release it in the first place. Speaking for Google, we exist by virtue of the trust of our end users. So if we were to make a mistake to release private information that could be used against somebody, especially if it could be used against them in a way that could really hurt them in a physical way or something like that, it would be a terrible thing. We have lots and lots of systems in the company to prevent that.

It's funny that we talk about the company being more transparent. But there are many things inside our company that are important that we don't share with everyone, starting with everyone's queries and all the information that that implies. I've always worried that the query stream was a fertile ground for governments to randomly snoop on people [for example]. We had a case where we were only a secondary party, where the government gave us a subpoena, which was in our view, over-broad. And this over-broad subpoena we fought in federal court – one of the great things about the American system is that you can actually have a judge make an impartial decision. And the judge ruled largely in our favor. So that's an example of how strongly we take this point.

Danny:

One of the worries has been that even with the protections, the longer that the data is out there, the more likely that it can come out in some way, shape or form. And to my knowledge, Google's policy is that you simply don't destroy it, that you do protect it. Does something like this [AOL matter] make you think, gosh, maybe we really should just go through and say we will keep this data for a month or two months, and then we will destroy it as a way of protecting it from ourselves or others?

Eric:

The information is in a sufficiently non-transparent way that it would be difficult to have somebody steal it. And so we're reasonably satisfied – and we obviously work hard to make sure that this remains true – that this sort of thing would not happen with Google, although you can never say "never." Again, the more interesting question is not [about] accidental error, but something where a government, maybe not even the U.S. Government, but a non-U.S. government decided that it was time to get in there [to our systems]. And we do worry about that.

Danny:

But beyond the idea of the privacy of the searching that we do, there is growing concern about the privacy of what people can find out about us through search engines. I mean you've personally had this sort of experience, where you didn't really want that sort of thing to get out. Everybody has encountered this thing, sometimes perhaps in horror if they've done an ego search and [then] they say, how did I get information? And when I've been asked about it, I explain it's not particularly a search engine problem, it's a problem that these public records are going to be out there with our without you. But it's fair to say that search engines have suddenly become a much easier conduit to [this information], that you no longer have to be a public investigator or newspaper reporter or some super researcher. Do the search engines need to do more to try to pull back on the public information as being released from other sources, or to act as a middleman to at least filter – to make it harder to get to through them?

Eric:

I hope that the outcome will be good judgment on the part of all of the players. In California, for example, there were laws that said that all the DMV records were public, which of course included people's home addresses. And so an actress was stalked and killed because a terrible, terrible, idiot criminal murderer used this information against her. And it's not okay. So that's an example that not everyone in the world is perfectly honest and honorable. There are crazy people, and there are criminals, and there are evil people in the world, although the vast majority of the people are not. We would be very, very concerned if the information that Google managed to search was used to harm someone.

We do work very hard to prohibit credit card numbers, which are relatively easy to get – we try to not send those forward. Similarly with phone numbers. Many of you probably know that phone numbers seem to be routinely printed on the Net for one reason or another. And we've made it easy for you to delete a phone number without us asking you a lot of questions about why, so that your phone number can be protected for obvious reasons. It's harder for things like home addresses and other kind of personal identifiable information, and it would be nice if there were a systematic way to give people that choice.

Remember is that Google is simply an aggregator of information. And the people who publish that information had better have a pretty good reason for publishing it. I'll give you an example of one of the worst sites I can think of is a site of people who list abortion doctors and their home addresses. And they start by saying this is not a site by which you can use to kill them. And the reason is that it's illegal, thank goodness, to encourage murder. So they put the disclaimer in, and then they put the home addresses in. Now come on, guys. This is not a hard question. So again, a little bit of judgment helps a lot here.

Danny:

Is that the kind of site that you would say, we just dislike the whole idea of that site that we will take it out [of the index]? We just won't provide a link to it?

Eric:

It's an example of something that we have debated. The good news about these sites is that they're not on the first page [of search results]. They rank poorly, thank goodness. The smart people are ahead of the crazy people. But the problem we have with that kind of information is it's easy to hide it. It's easy to obliterate it. So even if we were to eliminate one category, the terrible person could do it again [on another site]. So we worry a lot about this because we want Google to be used as a positive force in the world. And we are convinced that the overwhelming value of having all that information available to you whenever you want it and so forth really does justify what we do. And we take that as our mission.

