Fred Vogelstein: When you and I talked a month ago, you mentioned that you thought about the Dell model a lot when you run Google. Can you explain in more detail what you meant by that?
Eric Schmidt: Well, there's a couple of positive analogies for Google, I think. One is Intel, one is Sony, and one is Dell. And I think each of them have strengths. Intel is easy. When they came over to visit with us, Larry and Sergey and I looked around and said we want to have our executives be as precisely knowledgeable as these people. In every case they were very precise with respect to facts; they were on top of things. If you asked them a question they'd give you a crisp answer - noneof this wandering around that goes on in the industry. It's characteristic of the Intel style, so that was very impressive. I'm not talking about the other aspects of the style - the paranoia and all. I'm just talking about the fact that they were very, very, very knowledgeable. And in order to do that, you have to have a high IQ, and you have to have a discipline around knowledge.
Sony is the best example of what Google is trying to emulate in innovation.
That seems counterintuitive given the kind of problems Sony is going through right now.
But if you look over a 20-year period - and a lot of it had to do with (Akio) Morita, who was the CEO - they built a culture on innovation and people -- and I remember this in studying it. Maybe they're having trouble now, but if you integrate Sony over the long time, they're the ones who always bring out the slightly different product that has a different solution that's interesting. You can debate whether memory disks, memory cards make sense, right, but they came out, they saw the memory stick; they solveda specific problem. And they built this odd little book reader thing which didn't work, and then they had this odd little new format around recorders which didn't work, and, of course, they have the Walkman and the watch.
You mean they're not afraid to fail?
Thank you, that's the point. Sony says it's perfectly fine for product A to fail because we're all going to work together for the next 50 years, so we’ll just try something new, just keep going.
Wouldn’t you be more interested in the Intel innovation cycle than the Sony innovation cycle, because a big part of Intel's success has been the speed of its innovation -- a new chip every 18 months to two years.
Well, I'm not sure I completely agree with that, because that's largely determined, not by product innovation but rather process innovation. Intel is particularly good at implementing Moore's Law at scale with fabs, and they really are amazing with capital use, deployment, planning and so forth. And much of their success in the last ten years has been basically because of the people who (former Chairman and CEO) Andy (Grove) put in place actually knew how to spend 2 or $3 billion prospectively on a fab line.
Google is different?
Google does not have that (manufacturing process) aspect to its business.
The third management example is Dell. If you think about the history of the PC industry, the PC industry has essentially been nothing but acquisitions by one company or another. Dell is the outlier. Dell built its own culture. They automated themselves to be the most efficient manufacturer. They have a relatively simple business in the sense that it's a high volume manufacturer without a direct sales force. You had to call them or send them a fax. From the standpoint of the consumer Dell was a mail order company.So they started off with this very simple model and numerous, numerous people have suggested that they buy this company or that company, that they broaden their mission or so forth. But in every case they stuck to their core differentiator, and Michael understood that this was a market that would have a cost efficiency benefit of scale. People use the word "commoditized," and that's not the right word. He understood that it was a scale business and he was going to be the low-cost, scale provider.
He understood it was a manufacturing business, not a technology business, right?
Disagree. Technology is used at Dell to be the low cost provider.
What I mean is he understood that it was a manufacturing business and not a technology business compared to the way the rest of the industry thought about it. The rest of the industry talked about how cutting edge their PCs were.
That's right. So then when we judge historically, Michael's view was more correct. But the real lesson to learn about Dell had to do with the fact that the principles by which the company is run have been constant for 20 years.
It's basically this incredible focus around cost, incredible focus of scale, incredible focus on automation, right? And that transcended all the different distribution things. And they were willing to trade off innovation. They didn't do a RISC processor. It made no sense.
So the analogies go like this: The Intel example is the kind of executive we want. The Sony example is the kind of innovation we want, which is the allowing failure. The Dell model is the homegrown model; they built the company within its own culture and it has scaled. They did not build Dell by a series of acquisitions. Google is very much a not-invented-here, build-it-ourselves culture. We wouldn't want to go bring in some large other group with a different culture and then screw that up.
There is a science to managing high tech businesses, and it needs to be respected. One of them is that in technology businesses, leadership is temporary. It's constantly recycling. So the asset has limited lifetime. So as a management discipline you have to be always paranoid in the Andy Grove sense, waiting for the next thing that will come kill you, which means you have to be open to listening and talking. So that's sort of rule one.
Rule two is that it's very, very easy to lose control of the details in these businesses.
You mean you start reading your press clips?
Yeah, you read your own press, you listen to your own speeches and so forth.
And so it’s the analytical nature of Google, and the fact that we're a Web-based company and that we have instantaneous data that helps us be disciplined about management. We know the status or our projects on an instantaneous basis; we know our revenue; we know our clicks. We don't show this to anyone, but we can see it. We have internal dashboards, which I see every day. So you can just keep your eye out on it. As you see a problem, you raise an alarm. It's a discipline.
How do you get these companies to actually sit down with you and have this dialogue?
