Web 2.0 - A Conversation with Eric Schmidt

by Daniel H. Steinberg

Google CEO Eric Schmidt joins Web 2.0 Summit 2006 program chair John Battelle for a conversation. They talk about Google's recent purchase of YouTube and the importance of video as a format. Schmidt explains why Google stood up to the request to provide end-user data. Batelle asks if Google is building an online suite to take on Microsoft Office?

You can download the audio as an mp3 or download the video as an mp4. Here are all of the Web 2.0 Summit Podcasts released so far. You can also subscribe to the audio podcast or to the video podcast.

Intel Software Network Intel Software Partner Program

This episode is sponsored by the Intel Software Network.

This transcript is created by Casting Words

ANNCR: This week "The Web 2.0 Summit Podcast" presents a conversation with Eric Schmidt, Google's CEO and Chairman of the Executive Committee. Here is Web 2.0 Summit Program Chair John Battelle in conversation with Eric Schmidt.
John Battelle: Eric, thank you for coming.
Eric Schmidt: Well thank you, and congratulations to you and Tim for quite a show.
John: Yeah, well it's...
Eric: We're glad to be part of it.
John: It's good to get it rolling you know, get started here. So let me start with just a very short, simple question. Why'd you buy YouTube?

[audience laughter]
Eric: Cause we like them.
John: Well wasn't Google Video doing well enough?
Eric: Actually Google Video was doing extraordinarily well. When we tracked it's growth, there was a breakout in the last six months that I don't fully understand of user generated video content.
John: Was it the link you put on
Eric: No, I think it was actually people using camcorders and things like that.
John: I've got to give you a little...
Eric: I know. And we studied this for awhile. We went over and visited the YouTube folks, and their business was growing even faster. So something happened this year where all of a sudden video became a fundamental data type on the Internet. And we want to be part of that, and that's fundamentally why we bought it.
John: So YouTube kind of feeds the index, which is you know the core, this data base that Tim mentioned. But initially when the announcement came out you said, "Well we're going to keep it you know, we're going to keep it a separate company, we're going to..." Do you think that that's always going to be the case, or?
Eric: At Google it's hard to know what will happen next year, as opposed to infinity. The most likely scenario is that we're going to keep YouTube business separate property. The businesses have been integrated. And the reason is they're very much focused on the community and social networking aspects of video, and they have some special and unique ideas about how to address that. And it's clear that there's a viral component to use a set of overused sound bytes in one sentence. But that the underlying draw of being able to see what people are doing, have computers suggest adjacent or related kinds of videos, is a whole new paradigm, and one that's important for end users.
John: Now after the purchase had came out and, I will allow you to address the accuracy of these claims, that a very large sum of money in the acquisition was set aside to in essence by peace with the media companies. How's that going?
Eric: Well, the former is not true. The latter we have talked to all of the, we've visited with as many as the media companies as we can, because we have to respect copyright. And in many cases the explosion of video gives us an opportunity if we can construct the right partnerships, of course is the goal, to actually get the correctly licensed content into the right places. And it could ultimately result in a much larger set of viewers for that content, viewers, listeners, and so forth. YouTube had been on this path, and we have been on this path, and I think together we'll move even more quickly. So we've certainly talked to everybody. It's very complicated because many of the media companies have very sophisticated rights management systems, and very complicated rights structures, and they have to go through something called the "Rights Clearing" process to do that, then they have to determine the right monetization, and they have other customers who's interests they have to respect as well.
John: But you're pretty good at that monetization thing so are you helping them out?
Eric: Well we're working on models of how you would monetize Copyrighted Content. As part of our strategy to provide an advertising model for things other than Text Adds, we have a pretty good idea of how to monetize other kinds of traffic. And part of the secret at Google of course is that we did that really. really well, really early. So for example when we did the Mice based deal, which is why the disgust, that was fundamentally because we had some breakthroughs in our ability to monetize social networking traffic, some new algorithms, some new inventions.
John: Which leads me to another question. Google has what some would claim to be an "unfair advantage" in that you have a critical mass and a critical amount of data that you can stare at long enough and hard enough, and do enough sort of clever things with. Where you can pretty much price a deal better than anybody else - MySpace being an example, YouTube being an example. You can say, "Well we know how to monetize that better than anyone else." Are you worried about, and I've heard this from some folks that I've talked to...
Eric: John I'm confused, you're saying our advertising product is better.
John: That's right.
Eric: Good.
John: That's just unfair.