Danny:

You talked about this move to being more transparent that's been happening. And we've seen that – we've seen you out talking, or we see Larry and Sergey out talking, there's been more open Q&As and such like that. But then we also see other things where I think [you're] not being transparent enough sometimes – a huge debate over click fraud. Google must know the rate of click fraud that is escaping [you], or have some real estimates, but we never see numbers. It's fair to say Yahoo! doesn't give me a number either. Yahoo! will say "we're better than industry standards." And I'll say "what are you at?" [And they say] "Well, we're better than industry standards."

Eric:

Just so you know, last week we announced a program where our advertisers can see our estimate of the click fraud rate. So that's an example, and hopefully a good response to your question. To the degree that our systems can detect click fraud, which is a very serious problem, we detect it. We believe we have it under control. We believe that that is getting better. And so we're trying to get advertisers more of a sense of how it's working within the limits of how our systems work.

Danny:

That's a great example, because on the one hand, that is more transparent— but then what you don't tell us is, here's the total amount of money that was spent, here's the total amount of money we issued in refunds for click fraud that got past our system. You do the math, that's your rate. I had this in one of our [online] forums, where somebody who is buying AdWords in a currency other than dollars, is being changed into dollars. And so they're saying well, what's the exchange rate? And Google's response is, "We can't tell you, that's competitive." How competitive can it be? It can only be a small range.

Eric:

I'm sure that we can address that one. We're trying to find ways where we can give information to advertisers that is truthful. One of the problems here is that we have so many advertisers, and have so much going on, we want to make sure you can give people accurate information. And that's why it's been so hard to get this out. I think we will eventually do the majority of what you're suggesting. And the exchange rates are presumably knowable things. You can use a Google search to find [them].

Danny:

Somewhat related: my understanding is that I still can't go to Google's financials and know how much money is going into content ads versus search – and I care about the search. I mean, to me search is a different intention and contextual. And so when people say, "we're going to measure the health of the search market," I want to know how the health of the search market is from a leader like Google. But I've always felt like when those figures are mixed together, it pollutes the data. For all I know, your contextual network is suddenly tanking, a whole bubble is about to burst out there, but search will be healthy. But the whole search industry might go down with it.

Eric:

None of that is going to happen..

Danny:

None of that is happening. And I was going to say, alternatively, everything has been doing great.

Eric:

Since we're on the record, since we're on the record and it's a public company, I want to make sure that what you just said [that the contextual network is "suddenly tanking"] is not true.

Danny:

Right. But that's the opposite to what could be happening. But contextual might be doing wonderfully, and search might be [tanking] …

Eric:

They're both doing well. Again, we have a whole bunch of people who are trying to reverse-engineer the economics of Google. And we have historically not wanted to give out the detailed information that you're describing. These are clickthrough rates, CPCs, RPMs, and so forth. There are a number of reasons [not to split these out]. One of course is competitive. But there's a more fundamental reason, which is that anybody who looks at how Google actually runs the ad network makes simplifying assumptions that are not in fact true. And it's important that we, Google, not give out information that can be misused or is essentially false. So we've chosen, to the frustration of many, to not reveal the underlying economics of the ad box. Partly because it's changing so quickly. And all of the estimates that you see are based on smart people making estimates without our assistance. We think for numerous reasons that's the right decision. It's how we run the business.

Danny:

Do you think that might change as you have more and more ad channels? It was search, then the search and contextual, now we've got print ads, radio … and what about the video ads?

Eric:

You're asking the question the way an analyst asks the question. We are much more focused to the advertiser questions than the analyst questions, because the advertiser wants to know, when I do an ad, what kind of places is it going to go. And we're working hard to get people more ability to target places where their ads go. As we expand, and we were talking earlier about a whole bunch of new deals this week – MTV deal, the MySpace deal, many more coming – these are big expansions of new content, which we think are going to be very successful for our advertisers. But [as an advertiser] you should have the ability to say hey, I don't want to play in that space. And that makes sense to us. We're putting those tools in place, many of them already in place.

Danny:

Let me go to that too, because, again – [Google] started off [with] search, it expanded to contextual, then you started having image ads. Now we're having Google radio ads, soon, buyers will be saying we can put your ads on radios if you want.

Eric:

If you think about radio, here you are, you're in your car, you're listening to the radio. You really are a good advertising customer because you're not going anywhere. So we think actually radio is going to be a very, very big advertising business. And I interrupted you, I'm sorry. I just go too excited.

Danny:

Excitement is good.