You just call them up. Paul (Otellini, Intel’s CEO) is a long time friend. This is before he joined the (Google) board.
In Dell's case, I talked to Michael at a conference. He said I'd like to sell you some servers, and I said, well, we have some interesting technology, why don't you come by for the tour. When I joined Google, I sensed that Google was pretty interesting so I decided to sort of show it off. So as I encountered all my various friends, I would invite them over for the tour, and I would take them to show them the company – this is when it was in one building.
So, we'd walk around, they'd see the traffic chart, which no longer exists in its form because we're a public company. If they were a really big cheese, they'd sign our chart. It was fun.
I think to some degree one of the strengths of the high tech industry is that people are actually willing to tell you things. When I went to Novell, I didn't know how to be a CEO, so I went in and I called all sorts of CEOs I knew. I called in a favor. I wanted to come by and listen to them tell me what it's like to be a CEO.
When you got here in 2000 it was a relatively new company run by people (Larry Page and Sergey Brin) who were at that point probably 26 or 28 years old. How did you convince them that this (inviting other executives to Google, soliciting their management advice, and installing a more systematic approach to running the company) was a good idea?
I don't know. When I came here, I came because I liked Larry and Sergey. It was an interesting little company. I had no idea it would be successful. It was not broken. But we needed leadership.
For example, we had an accounting system which was an Intuit based system designed for five users, and they were using it for 20 people. It was too slow to use, so I suggested that they implement an Oracle system. It was a huge crisis. We ended up spending $100,000 for this. Larry and Sergey nearly had a cow over it (because they thought it was so expensive). A hundred thousand dollars is the cheapest Oracle system ever implemented in history I think.
How did you convince them that you needed to do this?
Well, it was actually very interesting. Larry and Sergey suggested that we should build our own, because most of the existing accounting systems weren't any good. And I said, "I'm sure that's true, but you'll never get it audited," And I thought that was a pretty clever argument. The auditors would never pass financials (generated out of software) that we built ourselves. And Larry and Sergey today will complain about the Oracle system, but they'll also say "We had to get one that was auditable."
I've always said that if the company were founded today on an empty lot, we would build the buildings brick by brick. We can't imagine someone else building our buildings, we'd have to build it ourselves. This is a build-it-yourself culture. The good news is there's no free land, and so we have to rent the buildings, rather than build them. But the culture is around building things. In that sense, by the way, it's similar to some of the companies (Intel, Dell, Sony) that I mentioned earlier.
You talked about how Google was different than Novell. How is Google different than Sun (where you worked for 14 years before you went to Novell in 1997)?
Well, Sun was built 20 years earlier than Google, so it's not an apt comparison. And hardware companies are different than software companies. The most important thing about a hardware company is that it can, in fact, run out of cash. Hardware companies have large manufacturing pipelines; and the company nearly ran out of cash in 1989.
Software companies have the property that once they're profitable they are cash flow positive. So in that sense, software companies are easier, and Google is easier than Sun in that sense.
Culturally, Sun is more hierarchical. Two reasons for that: one is the age of the company. These newer, network-based models were not invented back then, and once a culture is set, it’s hard to change. The second is that Google can use small teams and make small, incremental improvements to a common, cohesive system. We do that very rapidly.
Hardware companies, in particular Sun, would have 500 people building a single chip. I was in charge of chips for a while. You have your 500 people, and they make a chip and they spin it (produce a prototype). Then they do it again making modifications, which are based on errors. That takes three months. And then they spin it again, which takes another six to nine months. Then they put it in the (system) board and they ship a million of them. If that cycle is violated, you fire the entire management team.
Because it has to be so precise?
Yes. And by the way, if they could do it without a second spin, you give them a huge bonus, millions of dollars. Now, that is so antithetical to how software works. It's completely culturally different. Sergey gives this nice speech about what is the biggest cost to Google. And everybody assumes it's like engineers or something. He says, "Opportunity costs." My answer, by the way, is taxes.
I always thought it was electricity.
Yeah, well, electricity, but whether it's taxes, electricity or whatever, he says, no, it's opportunity costs. So the boys understood this (difference between hardware, software and web companies) in their own way very young.
Did you understand that Google was a different kind of company when you arrived?
Yes. The key thing that I understood was that because Google was a Web Service company, its development cycle and its economics would be different, because I had lived in these multiple models.
You understood this five years ago? No one understood what a web service even was then?
Yeah, but I don't think it's a hugely surprising result. I mean, if you work in the industry, you think about the models. I think that for me the constant positive surprise about Google is having such smart people work for you.
I used to give this speech about how you control risk, and the best way to control risk is to have the smartest people working on the problem, and businesses are about risk. An example would be in my first year the question had to do with the development of these advertising networks. Overture was doing well. Would we partner with them? Would we compete with them? We had some preliminary conversations about partnering with them. They wanted to buy us.
And there were a number of times when I frankly had no idea what to do. But the solutions didn't come as a result of me sitting in a room and inventing it. I don't think Larry and Sergey sat in a room by themselves either. It was an iteration that involved the interactions of five or ten people over a period of time. And the characteristic of such group decisions done right is enormous buy-in.