[audience laughter]
Eric: No seriously, I mean we built a very, very good Targeting Engine, and a lot of our business success has come from that.
John: Right, and so there are been many examples of companies that had better products that have been rapidly adopted, and then it become sort of the standard, and then sort of something sets in. The community starts to worry that that company has to much power, people start to plot against the company, people start to try to figure out ways to compete around to that company, and so on. Do you worry about that?
Eric: Sure, absolutely, and what we do is we run the Company around in users. And so the way we think about that question is - as long as we're respecting the rights of end users, who we care a lot about, as long as we make sure we don't do something which is against their interests, then we should be fine. If you look at the historic large company behavior, they ultimately do things which protect their own business, practice, monopoly, what have you, against the choice of the end user. So the more that we can for example - let people move their data around, never trap the data of an end user, let them move if they don't like us, the better we think we'll be both genuinely treat our end users, and be perceived as well.
John: I wanted to ask you about that data, this is a big topic, I think you know Tim said, "Data is the Intel inside", you've got a lot of Intel. Your approach towards users has been, "We will protect your privacy." But what you said is actually interesting to me in that if I, can I get all of my Search History and export it to another, to Yahoo? Can I...
Eric: We would like to be able to do that. We are working to make sure that as long as it's yours, so it has to be appropriately authenticated, you have to have a Log in Name and a Password. We want to be able to, we want to give you essentially the equivalent of member portability. Because we think that that is ultimately, if you will, the "pressure valve" on bad business practices on our part, and on others as well. As long as people feel that if they don't like what we're doing they can just switch, then that keeps us honest and keeps everybody else honest as well.
John: Let me give you a scenario, and then we're going to move into another thing about Web 2.0 where we had a conversation back stage and I want to talk about that. Let me set up the scenario - Google has done two things in the last year and a half or so that I think were, I guess you could say, "Almost honorable." Number one - you stood up to the Department of Justice, and made a case out of their subpoena of 35 or so companies. They asked for "your entire search index", and they asked for as I recall, "a week's worth of queries", and they wanted to go fishing with that. And you said, "No." And they then...

[audience applauds]
Eric: So, but I mean you're describing it in this very nice way...
John: No no no no no, I got another angle on this.
Eric: Let me just say that if you think of it from an end user, why would an end user want that done to them, unless they're at all near some crime.
John: Right.
Eric: It's a complete violation of people's privacy. And a Federal Judge, thank goodness for Federal Judges, saw the wisdom of our argument and ordered a Restriction of that. It was a very good outcome.
John: Now had you not done that, it would be entirely possible that no one knew about this, because actually most of the other Companies involved didn't make a fuss. And so Journalist's like myself couldn't actually, there was no Court Case that we could go look at and find out what was going on. And so that in my point of view, a great thing, before that when you went public. So basically you on Principles stood up, and obviously there's good reasons to do it as well, because you don't want your Index in the hands of anybody, it's your Index, but you in Principle stood up to the Department of Justice, to the U.S. Government. When you went public on Principle you stood up, not to Wall Street but to a lot of the Practices of Wall Street. The way that you approached the IPO, you stood up to a very powerful force in the way Wall Street likes to get things done, and you did it differently. So here's my scenario - Under the Patriot Act, under the principles of which you are not allowed to disclose that you are being asked for information, you are asked for something that you believe is morally wrong. Would you stand up to it?
Eric: We will do whatever it takes to follow both the Law and the Principles of the Constitutional Government of the United States.
John: But what if it's against the Law?
Eric: Well then we'll have to talk to a Judge. Ultimately "we" and all of us in this room are subject to the Law of the United States, as we sit here. And we have no question as to whether we're subject to that. We don't think we're citizens of some other Planet. So the fact of the matter is, we are subject to that. But one of the great things about America, on a day in which hopefully everybody had a chance to vote or voted by absentee ballot, is we have a System of Law, we have a System of Review. And we won't win everything, but we're going to try very hard. And we will absolutely follow the Order of a Federal Judge.
John: I very much hope that we will have a case that breaks this out, because I think...
Eric: Hopefully it won't be us.