Eric:

But I'll answer a different question just to finish that. So we did the radio thing, and now we're also doing two new areas, which we think will be very, very large. One has to do with adding video ads to video content on the Web. Going back to this "don't bet against the Internet," it's obvious that much of the world's video will be repurposed on to the Web, or developed for the Web.

It's also obvious that people who own those copyrights have a proper and legitimate need to monetize them, to make money from them, to build businesses around them, and it's a good thing. So we [just] announced with MTV an ability to take video ads using our targeting engine to target them for video streams and then deliver them to third parties, very similar to the way AdSense works. Similarly, with MySpace, if you look at nearly 100 million users spending hours and hours every day, it's a whole new category of people who we think advertisers will want to reach. To put it another way, it's an under-monetized opportunity. It's an opportunity for you as advertisers – to the degree those in the audience are advertisers, I think many of you are – to reach an audience that you're not reaching today.

Q:

It's good because you've described these new things, and they're exciting things that this audience is suddenly going to have. And I can tell you through talking [to people here], the same audience is going "I don't know what to do. You're throwing these things at me, and I don't see the unification." And one of the things I was explaining [at a panel yesterday], one of the things that [Google VP of Sales] Tim Armstrong had explained when we did this in New York was [that Google is] looking at measurable ads. If you're looking for the underlying unification, it's that whatever [Google does] is going to be measurable. And with something like radio and television, the metrics that [they] have compared to the metrics we have with search, they're laughable.

Eric:

That's true. But remember, these are early days in these new markets. The search advertising market – a tremendous credit to you and to the organization that built this conference – we've been around, all of us in this business, for six or seven years. [But] targeted, measurable radio ads are starting now. Targeted measurable television ads or video ads on the Internet are starting now. We at Google have thought about maybe we could do targeted measurable television ads on real television. So we're thinking about using our advertising system and our targetability for every form of advertising. So you [ask], why would we do that? Because it's a big opportunity to provide greater value to advertisers. It's also a better value to provide value to end users. When you watch television, you see ads that are clearly not targeted for you. When you're driving along in your car, you hear ads that are clearly not targeted to you. It's a waste of your time. So one of the outcomes, if we do this right, and obviously this is a long and big investment on our part, is that you should end up with fewer ads that are more relevant in all contexts. So that the benefit to the advertiser is that they now have real cost per acquisition —CPA as it's known – they can really see that the ad produced a real conversion. And our analysis says that at least in the areas that I have identified [here] so far, we've got a good shot at it.

Danny:

It's been about 10 years since I've been doing this and it's also been that long until we finally had the search engines reunify around any kind of standards. We have the no-follow attribute that was developed to try to help combat link spam; that happened last year. And I was almost amazed – I had kind of given up. And there are talks that are going on to have other standards that are finally coming together on a Web search, on how search engines operate. To me it's encouraging, because it will be so helpful to many of us. But I'm wondering why is this happening now? Did something suddenly change, and everybody [among search engines] said "okay, enough. We know where we're at. We can start working together again."

Eric:

It's always been our goal to do this. Sometimes it takes time. I think, if you think about it, all of the companies in the search space are benefiting from this conversion I was talking about earlier, to this new cloud model where people are living in more and more online. So it's in all of our individual interests to collaborate around things like, as you mentioned earlier, click fraud, all of these issues about link spam. Anything involving the indexing and preparation and aggregation of more information and doing it more routinely with lower load on the sites that we're searching. Standardization of ad formats through the IAB, the Internet Advertising Bureau. We've already agreed as an industry to a set of standardized formats. And there's a lot of work. You mentioned Tim Armstrong earlier, Tim is running this for us to do exactly that. For you all who probably advertise on more than one advertising platform, you should see more standardization, which should make your work more efficient. And I think that's a good thing.

Danny:

AdSense has been this huge boon for publishers – there are really people who just couldn't make a living off of their Amazon affiliate links that have suddenly discovered that they can have real businesses because of AdSense. So it's done a huge, huge force of good. But this also fuels a lot of the junk that's out on the web. And you're having to fight it to keep it out of your organic, free listings – you're having to change it to try to keep it out of your paid listings as well. That kind of starves it of the Google oxygen, but it doesn't go away. And some of the people that seem to be abusive or are generating this stuff, you seem to be happy for them to still get traffic in other ways. Should Google be stomping down even harder on some of those sources?