The characteristics of most companies, in my observation, is that you don't have a lot of buy-in. The decisions are top down or they're driven from some external thing. People are like, “Oh, screw that, they don't really buy in.” But a decision-making process that is a sum of its iterative process, which is how Google works, generates commitment and passion that makes my job very easy.
Did Larry and Sergey understand all this stuff when you arrived?
They're very, very smart. Their instincts are very good. Whether they could articulate this concept then, I don't know.
Remember that their judgment is remarkable for people of their age. Remember, at 26 they could have taken this company public in 2000, made themselves what would have then been a fortune and had a really good time ---- and they actually decided not to do it. Their peers all went one way, and Larry and Sergey went the other way. And this predates me; that's a very important fact.
So the characteristics of Larry and Sergey that I don't think people really quite understand is that just like Bill Gates or Larry Ellison or Steve Jobs, they have a new and distinct view of the structure of the industry. They wouldn't say it that way, but if you go back in the history of Apple and Steve Jobs with him in the conference room, he's got his shoes off, he's wearing his black shirts, what was happening was that he was making a distinctive contribution to the industry. I don't think he would expressit that way but his biographers did. The same thing will be said about Larry and Sergey.
I am not like that. I am a more normal person. They are the founders for a reason.
Right, because on some level they're nuts. I mean, in the best possible sense. When you are an entrepreneur you have to be unwilling to take “No” for an answer – ever.
But a better way to say it is that they did not have a choice but to do what they did. I had a choice.
How would you articulate the new management model you introduced when you arrived?
When I showed up, I had been used to all-day staff meetings, and this company had no meetings at all. And I had just finished reading this book on the brain. And one of the results in the book was that people actually do better with more frequent, shorter meetings. This had been studied in numerous psychological tests. So I thought, well, why don't we have a meeting every day for an hour. It's a little company, we're all in one building. And I know that sounds stupid, but it's, in fact, countercultural. All thecompanies I'd been in had one day long meeting every two weeks starting at 8 am. Well, nobody showed up at 8:00 in the morning here anyway, so I decided that’s just not going to work.
So we eventually worked out the following schedule: Monday was a normal business meeting, Wednesday was products and Friday was ads, and they were in the afternoon. And the idea was that I needed Omid (Kordestani, senior vice president for sales and business development) on the financials on Monday and I needed Omid on the ads on Friday. But on the product stuff, although he had an opinion, it was better to ship him to New York or wherever (to make sales and business development calls). And that worked for himbecause he could fly back from Japan or whatever and show up Friday morning and come to a meeting and then go to his house to sleep. So it sort of all worked.
We did that for a while and eventually that stopped working because of the scale of the company. So now we do a Monday meeting, which is a business meeting, and then a Tuesday meeting, which is the product reviews, called Google product strategies, and Wednesday, which is more product strategy meetings. For a while we had meeting free Thursdays. Fridays tend to be catch-up meetings. So there's a cycle to the week.
There's a couple of rules. One: we don't have small meetings.
As large as we can fit in a room - 20 people. And the reason is that in a culture which is consensus driven, people need to see the decisions being made. So I don't say, "David, let's take it offline." We make the decision right in front of him. And he may or may not care, but when somebody comes up to him and says, "Well, were you there when they made this decision about" X, David can say, "Yeah". And it was a really, really loud argument, and I didn't really understand it, but they really had a big row aboutit.” So he's going to be a great witness inside the culture.
The trick is to have everybody participating in the decision and make sure everybody has been heard. The book Wisdom of Crowds (by James Surowiecki) says that you have to have two things to make better decisions in groups. One is you need a deadline, which someone -- at the end of the day is me or some external factor, and the second is that you need a dissident. There needs to be one person who will sit there in the room and say I disagree, and then the person next to him or her will say, "Well, yeah." See,everyone waits for the first person to be the dissident, and then they pile on, and that's how you have a healthy argument.
Most would say that sounds great, but it's also utopian. How does it work in practice?
I'm glad to hear that your ad campaign for Google is that Google is a utopia.
For starters, I hope I pay enough attention that I know who the dissident is in any meeting. It helps that I have a good memory for what Larry and Sergey care about.
So let me give you an example. We have an internal crisis at the moment, which has to do with the nature and structure of the internal databases. I'd rather not go into the specific technical details, but we have these large datacenters, and because of the speed with which we work, we end up with a multiplicity of databases. And the more databases you have, the more complexity you build into the system. Now, it seems obvious that if you were going to build Google today, you'd build it with one database, whichhad everything in it. It's obvious to me, obvious to you, seems obvious. So why didn't we do this?
So I happen to know that this is Larry's hot button. So if the meeting is getting boring and somebody talks about this, I'll say, "Larry, would you repeat your standard speech on the subject," and then Larry pipes up, because he cares a lot about it.There's a category of such examples. And if you work with people long enough, you know what their hot buttons are.
