Shishir Mehrotra on Scaling YouTube and The 10 Things That Matter | Blitzscaling 13

This is session 13 of technology enabled Blitzscaling, a Stanford University class taught by Reid Hoffman, John Lilly, Alan Blue and Chris Yeh. This class features a guest lecture by Shashir Mehrotra, who helped guide YouTube through hypergrowth after its acquisition by Google. At the end of the class, Shashir is interviewed by Allan Blue. This podcast has been produced by Greylock Partners. For more podcasts, class notes, slides and videos, please visit greylock.com first off, just I'll introduce myself a little bit. So I'm Shashima Rotra. My background I went to school out east at mit. Don't hold that against me here. I left school to start a company called Centrada that was back in the 2000 cycle, then spent a few years at Microsoft, worked on Office and Windows and SQL Server, then spent six years at Google mostly working on YouTube. That's mostly what we'll talk about today. And I left Google about a year ago to start a new company and I don't actually plan to say much about that today. So let me tell you a little bit about what I'm going to talk about. So over the course of six years at YouTube we learned a lot of different things and what I did was I tried to pick out some lessons from that period just to give you context. I joined YouTube 2008 and people may forget, but YouTube was in a very different state at that time. YouTube had just been purchased by Google, a little over 100 people at the time, actually not seen as a success. It was seen as a great exit, but seen as a company that may still fail. There's a lot of people who thought that Google made a big mistake in the acquisition and six years later it looks very different. We'll talk about some of how those changes happened. I wrote down a couple notes of preamble before we get started. Let's see. So oh my role at YouTube. So my role was a little bit non traditional. So I thought I would explain a little bit how that worked. So officially I was responsible for product engineering and user experience. We actually ran YouTube as a triumphant so a guy named Salar Kamangar was what you would call a chairman CEO and then day to day operations were run by myself and a guy named Robert Kinsel. So Robert generally handled the business side. So he'd handle sales, marketing, content partnerships, so on. And, and I handled the product side and the three of us did a lot of stuff together. So these stories are mainly coming out of those experiences. One other caveat is I made no attempt to be Comprehensive. I think there's lots of things that matter as you scale. I prioritize counterintuitive and interesting. In fact, the way we put the talk together was last night I sat down with a friend with 20, 25 stories and he stack ranked them for me. So you're going to see the 10 that are interesting. They may seem somewhat random in different order, but hopefully stuff that is counterintuitive and then goals. So I have two goals for today. I imagine this room has, Reid was telling me, has a wide group of people at various different points in their career. The way I think about this is as you hit those moments of scaling, my goal is to take some of these stories and a, help you recognize when you see one of those patterns and B, give you a set of tools and lessons to apply at those times. So I'll try and do that throughout the talk. And I think that's it. Let's get started. So these are the 10 we're going to talk through. I said they're not sort of well structured, but I did find a natural grouping. So the first two I call choose wisely. And the general theme there is, I don't know how many people here are poker players, but every poker player will tell you learning how to play poker is partially about learning how to play the game, but mostly about learning how to pick which table to sit down on. So I'll talk a little bit about how you pick which table to sit down on. Charting the course I think of as the core of running a business. How do you make hard choices, how do you think about ecosystems, value, so on. And then the last three, I'll focus on some team related things. How do you pick people, how do you organize them, what's your own role in those situations and so on. So that's gonna be the 10. Good. All right, let's start. So tailwinds. So let's start with tailwinds. The main less the story I want to tell here is a bit about the YouTube tailwinds. And really the main point I want to talk about here is how to pick businesses that have great tailwinds. So we'll start by looking at how we used to describe the YouTube tailwinds. So this is how we described it for YouTube. So back in 1988, number one show on TV was the Cosby Show. Cosby show got a 25 point rating, had about a 50% household share. For people not familiar with media metrics, what that means is if your TV was on on a Thursday night at 8 o', clock, there was a better than average chance you were watching the Cosby show. You fast forward 10 years, the number one show on TV was Seinfeld. Very similar ratings, very similar household share. But if you go 10 years past that, you get to 2008 and you see American Idol with a 12 point share. Now, one interesting side note in this story is Seinfeld in his first few episodes actually didn't do that well. It's piloting a few episodes after that. And so Seinfeld was almost canceled by the NBC executives. And if you go back and look at its ratings during that period, Seinfeld actually had better ratings than