Eric:

These are the problems of success. AdSense for Content has far exceeded what we thought it could do. And what we're particularly excited about is that a large number of people who did not have an opportunity to make money based on their ideas, their information, their hard work – often [these are] very, very small operations, one person, two people – now at a minimum have enough money to fund their data center work, the computers that they need, their specialized tools, a little bit of an income. In some cases it's even much more successful than that. So I think it's overall great.

[But] as you pointed out, whenever you have one of these big systems, the successes, you can have abusers. You can have people who take advantage of it. And we anticipate that. We automate our search. We automate our advertising targeting. We can test for the kinds of uses that you're describing. Are we doing enough? We're never doing enough, because there's always a new person abusing the system. These are very large systems, many, many, many, many users of these systems. But we work as hard as we can to prevent it. And I think you'll see more and more sophisticated tools to try to take people who are doing ad spam, click spam, link farms, all of the categories of attacks that you've covered in your work, we're getting better and better and better at detecting them. The attacks are more and more subtle, and we're also getting smarter.

Danny:

In some ways search feels like it's marginally boring in that we've got these new…

Eric:

Boring to you.

Q:

Well, not boring. It's never boring to me, but the experience of, I go to this page, there's still a box. I put in the words. I know we've got new verticals that are coming out and the results have gotten much better. But the way I dialogue with a search engine on Google might as well be Open Text or AltaVista in the basic of I'm putting words into a box, I push a button, I get results that come back.

Eric:

People actually like that. They like the simple interface. They tell me this all the time. Do you not?

Danny:

That's my question – I agree with you, people are largely comfortable with it. And when I'm asked about that, all of the analysts and reporters, they'll say to me, "Well, what's the next big thing?" And I'm like, "Well, it's like saying what's the next big thing in television sets." You went from black and white. And then you went to color. You had a long period where we didn't feel like it was working. I could put all sorts of things but then they'd confuse you. And we've made a jump to HD, but a television set largely works as a television set and it feels like that's because people are comfortable with it. So I'm wondering, is search like television sets where we get HD search or are we going to have an "Oh, wow!" moment which just completely changes and we have something different in how we dialogue with them?

Eric:

For many people, many people are very, very happy with the very simple Google search. And we think that's wonderful. There's a set of people who want to personalize it. And so we have a version which is currently known as iGoogle, where you have a login name and a password, and then if you personalize your homepage, you can add what are called gadgets. We started off with a number of gadgets that we wrote. But more importantly we created a developer program and the developer program enables people to build gadgets. The first gadgets were the obvious ones: What's the weather in this town and the stock prices and so forth. [But now] the gadgets are getting much more sophisticated. So my personalized Google homepage, for example, has information sources, newspapers, magazines that I care about. It has pictures; it has video and so forth.

So it's possible to take the personalization engine that Google provides and have a uniquely complicated experience rather than the simple one. So that's the first part of the answer to your question. Many companies who are our partners are taking it a step further. They're taking Google search and they're embedding it deeply in whatever their application is, the most recent example being in MySpace and IGN, which is their other subsidiary in our News Corp deal which we now announced yesterday, where today users in those communities leave those communities to do a Google search. And they're going to integrate Google search in many ways inside of the online community. So we think that you'll have a choice. You'll have the simple Google search bar, you'll a personalized page in Google, and then you'll have these interesting integrated online communities. And I think people will choose the one that they like the best. I don't think there's a single answer.

Danny:

One of the things in terms of changes in the results that has been happening is that Google's claim to fame really was this greater use of links and counting the links and analyzing the links. And in the wake of that, an entire link economy has emerged, to the point that we have things like no-follow. Well, people don't know they can buy and sell, don't do that sort of thing, and Google tries to regulate it. Google tries to go through and say, "We don't like this. We do like that." But is it too late? Should you maybe step back and say, "All right, do what you want with links. We understand there's an economy there and we can't control it nor necessarily influence it."

Eric:

Well, remember, our fundamental goal is to derive the most accurate search results. That is the paramount goal. That's more important than any of the other things that we're talking about. So we always make these decisions based on coming up with that perfect result to that query that the end user gives. And so we would make a decision on any of those technical questions based on [whether] would it improve or make worse our standardized results. So the example of the "follow me" that you were referring to is an obvious example of various forms of abuse. And we have many, many new ideas, most of which are highly proprietary, on how we can use information that's not link-based to improve the quality of the answer. So although we publicly talk at great length about links and so forth, and there are a number of people who've tried to build link farms and defeat it, and we detect those and we try to eliminate them, the truth is that we use link farms as only one of the signals that we use to do ranking, and that the precise formula of ranking is so proprietary only a small number of people actually in the company know it. And I've chosen not to even know it.