John: Well as I...
Eric: The question you're asking is, "There's clearly going to be somebody who's going to test the limits of the Patriot Act."
John: Right.
Eric: Patriot One and Patriot Two, and it's an important constitutional issue.
John: Yeah.
Eric: It's not clear to me whether it will be Internet or libraries, or something. But everybody in this room understands that the law, that there's always this tension between Freedom of Speech, terrorist rights, secret trials, that kind of stuff.
John: Right.
Eric: One of the great things about America is this debate occurs in the open, and we as a society as a whole can come to a rational outcome. And if we don't like it, we can replace the people who passed those laws.
John: Keep that in mind if - the next election. So just, there's like a 150 bloggers out here so it's the beauty of this conversation is, it's going to be well reported. Now...
Eric: There was a blogger conference where no one looked up because everyone was typing. It's very strange.
John: Yeah, it sounded like it was raining...

Eric: There's a lot of noise in the room and they're not talking.
John: Yeah, yeah. By the way when we first spoke in 2002, in early 2002 when I was working on the book, I said, "What do you think is going to change the next couple years?" And you said, "I'm going to hire a lot of lawyers."
Eric: Well we've done that, I'm proud to say.

Eric: And some of the best in the world.
John: Yes, some of the best in the world.
Eric: We're very proud of our Lawyers, we have a lot of them.
John: So let me switch gears here and ask you, back stage you said, "You had read all Tim's papers on Web 2.0, and you took issue with one thing, which is that you didn't think the definition was broad enough." Can you speak about it?
Eric: You know the most interesting thing that's happening now is that as the "Internet Phenomena" occurs, and all of us are part of that, it's beginning to effect Industries that really don't know what to do about it. And it's worth reminding yourself, that it's a mistake to bet against the Internet. Don't bet against the Internet. As the Internet has rolled into Industries, it has really upset things. And it's upset things especially with Businesses or Structures that withhold information, that prevent communication. The Internet is fundamentally about sharing, it's fundamental about branding globally, it's fundamentally about global businesses, and so forth. There are many examples of this. The most interesting one, and the one that has recently I think emerged as the "Platform Theme" that you all were talking about at this conference. And the easiest way to express it is, do you think you'll get better software out of a programmer if they're building software that's packaged software that you can buy off the shelf or online, verses do you think you'll get better software off of a running data center that's run 24 hours a day? And all of computer science says the better answer is the latter. And the reason is - that it has to work. It has fewer variations, fewer bugs, it's more robust, more reliable. They don't break cause if they break, a million people give them feedback at that very moment, and all of a sudden off they go. So one of the transitions that I think people have not focused on in this Architectural Transition, which is generally around technologies called, "LAMP" and "REST." LAMP stands for Linux, Apache, My Sequel, PHP, Pearl, Python. REST is a particular data architecture around APIs. Is that, that transition enables this area, this sort of era of huge massive servers, which Google and many other Companies are now building. And these massive server farms are fundamentally going to be more reliable than the things they replace. And that shift, which is a very user centric shift, means that users can get back to whatever they were doing, rather than debugging their software. It's a big change. It also means that those users can use other kinds of devices, mobile phones, new kinds of inventions. It's a real change in the paradigm. So if you take the sum of the changes in the economics of the Internet, the fact that everyone is now living online in these social networks. And this emergent new platform, you can see an industry which is larger probably in scope than the size of the PC industry 15 years ago. And certainly has as many players, if not more. And many of them are going to be winners. That's why it's so exciting.
John: Now this insight which you've actually been, you've been on about this since Sun really.
Eric: Well actually this is the same speech that we've given for almost 20 years, architecturally.
John: Right.
Eric: The difference is Moore's Law. Anybody here use something called the Network Computer announced in 1996 by Oracle and a number of people including myself representing Sun? I'm sure you didn't because it didn't really work.