American idol had in 2008. So the new normal was set in a very short period of time. So one obvious question is what happened? Where did all those people go? Did they stop watching television? The answer to that is no. Actually, through that whole period, the number of hours a day the average American family spent watching television actually went up every single year. We what seems to have happened is this bottom line, that the choice all of a sudden changed. And this chart is just looking at cable networks is about in that time frame, 2008, 2009, and this is just plotting cable networks on one axis and audience share on the other. And just shows that about 50%, if you take all the networks that had less than 1% share, they accounted for about 50% of viewership. And ones that had less than half a point share counted for 30% of viewership. Half a point is an important mark because it's actually where Nielsen stops measuring. And so that's about the amount of viewership that's sort of unaccounted for on television. So how does this apply to YouTube? So the YouTube thesis mapped directly to this narrative and the way we would talk about it is if you look back over the history of television, there was a long period of time where we had three major networks, Fox got added in later, and then over a short period of time that turned into hundreds of cable channels. And, and the YouTube tailwind was this trend we saw that we thought, or think that the same transition was about to happen in the online video world, or to say it differently, we believe online video is going to do to cable what cable did to broadcast television. We were going to go from tens of channels or ones of channels to tens of channels to thousands or millions of channels. Ok, so why is this? Why, why does this matter here? One of the things I've been able to do across my career is work on lots of different Types of products, some that have great tailwinds, some that have great headwinds. And if I had to give one clear lesson, it's work on ones that have great tailwinds. There's a few reasons why. One is that it accelerates everything you do. You can do little things, and they become amplified. But probably the biggest reason I say is it covers a lot of mistakes. And in the period of building something, you're inevitably going to make mistakes. I think picking things that have those tailwinds makes it much easier to weather through them. So that's lesson number one. When you're picking which table to sit at, pick ones that have tailwinds. All right, number two, let's talk about purpose. So Daniel Pink wrote this book. Anybody read this book? So it looks like a few. It's called Drive. It's about what motivates people. The book's pretty good. If you aren't that interested in reading it, go watch. There's a YouTube video of it. And if you're even less interested, I can summarize for you quickly now. So the book is basically, he says what motivates his theory, what motivates people is three things. Mastery, autonomy, and purpose. Mastery is the being good at what you do. Autonomy is having the room and space to be able to do it well. And purpose is feeling a connection, feeling a connection between that and impact. And he tells a lot of great stories about how to look at each of those different things and how it impacts motivation and so on. Now, over the years, I've come to the conclusion that of those three, the one that matters the most is purpose. And so I want to tell you a little bit about how I thought about purpose at YouTube. And the story I want to tell is about this fellow who knows who this guy is. Okay, Everybody knows. So that's Sal Khan. This guy, I assume less people know. His name is Imran. That's Sal's eldest son, and he's actually involved in the story, too. So let me tell you a little story about some of the early days at YouTube. So Sal and I actually went to college together and stayed in pretty close touch. We both ended up marrying our college sweethearts, actually. Both our wives are physicians. Moved across the country about the same time. We went different paths. He ended up in the hedge fund industry. I ended up in the technology industry. But we would get together regularly. In 2008, I had just joined YouTube, and he and his wife were over for dinner, and so conversation started, and I happened to mention that I had just joined YouTube. And he said, oh, you know, I actually use YouTube a lot. And I said, great, that's wonderful. And he said, no, no, no, I use YouTube. Like, I upload to YouTube a lot. And he was trying to explain what he was doing. And for people that don't know the story, he's got this cousin in Louisiana that he was helping with math homework, and the easiest way he could find to tutor her was to have her send questions, and he'd post answers as YouTube videos and send them back. And really he was using YouTube because it was an easier way to send a video file. He said, you know, it's kind of interesting. I've been doing this and a lot of people seem to be watching it and. And the conversation kind of moved on. Next day I go to work and we had this dashboard we could use to look up how channels were doing. And this is 2008, so this is Stanford and MIT had both just started uploading all the footage of all their classes to YouTube. And I go and I did some quick math and realized that actually Sal was getting more viewership than Stanford and MIT combined. So I write him a note. I say, I don't know if you realize this, but you're kicking