Danny:

That makes you safer. You're in this fight with book publishers over the book search service. Not all of them. You've got plenty of book publishers who love it and they're very happily working with you. But some don't. And you've stood your ground that you're doing something that's correct there. The AFP [Agence France Presse] is fighting you on new search, saying that they just disagree with it and you defend your ground there. And yet, this recent AP [Associated Press] deal – while Google has never said, "We did [the deal] because we're trying to solve a legal issue," it feels that way. It feels like it was convenient to get other things with them and sign an agreement.

Or with Google Video, at first you had some broadcasters that didn't seem to be happy with you doing the taping of television programs. And then when it was sort of relaunched and you were working with them directly, that seemed to [gone] went away. So it feels like in some senses Google is having to think more about these publishers and content owners and maybe not hold [to] hard ground. Will that be the way forward? Will you not just sort of launch products and then see what things fall out afterwards? Is there more of a change of if we're going to do something new, let's find out what more of the ecosystem is and see how we can do it with everybody on board?

Eric:

Because of our scale and because of the amounts of money that we have, Google has to be more careful with respect to launching products that may violate other people's notion of their rights. But also, frankly, we find ourselves in litigation and the litigation was expensive, and diverts the management team, etcetera, from our mission. In the cases that you describe, most of the litigation in my judgment was really a business negotiation being done in a courtroom. And I hate to say that, but that is my personal opinion. And in most cases a change in our policy or a financial change would in fact address many of the issues.

Without commenting specifically on AP or AFP or the book publishers, we have to respect the copyright owners' information, and it's okay to disagree on the precise aspects of the law, but no one at Google is suggesting that we are not subject to copyright law. In the United States there is a fairly well-established doctrine of fair use. And depending on which graduate school or legal school the lawyer went to, they disagree on precise details. The ones who went to this law school agree on one thing and the ones that went to this law school went to another. And I've learned that the law is not as crisply defined in this area as you might want. So in our case, we've analyzed this pretty carefully. We believe that the library work we're doing, given that we're not, in fact, reproducing the book but rather simply a snippet and then we have a pointer to the book, is absolutely permitted by fair use. Reasonable people can disagree with that, but that is our view and we spent a lot of time on it. And I don't think we're going to change our tune on that.

Danny:

Sometimes I feel like Google wants to say, like the old Visa card line, "Google, we're everywhere you want to be," in some ways. You're in our email. We search with you. You're on our desktops, ads in print newspapers, you're going to be on our radio, you're giving us free WiFi, you accelerate our web, give us analytic spreadsheets, maps. I joke, but seriously, I want to say, "I surrender. Give me the Google implant and we'll just get it over with …"

Eric:

We're working on that but it doesn't work yet. Would you like to be the test case? [Laughter]

Danny:

Well, there might be some privacy issues, but I'll have to take a look at it. But seriously, the mission is to organize the world of information, make it universally accessible and useful. And that doesn't seem to have any limits. Are there hard limits? I mean it seems to go anywhere. And if you expand, how do you keep people not feeling like they should just fear you, that you're too far into everything?

Eric:

Part of the reason we're trying to be more transparent is because people are concerned that our mission is over-broad. And we try to do what we do very systematically. The company is run essentially by computer scientists, so if we're anything we're systematic. And in terms of the explanation for why all these products are coming out, there is, in fact, a master plan. The master plan is to solve the problem that people have online.

Gmail is an example of a huge success, because it really solves the problem with what to do with all that data online. So it's in your email folder; you can search it. People love it. It's always at capacity, which is wonderful, and that's why we keep spending more money on capital. So the test that we apply is not whether we think the product is great or whether we love it or whether we use it, but does it fundamentally affect in a positive sense people's online experience. We use search as our unifying experience.

If you think about it, all the world's information includes personal information. Personal information is held in online word processing, online spreadsheets, online calendar, online email. If we can unify those in a way that's new and make them shareable, which indeed we announced we're doing; we're delivering the first versions of those products now. You all can test it and see if you agree with it. We think it has a really big impact on end users.

Now, the second question we get is, "Okay, well how are you going to make money on that?" That's not our primary goal. The primary goal is to make the end user experience of living online that much more powerful, that much more useful, that much more fundamental to making their lives better, and to improving the state of the world.