And I tried to explain this to you before with how Larry and Sergey and I work, but in this context it makes sense. We've worked together long enough to know our hot buttons. And we have a rule that you can't cut and run.
What is Sergey's hot button?
He cares a lot about the culture. He cares a lot about deal structure. He is the better negotiator of the three of us.
Larry is the deepest of the thinkers in terms of the strategic technology.
For example, today, I said, Larry "Larry, what are you doing tomorrow?" And he said, "I really need to work -- I need to go talk to these three teams." And he has this face - you work with people long enough you can sort of tell - he’s thinking “ I don't have good enough touch (with the teams) on an issue.” What that really means is Larry is going to go find the engineers and torture them until he is really, really sure (they are all in sync). And he's very good at that. He's very, very good. I mean, he is reallyextraordinary.
And that talent is a useful leadership skill because?
Well, here's another example to illustrate what I mean. We're arguing over the structure of our datacenters -- this was yesterday -- I thought the presentation was excellent. They had the datacenters, the components -- absolutely brilliant. And Larry says, "I just don't think that's a very good strategy," and I go, "What's wrong?" I'm a computer scientist. I think it's pretty good. And he comes up with this thing that nobody had thought about -- that he and one of the engineers had been cooking up, which involvedchanging the way the data centers are built that was the next logical evolution.
The point is he ignored every possible chain of command. He didn't tell anyone. He went to this particular engineer who's experimenting with this idea in his 20 percent time. So what I've learned here is you don't say “no” to that. As a manager what you say is I think that's a great idea, which indeed it was, and I'd like to see a report on this idea as part of our next review. So I've now closed the loop. Larry got it started. A month from now, when Larry is onto his next thing, I'll see a report on this particularidea.
The obvious other question is when does it break down? When it breaks down is when people are not around, when people are traveling and there's a crisis. There have been a couple of cases where we actually don't agree on a policy and we need to make a decision. Jonathan (Rosenberg, senior vice president for product management) once sent this note, where he said, "Look, this (indecision) is driving me crazy. And so I said, "Jonathan, here's the rule: when you see this (indecision), send me a mail message whichsays, 'You Eric have to force an outcome, because it's driving us crazy.'"
And so that's the other mechanism for the management team, which is when it doesn't work because there's real disagreement, then they give me the action item to force it. And I keep a list of such things, and I'll sit down privately with Larry and Sergey and say “What do you want.”
There was one case in February (2005), which is a very, very difficult issue, and I'd rather not go into the specific product issue, but we were in violent disagreement, and for very good reasons. Actually, Larry and I were in agreement. Sergey was wrong, in my view. Sergey came in and he was just on a tear. He had convinced himself of this and he was absolutely sure. And I said, "Look, I can't take this, guys. You guys have got to sort this thing out. If the two of you can agree, I'll agree with you. But youhave a deadline of tomorrow” Again here I am structuring it. "And you don't need to tell me the decision, you have to tell them the decision."
Tell the engineers. It's like 20 people who are completely demoralized because they're not getting clear direction. So what happens is the next night they get to the deadline -- and I remember it because this is an emotional issue -- and I called up Sergey, and I said, "What did you guys decide?" And he said, "I'm going there now to tell them," and he then described the solution, which was different than the three of us did -- and better.
So again the generic model is consensus building with dissent with a deadline. If you don't have dissent, stimulate the dissent, and inspect everything. That's sort of the default model, and then you manage the exceptions.
So the exceptions are when there really is a disagreement among the principles, and certainly I've encouraged people to say “I just don't agree,” and then every once in a while I've had to actually be a real CEO and mandate something and really force it and not listen to anyone else. Those are cases where it's legal or regulatory where I've just said, "Look, I'm just not going to participate in anything other than this outcome." And people know me well enough to know not to challenge me on it.
And at a certain point you have the responsibility to say, look, I'm not going to participate in this as the CEO because it's my head on the chopping block?
There is a position -- I'm not trying to toot my own horn here. There is a role of the CEO that is distinct, which is the CEO does have the ultimate responsibility for the outcome, and so there are some outcomes that are unacceptable, even if produced by consensus. So the good news is they're extremely rare, and none of them involve the product stuff and the deal stuff and so forth.
They sort of involve human resources and legal issues?
Yeah, being sued, being sued.
What did you do before Wisdom of Crowds came out?
Well, again, this model that I'm describing evolved. There's no single event that caused it. And one of the characteristics of high performing teams is they constantly are evolving the way they work. When we started the company the culture was really quite nondirective. Very “take its time”
I remember a little less than two years ago, I had worked myself into a state over the weekend over some list of things, and I decided that I was just going to go at it. So I had written this memo, I sent it out on e-mail over the weekend, which is when you got a chance to think, and Monday I said we're not going to do our normal meeting, instead we're going to do MY meeting and this is the list.
So I went through it top to bottom and everybody was silent because I was in my "I'm in charge and you're not mode." And so then at the end of it I said, "Okay, I'd like some feedback on this," and it was limited feedback.