Eric: But 10 years later, which is by the way a factor of 100 under Moore's Law, all of a sudden the computer's faster, there's a lot more DRAM, the data networks are faster, you have the development of 2.5G and 3G Networks, etc., etc. Finally now the architecture works. It's fundamentally better to keep your money in a bank, than to keep it in your pocket. It's fundamentally better to keep software applications, data architectures, all the data stuff that you're talking about in your conference. It's fundamentally better to let a 24-hour data center handle that, then you can use it as you want. This is a very large architectural shift, and one we're just at the beginning of in terms of its impact on how we'll consume applications and information.
John: So when we look at a major areas, you know development at Google, you know one of them is the Office Suite. Last year Sergei was a bit more coy about Google's desire to go into this area, but I don't think that you can be coy this time, it's not allowed.
Eric: It's interesting that you use the word Office Suite, because that's not what we talk about, we don't call it an office suit.
John: Well, formerly known as...
Eric: But we never made that claim. This is an important distinction. The argument goes something like this. For many people it would be just as easy to have the computers that are in the cloud, if you will, and Google's data center is one of those data centers, store the information that you use everyday. We can search it, we can do all sorts of interesting things with it automatically, and you don't even need to think about it. Furthermore, if you have that model, and if you put that information out in the cloud it should enable very rapid sharing, it should make it very easy for people to share information.

So we embarked on a strategy more than a year ago to build applications that were search centric and highly sharable, Gmail being an example of that. But if you look at what is now docs dot, what is essentially the docs suite, spreadsheet, calendar, writely, and there's other aspects of that coming that have not been announced yet. The sum of that is a different way of managing information.

We don't position it as an office suite, we position it as something that you would use everyday, casually as part of your normal life. With your family you'd have a shared calendar. You'd integrate all the information, you'd have people jointly editing documents that sit there. We're not arguing that it's an office replacement. We're arguing that it's a different way of managing information. And it should be judged on that basis.
John: Which ultimately replaces Office?
Eric: I don't think so, I don't actually think so.
John: You don't think so?
Eric: No and here's why...
John: Well we'll ask Ray Ozzie tomorrow.
Eric: Well, I should not speak for them, but I will tell you the focus that we have is not the focus that we have.
John: Can you define that a little more?
Eric: Our focus is on casual sharing, casual communication, casual interaction.
John: So it's not necessarily enterprise, is what you're saying, small, medium business?
Eric: Oh, not particularly, except insofar as they need these simple solutions. It has one benefit, [laughing], which is in that it is free.
John: The benefit that it's free, that's a hard benefit to beat.
Eric: You could pay people to use your product.

John: Well I've seen your cash flow, and your quarterly statements, you could probably do that if you really wanted to...
Eric: It's not, it's... No, but let me give you a simpler example, it makes sense, right, that mobile phones should eventually be free, because, and you should have free telephony and all that, because it should be possible to monetize that. In other words, it should be possible that if somebody has one of these digital devices there should be enough economics in that transaction that it literally makes sense to pay people to use your mobile phone.
John: Nicholas Zenzman from eBay definitely agrees with that sentiment. Let me move, now that we just talked about monetization, and then we are going to get to questions. And there are microphones set up in the aisles there.

About sales, when we, again, going back to 2002, I asked if you were a media company, and you said 'No, no, no, we're a technology company driven by media, media revenue'. And know you, I would say, you're a technology company driven by a really lot of media revenue, and so much so that I think it's hard to argue that at least a piece of Google isn't very media driven, and therefore very sales driven. And sales driven companies that have been very large and successful in a market often become a little slow.

You look at IBM, and IBM's sales organization.
Eric: We'll work harder.
John: Look at Microsoft, and Microsoft's sales organization. Both extraordinarily powerful sales organizations, built up over a decade. You have hired thousands of sales people in the last 18 months...
Eric: And thousands of engineers.
John: And thousands of engineers. But at the end of the day in an argument between an ad sales guy and an engineer, who wins?
Eric: The engineer.
John: The engineer. I just had to ask that question for the record.
Eric: Let me answer the question that you didn't ask. We run the company around solving needs of the users. We primarily run a company around our 70-20-10 model, which I think everybody here has heard about. You have 70 percent of your core business, 20 percent on adjacent business, 10 percent on other new ideas. Virtually, all of the really interesting stuff has come out of the 10 percent. You know, crazy ideas. We encourage our engineers to spend 20 percent of their time on new projects, new areas. That model has scaled extremely well and we don't see any current limits to its growth. I always worry that there's some architecture, you know I'm a scientist right, you always worry if there's some asymptotic growth limit from just command and control and scale and so forth. We have not seen that. Doesn't mean there isn't one, but we have not seen it yet.
John: Do you worry about going in so many directions at once?
Eric: Well the way we handle that internally, aside from working even harder, is we have a way of classifying innovation. So the company may appear to be chaotic, but in fact the strategy is very, very coherent. And the strategy is around innovation, around these markets to solve these problems. The chaos that occurs internally, is the creative process, which we think works well for us.