With respect to discussion of fear and so forth, we've talked a lot about the limits to our growth and issues. The most obvious one is that we are literally one search click away from you going to our competitors. So if we do something that makes no sense to you as an end user, we have no way of preventing you from moving your information.

We've also adopted a set of policies which we think are very Internet-friendly where we're not going to keep people's data hostage. We're going to allow them to, if they become dissatisfied with us, to move [their data]. So we're trying to address this question that you ask from the standpoint of how do we solve the end user's problem and keep them happy, give them choices.

Danny:

I'm going to hit you with a lightning round of very quick, easy questions. All I ask is if you're traditional media or if you're a financial analyst, don't put your hand up because you get opportunities to ask those questions. We just kind of want to make it more rank and file. If you are run your small blog or something like that, that's cool. [To Eric] Do you ever say you "Googled" something?

Eric:

If I do, I discover I made a mistake.

Danny:

You get a letter from the trademark lawyers. Don't say that.

Eric:

We have trademark lawyers and they're very smart. I have to listen to them.

Q:

How often do you personally search?

Eric:

I don't know – 50, 100 times a day.

Q:

For all sorts of things or…?

Eric:

When I started with Google, which was about a little more than five years ago – it's a different way of living your life, because if you always have Google around you can always ask questions. One day a politician came in and said in this politician's voice, "Did you know, Eric, that there are more outhouses than TiVos?" And that's an odd question. And by the way, I didn't know. So I did a Google query and the answer was that what he said was true a year earlier, but in fact the growth rate of TiVos – which is, by the way, at the time was 100 percent a year – outpaced the growth rate of outhouses, which was negative 0.1 percent per year. They're declining at some predictable rate.

Now, you fill your brain with this stuff, I guess you get paralyzed and don't get any work done. But I thought it was interesting. And we live in a world where people make all sorts of claims and I just wonder are they true. And although Google is not a perfect truth filter, I personally try to inform my knowledge about the world or business decisions we've made by doing searches. So, for example, when we run the company everybody always has their computers open, and people make all sorts of claims, and we check. We check right there. [Google search results] say "no, that's not true, or here's the link." It's a nice way of living your life, as long as you're connected and as long as you've got a good broadband connection.

Q:

When was the last time you clicked on an ad, and why, on Google?

Eric:

I do it all the time, probably because I want to make sure that everything was working. But I purchase a lot of stuff online. And now with Certified by Google [Google Checkout] (if I can get a plug in for that service), we actually have a system where my credentials, my credit card and address and so forth and so on are stored at Google. You can click on the advertiser and you can immediately purchase the product. So the velocity by which you can purchase a new product is so quick that the moment somebody says buy something, you can buy it.

Q:

And what's your favorite Google product?

Eric:

Well, the one I use the most is Google search, for obvious reasons. Probably the one that is the most interesting right now is Certified by Google [Google Checkout], because it changes the way that you think about advertisers. We were working on this yesterday, and we were thinking now that we've got the ability to get these advertisers to include and the stores included, what can we do with all of that information? Can we search it? Can we index it? Can we make it even more powerful for end users? Larry Page [always] says to remember that the ads have value in and of themselves. And I like that philosophy a lot.

Q:

I'm going to run us five minutes over as we planned, just so we can get a couple questions in there.

Eric:

I see a couple hands up here. One over there, two over there.

Q:

Good morning. Quick question. With Google's desire to be more transparent, do you have any plans or foresee any plans to allow the user to control their click stream data so they can either allow it out, or could actually monetize it for their own benefit?

Eric:

By click stream data can you -- do you mean advertisers or do you mean websites?

Q:

What I mean is that a user of Google, as I go in and search, there's that data that comes back to you as well as the websites that I visit.

Eric:

Oh, I see, as to whether we would see the information?

Q:

Well, you already see the information. The question is, is there going to be an opportunity for me as a user to monetize that by allowing advertisers or other people to get access to it, so I become the controller of my click stream data.

Eric:

That's a clever idea.

Q:

Well, if it's a clever idea I'll give you my business card and you can send me my royalty check. [Laughter]

Eric:

It's a clever idea. We're already doing some of that because in our personalized search, if you're logged in, we actually remember where you went and what you clicked. So we do, in fact, have ways of aggregating information. I guess it hadn't occurred to me that that in and of itself, the end user could then decide what to do with it beyond that.