Larry and Sergey's offices were next door, and I said to them after the meeting, "Well, do you guys agree with my list? Larry said, "You just gave a speech, Eric." Here's Larry, right, who's not exactly your professional CEO, right, and he's watching this and he agrees with me, but he completely disagrees with the style. So he waits till I'm finished and in private says, "You know, you could have done that differently." I remember this because I thought, “Wow, not only is Larry giving me feedback, but he's right.”
So the next week I said, "Okay, you guys didn't like my tactic last time, so next week we're going to have no list at all, you guys make the list." So I stood at the whiteboard and we wrote all the list down, had a nice conversation. So I said to Larry and Sergey, "What did you think of that," and they said, "Oh, it wasn't direct enough."
So we ultimately moved to this rule where I send the message out about the things I'd like to talk about and then we just go through them, but there's plenty of time for conversation, and it's not a lecture. I didn't ever have to do those things initially because the company was so small, but now it's big enough that you have to have action items and reviews and so forth and so on.
Another example would be when we had our first earnings call (Fall 2004). Well, everyone was upset about the earnings call, although the numbers were good, because we weren't organized and so forth. So my secretary goes, "You know what we do, we do product launches. Why don't we have an earnings launch?" And I go, "What's an earnings launch?" So it's an earnings launch review. So she controls the corporate calendar, so she puts a one-hour earnings launch review, and everyone shows up the next time and says, "Whatare we supposed to do in earnings launch review," and I said, "It's just like a product launch."
I'm not sure if this is helpful, but you asked me a question of how did this evolve; all of these are things that came out of a constant debate within the culture, and they're unique to the company.
Most executives of your caliber would have had a hard time making a triumvirate like Google’s work. How have you done it?
There's a bunch of stuff. Again, this is the benefit of having been around for a while. The situation where the outsider is brought in to work with the founders always results, in my experience, in one of the founders or the CEO leaving. The CEO packs the board with his buddies and then it causes the founders to leave. So I knew this coming in, well before I showed up on my first day here.
When I interviewed the board said this is a very good company, it has tremendous potential, don't screw it up. We need a CEO, but they were very clear that this was different from bringing in a new CEO to make some big change; this is an aiding and abetting the growth and success of the company.
And finally, when I was hired, and Larry and Sergey and I literally negotiated my package and what I was going to do and my role, at the end I said, "What do you guys think" And Larry said, "You know, I don't know that we need you now, but we're going to need you in the future." And that was exactly right. They were doing a great job at a 200-person company, but they didn't have the background to build a much larger company, and I wanted to work with them because I liked their technology, so it was a perfectmarriage in that sense.
How did you develop the kind of trust that you guys now have. It is not something that happens overnight, and you guys like to argue. How did you keep feelings from getting bruised?
What I liked from the moment I showed up here was there was never a situation where Larry and Sergey's incentives were different from mine, from a financial, a fame, impact, mission, you name it. I knew that whenever we would argue about something, we were arguing based on what was the right outcome for the company.
The most acute disagreement was over the original deal with AOL (in 2002). The reason was that Google did not have enough cash to fulfill the guaranteed payout in the contract (in the future). And the actual sequence is quite interesting because you see the beginnings of the management structure. So in the first place, the deal itself was a very good idea, we all agreed to it. Because the AOL people were very smart, and because Google was a new business, they negotiated very, very cleverly.
It was very shrewd. They got Google stock AND a minimum guaranteed payout from advertising revenue.
And by then we had expanded to two buildings. What I figured out was that this was a big enough deal that everyone had to participate. So every afternoon, at 4:00 we had the usual suspects, which were Larry, Sergey, myself, Omid, the deal team, and we would go through the latest iteration of the terms.
And it was clear that the three of us had very differing risk tolerances, I being the most conservative, as you can imagine. And I figured that the board would have my view of the world, because these are normal board member type people.
And I remember one day where we argued to the point where I actually stopped the argument. I thought it would split the company. It was the closest point -- and this is now four years ago, three and a half years go -- it was the closest point I ever saw literally the company falling in two.
In what sense?
Because the arguments got so heated. And I remember who was on my side and who was on Larry and Sergey's side.
Let me not name them. But it was literally half and half.
But what was heated about it?
Because of the guarantee …. we would have gone bankrupt (if we couldn’t meet it). I was furious. This is as mad as I've ever gotten, right, because it was so obvious to me (that this was wrong) So I stopped the meeting, and I said, look, we need to talk about this tomorrow. It was just not constructive. So that's a case where I was playing both protagonist but also sort of organizer type.
The next day after everyone had sort of calmed down, we came up with a mechanism that would meet all of our concerns, and we signed the deal.
So what specifically happened to kind of get you there?
I think what happens is when you get very smart people in a room, sometimes they actually get locked into points of view, and in that particular case I was on one side. Larry and Sergey were on the other, and we actually disagreed. And we were having an awful lot of trouble seeing the other person's view. And the reason I tell you that story is that it's a very good example of taking it to the brink and calling it (backing away), because the next sentence would have been a personal sentence like, "You idiot",which we never got to. And so that's where the line was drawn.