The sales comment may be as much a comment about the importance of partnerships, and one change in the last year has been our recognition that we must, that we were not particularly good at working with partners.

That we needed to work very hard to work with them, to listen to them, to learn, frankly, to learn how to make money together. So we've been announcing partnership after partnership in different areas to essentially extend our model, to extend our reach. There are partners who have access, who have technology, they have media, and so forth. This partnership message is fundamental to I think, the transformation of the company. From sort of a technology upstart, in the way that you were describing us before, to a more mature company that understands that it has to work with everybody else in the industry.
John: Great. Well, we have mics setup. Anyone wish to ask Eric a question, please come up and do it now. You can't possibly not have any questions for Eric. Over here.

Mark Mehany: Hi, Mark Mehany at Citi Group. Today, maybe at best 5% of all advertising worldwide is on the Internet. Even 20 or 30 years down the road the vast majority of advertising revenue worldwide will be in formats off the Internet. Is it helpful to think about Google, 20 or 30 years down the road, having the majority, or a much larger percentage of its revenue in formats that are off the internet?
Eric: 20 to 30 years is a long time, even for anybody, right? The general observation about the advertising industry is that it has not had a lot of technology innovation, such as new targeted models for a pretty long time. A set of companies represented in this conference have been working on targeted advertising, and essentially better economics for the advertisers. It's worth noting that the majority of advertisers actually advertise for a reason, they want to sell something. So if you can show that advertiser a better return on invested capital, literally a better spend model, they will move their dollars or they will shift dollars from other categories, that is not advertising, into advertising. So we may very well be in a Bull market for advertising, in particular, for targeted advertising. And there's no particular reason to believe that that model is limited to one technology space.

So we, for example, as has been widely reported, are investing pretty heavily in radio, which we think is going to take off. It was announced this week that we have a partnership, which includes by the way the New York Times, where we are essentially aggregating advertisements and putting them into essentially an auction mix, that then goes, and that they essentially accept bids, and place it in the newspapers. We believe that the generalization of these markets makes them more efficient, more accurate, and so forth. There's no particular reason in our view to limit that. A typical example would be that we might be able to go to an advertiser and say, "Here's a set of tools." A more likely scenario is - we go to an advertising agency and we say, "Here's a set of tools that allow you to do essentially cross media planning." And then the advertising agency using the judgement and experience that they have from doing this for so long can use those tools to more accurately plan, plan revenue and expenses. I don't know the limit of that model but the Internet advertising business will ultimately be a very large component of overall advertising if you define it the way I described it.
Evan Pagen: Eric, Evan Pagen - Hot Topic Media. Question, you guys are famous for having attracted, and kept some of the smartest minds in business, period. What have you learned about attracting, developing, and keeping those folks that could help others?
Eric: The simplest answer is that people don't work for money, they work for impact. And if you can figure out a culture where, one - you listen to the people, and two - they have an impact, you have a winner. Google is run in an unusual way, it's very consensus driven, we don't have single decision makers, it's hard to deal with, and our partners have learned this, and we're working as hard as we can to address that. But ultimately the fact that everything is done in groups, people talk to each other, and so forth, and so on, has produced more robust decisions in greater volume. Another way of saying that is, "that you tend to get better answers if you ask a question as a Manager, rather than telling people the answer." So the way we do our strategy for example is - we produce 29 questions for next year, which are reasonably obvious I think to people in this audience, and we sent teams off to try and answer them. And we'll see what they come back with. Much better than telling them the strategy, it promotes volume.
Evan: Can you give us a couple of those questions?
Eric: How do we, there are lots of examples - What are the limits of technology in search? What are the limits of the advertising model that I was just asked about? How extensive can the mobile market move? How quickly can mobile be extended? What do we do with the platform layer with respect to cost structure, capital expansion? There's issues of running out of electric power, there's issues of, literally, literally if you take this model to the extreme, and "we think big, " you run out of things like "literally" space for the computers, and things like that. How do you develop a standardized way of set of Interfaces that can be used to build applications that are web centric in the "old term"? What does that look like, what does a developers program look like? Those are the kinds of questions.
Evan: Interesting.
John: Okay, we have time for two more questions.