Q:

Good morning. With regards to Google's ongoing trek to seek out and find the perfect result, as you mentioned a little earlier, algorithmic search is currently not very good at determining user intent. Is Google doing anything other than Co-Op to address this?

Eric:

User intent is very hard. And I assume by user intent you mean more of an AI component of what did they really mean.

Q:

Exactly.

Eric:

Today we understand that if we give people a sub-menu, a choice, and Google Co-Op is an example of that, where they can see a sub-taxonomy against their queries. Like in a health query, it says "did you mean medicine, did you mean recovery, did you mean drugs" – whatever, [searching] against this illness that you might have or you might be researching. It works in areas of information that are well defined, that have a well-defined taxonomy. We've not yet figured out a way to do that generically, to have our algorithms and our AI technology invent it and discover it. And I think that remains an important research problem.

Your question is probably at the root of where does search, algorithmic search go. Eventually, the computer wants to know what you really meant. And even with the history of the user, which of course we have, and they can give it to us with the click stream data, and then these sub-taxonomies, we still don't have enough understanding of what they meant, especially since language is ambiguous.

Q:

What is your current thought on eBay's response to Google Checkout and their lack of integration?

Eric:

They would have to comment on what they're doing there. eBay has turned out to be a big partner of ours. They're a big advertiser. We do a lot of stuff with them. And I suspect that we send a lot of traffic to them; they send a lot of traffic to us. It's their decision as to how they want to integrate Checkout. We've also told them that we would like to integrate PayPal as we can because we want to give people in [Google Checkout] a set of payment choices.

Q:

Hi, how are you? Let's see, I actually wanted to ask about the trust that users place in the Google search results. When a user does a search, they oftentimes come back and say, "Well, the first result, that's the most trustworthy one, the most truthful one. I put my faith in it." And maybe this is a problem of less-savvy users more than more savvy users. Do you think that Google will ever try to do something to say, "Hey, users, pay attention – you're not always going to be getting something highly truthful. You're not always going to be getting something that you should believe in." Is there some type of disclaimer that you guys would ever think about? I know that it doesn't exist right now, but…

Eric:

That's sort of like "objects may be closer than they appear in the mirror." I'm not sure that there's a lot of confusion on this. I think most of the people that I talk with understand that Google is not a truth machine and does not represent it to be so. We do the best we can. The evidence that we have and the testing and analysis is that we do the best job of any of the search engines. And we invest a tremendous amount of time and effort to have the most accurate, most comprehensive, largest, most sophisticated, most international index with the quickest result. That's what we do. That is our fundamental mission. Everything else that I've been talking about is secondary to that fundamental mission.

As we get better we might have the problem you're describing. People might say, "This thing is so good, it's a close to a truth machine as you can." But then I'm sure someone will come up with stuff, probably invent some way to make sure that untruthful things get there. So I don't think in our lifetimes we'll ever get to a perfect answer. But we'll get closer and closer to the model that you're describing.

Danny:

I think we've got time for one more.

Q:

You talked a lot about transparency, and you guys have talked about kind of being good, doing no evil. On the AdSense side, there are two things we hear from the publisher community. One is kind of they want to know what kind of revenue share they're really getting, instead of a black box. And also we hear a lot of feedback about your requirements and your terms of service about requiring contextual exclusivity at a page level. And both those things seem to be in conflict with some of the big themes that you've talked about vis-à-vis transparency and I'm just wondering if you could talk to that.

Eric:

Again, the general theme about transparency is to talk about what we're doing. The specific product details we've decided to keep largely proprietary, partly because of competitive issues and partly because they continue to change. And the particular things that you're asking, both the rev share and the need for terms of service, are as related to landing page quality as they are to the things that people are asking about. We're trying to find ways to give that information but without giving up the information that we don't want our competitors to see.

Danny:

Eric, I forget to ask you if you had a hands-up question for the audience. I'm going to do one for you. How many of you search with Google? And how many are Google AdWords advertisers? [Show of hands]

Eric:

Wow. Well, thank you very much.

Danny:

And how many are AdSense publishers currently?

Eric:

Even better. Wonderful. Well, thank you very, very much for coming. Thank you for being partners with Google. I really appreciate your partnership, your business and your ideas. And we'll continue to evolve this amazing, amazing phenomena.

Danny:

Thank you very much for being here today.

Eric:

Thank you.

end of session