After you had like taken a day and come back, how did you get comfortable, what changed? Was there a modification in the specific deal terms that allowed you to come around?
Essentially, we compromised. It was a specific deal construct about how the cash flowed and what we did and what the guarantee was, and I don't remember the specifics.
But what were you the most concerned about?
I was the more conservative. And the funny story about this is I had assumed (the board would require convincing) So a week later we have the board phone call and I'm saying, "Look, I need to explain to you what we've done, I recommend this, we're all in agreement and so forth," and I laid out my arguments and why I was concerned about this, but I thought this was a good decision and that we had come to terms. And a board member said, "You're being way too conservative." All of a sudden I discovered that theventure guys were used to taking as much risk as Larry and Sergey -- that I was the stick in the mud.
But the reason I go through the whole story is that it's an example where there was never a question as to the goals.We have not had any substantive arguments like that in the last three years. We may disagree, but they typically defer to me for the things of which I'm the expert.
When you and I talked a year ago, you talked about how Larry was in Italy with a bunch of executives and how he went to Bologna where you grew up to find your house and get a picture of the crew in front of it.
Yeah, wasn't that nice? That was a nice touch. It wasn't deliberate, it was just they were actually interested in where Eric had grown up. That's a very nice touch.
But that's the sort of stuff upon which relationships are built. Are there some other examples like that that you can think of?
Well, Larry got two Canon digital cameras, so he gave me one and he said, "You should take some more photographs." And so I said -- well, I really wasn't focused on photography. So he had his camera and I had mine, and then so for periods of time we would take pictures together, and his pictures were much better than mine. So I decided to become a better photographer than he, actually studying it. And then, of course, got a better lens. And then he got this incredible high definition camera -- you know he's atechnology nerd, right -- he got this incredible high definition camera, and he said you've got to try this out. And, of course, Larry has it turned on in his office looking at the whiteboard. And I said this is the worst possible use of a high def camera. So I used it and shot a whole bunch of stuff and then he said, "Why haven't you given me back my camera," and I said, "Because you don't use it, Larry."
They're all really stupid like that, but I think in deference to their privacy and my own, I should probably not go into them. But I think a fair description is that our relationships are personal and professional and they make a lot of sense when you think about the nature of who they are. Although we're separated by age, I respect them greatly, and I hope they would say the same thing in return.
The reason the relationship works is that we really do agree. For example, I decided to put in this strategy process last year (2004), because the company is big enough, we didn't have a strategy process and because informal strategies don't work in the long term. And I could not figure out what to do. And so I go into their office - and you know their offices are together - and I said -- and it's like 6:30 pm, and I said, "Look, guys, this is driving me crazy. I'm having trouble writing down a compelling strategyfor what we're trying to do. I can do the normal strategy, but I don't have something really compelling."
So Larry sort of looks at me and goes, "Oh no," you know, because he wants to talk about some specific product stuff. And Sergey says, "I have to go." And I said, "Well, where are you going?" He said, "I have to go to dinner." Okay. "Well, tell you what, let me show you what I have produced." So I write the first sentence. Well, that immediately gets them going, right.
So Sergey, goes boom, boom, boom, boom, boom, boom, boom. And I kept my mouth shut. I erase what I wrote, which wasn't very good, by the way, and I write down everything Sergey says, and Larry doesn't say a word. And Sergey says, "I hope that's a good start, I literally am late, I have to go."
So Larry and I look at this list and he says, "Well, how do we translate this into action?" And so I said, "Larry, what do you think the first team needs to do and the second team," and he writes it down, and I write that on the next whiteboard.
Meanwhile, Sergey was confused, right, he had the wrong dinner -- it was a different night. So about 30 minutes later he comes in and he looks at Larry's list and he says, "That's exactly right." So I said, "We're done", okay.
Now, I then took the whiteboards, I took my computer and I typed them up. Of course, I used a little bit of license in tightening them up and generalizing the points and so forth and so on, and I said Larry and Sergey and Eric have agreed to the strategy of the company, here it is.
The company defines its mission as all the world's information. We choose to work on problems that affect a large group of people and problems that matter. Simple.
Another example of Sergey’s observations is that our advertising network is very powerful because it's quite resistant to certain competitive attacks.
Because it’s an auction market you cannot under-price it. This point is lost on many, many people. Another example has to do with geospatial data. This is before we bought Keyhole (the precursor to Google Earth). And Larry said that he believes that physical location and personal positioning, devices and information will become as important as information in its virtual location.
In other words, information that has a GPS tag to it – a photo taken at a particular location - all those kinds of things are very, very interesting. And he had spent a lot of time, I found out later, looking at the question of GPS enabled phones. And all of that predated the map stuff.
So if you assume that every document should also have a GPS coordinate, then you get his point. It's every photo, every document. So that document's current GPS coordinate is this room. A more interesting question would be where did you author it. Because the GPS location of its authoring might be of great interest. Historically, when Fred wrote the defining document of his career, where was he? It's a very interesting idea, it's Larry's idea.