Questioner 1: Do you think it's possible to build another YouTube playing by the Rules of Copyright, in the current climate copyright, is it possible for a large Media Company to scale like a YouTube has done?
John: Interesting question.
Eric: I think there's a premise in your question that YouTube somehow didn't play by that rules, and I'm not sure that I agree with the premise. What appears to happen in these network markets is they follow a power law distribution, and they quickly get a 1st and a 2nd. This is not a new fact, it's been true in network markets for a hundred years. So what appears to happen is that somebody has a breakthrough, cause remember YouTube had many competitors, and today I think has many competitors, but somebody has a better idea, and they drive it. So I think the answer is "Yes". But the idea is driven around a complicated set of user tradeoffs, it fundamentally produce a better user experience. So I think that the true answer is - that all of these sites and all of these sort of breakouts that we're talking about here at the conference can emerge with a set of insights about how people will use information, form social networks, cross link to each other, and globalize their content and their ideas.

Questioner 1: What about for the large Media Companies where to compete...
Eric: No they certainly can, and I also think small Companies can do it.
John: Last question here?

Questioner 2: Yeah. So you're obviously investing a lot of money into data centers along with Microsoft and Amazon, and you're basically saying that, "Web Applications are much more reliable in a data center." But isn't there a problem right now of kind of activity being ambiguous you know? For instance if you're on a plane and you need those documents, what are you going to do, and what do you guys plan to do to sort of fix that?
Eric: This is a very good example of the limit of the current architecture. If you think about the architecture of 20 years ago, it was a disc centric architecture. As long as you carried your disc around you were fine. It's possible that the new architecture, the network will always be presumed to be around, and the disc is optional. And there are computer scientists working on that, so we're in a transition period. And by the way for better or worse, people are on airplanes are going to have broadband networks, and people are going to be making VOIP phone calls and drive you crazy. So this is going to happen. So at least in the western world it looks like ubiquit is broadband, which we've talked about for many, many years, he's really arriving. And for developing countries there's a lot of reasons to believe that the 3G wireless networks, as more competition emerges and so forth will provide that. The specific question which has to do with online access, there are people in the Ajax model working for technology that will enable essentially a cashed proxy, where you cash the information on to your local client, you can disconnect, it performs, and then when you reconnect it, it sinks. It's a reasonably obvious technical solution. And it's one that we should expect will be built out of an industry that has this much investment in it.
John: There are Companies doing that now as I recall.
Eric: Yeah, in fact there's a number doing it.
John: Last question, just cause I had to ask. You run the company as a triumvirate. Who wins in an argument when you argue with the founders?
Eric: They always win.

Eric: And that way they'll give you a different answer.

John: I can't even solve for that. So can you give us one example of the time when you won but it was a mistake?
Eric: Oh, I've in the triumvirate, I'm the one with experience, who's late. They're the ones with the inexperience, who are early. Virtually all of the interactions that we've had where I've had a concern and they've been right, the compromise has ended up being in the middle. So left to my own devices I'd be right but late, and left to their own devices they'd be early and right, but too early. And that's I think what makes it work so very well.
John: Well Eric thank you very much for joining us.
Eric: Thank you very much.

John: Great, thank you.

ANNCR: A conversation with Eric Schmidt at the Web 2.0 Summit 2006.

Daniel H. Steinberg is the editor for the new series of Mac Developer titles for the Pragmatic Programmers. He writes feature articles for Apple's ADC web site and is a regular contributor to Mac Devcenter. He has presented at Apple's Worldwide Developer Conference, MacWorld, MacHack and other Mac developer conferences.