So that little vignette is a very, very good example of how the three of us work. Here is Sergey, who's busy, and it just all pops out and it was brilliant, then he runs off, right, Larry looks at it, figures out all the details in his brain, I'm sitting there, I'm organizing and I understand my objective is to get something that then can be translated for the rest of the company.
Notice that it was their strategy that I adopted, but I also had the power of the pen, so if there was something that I thought was more important I could obviously emphasize it a little more. And when I showed it to them later, they said, "Yeah, that's roughly right."
Is one of them more of a conceptual thinker and the other more of a detail guy?
It has to do more with rapidity. Sergey is brilliant, pulsing, right. Larry is brilliant and deep. So Sergey is -- what is it, a supernova, he's exploding his ideas. So with Sergey you listen, you know, he'll be sitting there and he'll be doing something random, right, not paying attention at all, and then all of a sudden he makes this comment that is distinctly different from everyone else and is very incisive. It is how his brain works.
Larry would write a 30-page paper on his GPS idea overnight and not sleep.
And the more I work with them, the more I understand the nature of that brilliance. And it's not normal IQ. Let's be very clear here, there's a reason why these guys are a success.
To the degree that I'm successful, it's because I have been able to work with people like them in my career. I worked, for example, very closely with (Sun co founder ) Bill Joy for 25 years, who was just like this -- beyond intelligent. Bill and I did our thesis orals together, which is the hardest thing you do as a computer scientist, so I spent six weeks working, preparing, and it was very hard, and I did it and I passed. So I said to Bill, "Well, how did you do? Did you use the whole time?," And he said, "No,no, no, I used about half the time." And, of course, Bill got everything right.
He's just literally smarter than everyone. His processor runs faster. But he also was not as organized, right, so it's a different skill set. So you asked the question, how did I know how to work with Larry and Sergey. I've always worked in such arrangements. I'm used to this structure.
How is it useful to let others see you guys argue? In so many other corporations, as you've pointed out, none of that stuff happens.
Many, many employees in corporations essentially assume the senior management are jerks. It's just a true statement. And if you do a study, they say, oh, these guys are playing golf or whatever. So it's important that the people see the engagement. And in most cases the employees are perfectly happy to agree to follow whatever the leadership position is because they want to win, but they fundamentally want the leadership, whoever the leadership is, to be paying attention, so that's the first test.
The second is that I think it actually produces better outcomes because although I'm using the term arguing, the actual process is more than the three of us.
What do you mean?
Well, I'll give you an example. Omid is one of the world's most polite people, will wait and watch this, and then he'll say, "This is not okay with me, let me explain why," and you'll hear his voice. Jonathan is very loud in a good way and will just interrupt and say, "I can't believe this." So I talk about it as the three of us, but, in fact, the majority of it is an intense conversation. So if you think of this as conversation-based leadership where the ultimate authority is relatively clear.
How does it not wind up devolving into matrix-based decision-making where no one has enough authority or accountability to get anything done?
I think that's a legitimate criticism of the model and we do debate this. So the question is when we had -- you know, I don't know how many employees we have now but assume we double again, for purposes of argument. I think we have like 5,000 employees; if we have 10,000 employees (like today) how will the management culture change.
Take delegation. Historically almost all the decisions went to Larry and Sergey and me, and we're trying to now push every operational decision one level below or two levels below. And Larry yesterday, for example, said not enough people are getting in trouble here, you're not taking enough risk. His phrase was not getting in enough trouble, but that's what he meant. So we're aware of the issue and as the management team spend more time together, it should be possible to federate, literally to duplicate the model
And these guys should be able to federate these models and then talk to each other. Now, if they can't, it's because we didn't empower them, they don't understand the strategy, they're not comfortable with the culture or they didn't get enough of the other things that they needed.
What I don't understand is that the basic problem with matrix management is that it requires a huge number of people to get a project approved but only one person to blow it up.
Right, and that is the classic reason why consensus-based management doesn't work. So that may befall us as the company grows. Again, I make no warranties for the current model scaling It's working well now.
How does it wind up not becoming too hierarchical?
Well, the normal answer at our scale of company is that you divisionalize, you create general managers, you have different product lines and so forth. There are numerous problems with that; I went through this at Sun. The biggest problem is that part of Google's secret is that we get leverage across technology, the same datacenter, same distribution pattern, same sales force and so forth and so on.
When you divisionalize you wind up with no cooperation across divisions?
That's right. If we do any form of divisionalization, it would be a very light version of it in order to push decisions down, but it's clear to me that we're not going to divisionalize in the classic sense, it doesn't fit our model. We get such enormous leverage from centralizing these systems.
But I think your questions about organization are very appropriate, and I don't know that there are many companies that have grown this quickly in high tech. And almost all of the companies have ended up being one product companies with their secondary products not as successful as the first.
So it may very well be that we will hit such limits. Let me give you an example of where our model will break down.
The model that I've just described completely breaks down when you have a large amount of international engineering -- which we intend to do. So I project that everything I've just told you is going to break down.
Why does it break down when you have so much international engineering?
Time zones. You can't get everybody -- even on conference calls you can't get everybody in a virtual room to build a consensus. So if you go through, we have a big operation in India, Zurich, we're considering engineering in a whole bunch of other places in Europe, of course you have a very successful group in New York and Kirkland and Santa Monica already. So you now span time zones such that you can't fundamentally do an integrated meeting in that sense.
Why do it, given the downside?
Well, because we want the best talent all around the world. There's an enormous amount of programming talent in China, we want that talent to work at Google. And the fact is they want to work in China, they actually like China, it's their home, or the U.S. won't allow them in. So we have to adapt, just to be very clear here, we have to adapt the culture that I'm describing with this international development focus.
It's relatively easy to see how to manage international sales because sales is organized pretty much hierarchically. You have a country manager and you have somebody who runs all of Europe and somebody who runs Japan, somebody runs Asia, and we have that, and that's traditional.
Engineering, however, is done in this odd consensus way, right, unusual, and it's never been done with the kind of engineering, international engineering scale that we're proposing. It's a problem to be solved.
Is there something unique to the way Google is that enables you to have a flatter organization?
Yes, absolutely. There are a couple of things that we do that I don't know that we've talked about very much. We have something called a project database, which is visible to all employees, which lists all the projects, and we use that to manage a lot of the remote stuff. There's also something called the snippets database where people put in what they are working on. These two were Larry's, these predate me.
So one way, again thinking of solutions to the problem, one way you could solve this problem is by a culture which requires, if you will, people to write down what they're doing and then other people get a chance to see it, even if they're not in the same place. That seems kind of obvious but it's not true in almost any organization. At other organizations they can't see what the other organization is doing, and the CEO can't see either because the management prevents communication.
So you get a flatter organization -- flatness is not a function of reporting hierarchy, it's a function of information flow. And again it's clear that Larry deserves a lot of the credit here when he sort of came up with this.
How do people actually do 20 percent time? How do people actually figure out a way to actually get 20 percent of their time for that without working on weekends?
They work on weekends.
Do you compensate them in a way that encourages them to come up with these projects?
Yeah, but remember the kind of people who we hire are not here for the compensation, they're here for the impact. And there's essentially an internal draft system, that helps redistribute talent which is complicated and quite clever.
Do you actually have to declare what your 20 percent project is going to be?
People are encouraged to do so as part of the snippets.
Okay. That's the incentive.
But it's encouraged, not required. Again, there's things you measure and require and there's things that you encourage. The 20 percent is a cultural thing.
So you're encouraged to come up with an independent project, and if you're an engineer it's part of being able to sit at the lunch table with your peers and be respected?
That's right. Your peers all have one, so what's yours?
A couple of quick last things. You’ve talked before about how flying your own jet helps you manage. Explain. The reason I'm intrigued by it is because the only other guy I know who flies a jet is (Oracle boss) Larry Ellison. But you can look at Larry Ellison and understand how the machismo associated with flying a jet translates into Oracle’s culture. I don’t get that same sense of machismo talking to you about flying.
It’s helped me in the following sense. The difference between what I do normally every day and flying a jet is that when you are flying a jet people can actually die --- right now. So while it's very hard for any actual human harm to come in this hour with you, when you fly a jet that’s very possible.
And so in the jet training, and they do this whole thing called cockpit resource management, CRM, there's a whole protocol about how you handle life and death matters, what are the rules, and people pay attention and you would too obviously. I mean, it's a natural human thing. And by the way, people who don't pay attention don't pass, they literally fail them.
So I think the analogy is that it helped me have the confidence when I saw a situation where the right answer was clear to say it crisply. I know this sounds really stupid but intellectuals, and I'm an intellectual, tend to sort of wander, right. Whereas the rule in the jet is if you see it, call it, because we don't have time. Jet flying is all about lack of time.
And the other rule, and this is particularly true in difficult training situations, is it's “What's next, what's next, what's next, what's next.” Your eyes have to be constantly moving, and literally to the second.
And the problem, of course, is at the end of this you're very tired, but I think it's a reasonable metaphor for high tech management. The moment you let your eyes relax, there's something else coming. And by the way, it's okay for you to say you don't like that, but if you don't like it, then you can't succeed in it.
So we're always on. Everyone is on their e-mail 24 hours a day, you know, people work at midnight on a Saturday night, that kind of thing, and it's expected.
Did you learn how to fly a jet because of all that or because you just thought it would be fun?
In my case I just thought it would be fun, but it's an interesting corollary. I actually started it because when I was in Novell I needed something to distract me from the pressure of the job. And the great thing about flight training, and you know this because we've talked about this before, is that the great thing about flight training is that because lives are at stake you can't daydream. So another way of thinking about it is when we're in one of these situations like the ones I'm describing to you, nobodyis daydreaming, nobody is penciling, everybody is engaged. I try to personally model that behavior.