Stanford at MIT's butt. You have to get into the YouTube partner program. @ the time, you couldn't make money on YouTube without being in the partner program, and you had to get invited into that. So we invited him to the partner program and he accepted and started making money on YouTube. So fast forward a few months later, he and his wife are back over for dinner. His wife happened to be pregnant, actually, with Imran at the time. My wife was pregnant with our younger one at the time. And again, conversation turns to YouTube and he says, sure, thanks for setting me up in the partner program. It's been a great experience so far. And then he kind of pulls me aside a little bit and he says, you know, I've been doing some math and looking at the checks coming in, and I think actually we're really close to them. You know, I can see clearly where this is going to cross our rent. And I've been thinking about it, and I can see a little bit further out it might cross my hedge fund salary. I don't really like working at hedge fund, so I'm thinking about quitting and doing this full time. And he said, what do you think I should do? And this is the moment where his wife, Umaima, gave me the stare of death and said, answer this question properly. And Thankfully, I told him I just bet my career on it. I would say I would be encouraging of you doing the same. To be clear, Sal's achievements are all his own. And I think I had this one little tiny moment where I gave him a little extra nudge, but he did end up quitting. And the rest of the story, it sounded like everybody here knows and Khan Academy now reach millions of people and have tens of millions of users and so on. But the interesting part about that story is imagine if this had played out just a few years earlier. Imagine a sale had said, I have this great idea. And he'd gone to a cable network and he had said, hey, I want to create this new educational show. And they would say, great, do you have any experience as a teacher? And he said, well, not really, but I think I'm pretty good at math. And said, ok, ok, do you have any experience as an entertainer? And I said, no, not really, although he is pretty funny. And he said, ok, well, tell me about the show. And he said, well, here's what I'm going to do. I'm going to start at the beginning of the algebra book and I'm going to start solving problems. And when I get to the end of that book, I'm going to pick up the next book and I'm going to do the same thing. And by the way, I don't want to show my face in any of the videos and I don't want to have any storylines or plots or anything. Will you fund my show? I think he would have gotten laughed out of just about every cable network out there. Thankfully, he didn't have to pitch anybody. And this is a theme we used to refer to as a gatekeeper less world. And he ended up being able to start his career on YouTube. And now he's arguably one of the biggest educators on the planet. Now, the reason this matters here, fast forward a few years later to 2010. I actually was, let's say, not feeling so good at work and was debating making a switch, being recruited by another firm, thinking about leaving YouTube. I have an advisor, a guy named Dean Gilbert. So I called him up and said, what should I do? And he asked me to make three lists. And I think this would be helpful for you guys as well. So list number one, he said, what are you hoping to accomplish in the next year? List number two, he said, okay, take that, take list number one and figure out which of those things wouldn't happen if you were to leave. And list number three was take list number two and Say, which of those things would you feel really terrible about? Like, not terrible about the next day or the next week, but like months and years later, how you know, which of those would you feel really terrible about? And, you know, list number one, you could say is goals. It's focus. So on list number two, I now refer to that as the marginal utility list. If you're in an environment where you're not actually having impact, if all the same stuff would happen with or without you, then you have to question why you're there. But. So list number three actually ended up being the most important. And I call that list purpose. You could also. Sometimes people refer to it as a tombstone test. Only work on things that you want to have on your tombstone. I think that's a little morbid. So I like to use the word purpose instead. But one of my lessons was, I think no matter what you work on, whether it's a small, risky startup or what seems like a big, stable company, you'll have moments of lack of confidence and you'll see challenges that. That may not seem like ones you can overcome. And the things that get you through those periods are your sense of purpose. So when you're starting and figuring out where you want to sit, I think picking things where you feel that connection to purpose, I think is really important. All right, that's number two. Number three, let's talk about theses. So I'm going to talk a little bit about how I ended up at YouTube. First off, I just wanted to set some context on what it felt like when I got there. So these are some of the headlines that my mom would send me as I joined. Are you sure you're doing the right thing? And like I said earlier, it was far from obvious that YouTube was going to be a success at the time. Now, there's another book. I'm assuming a lot of people here have read this one called Zero to One. It has. This has a bunch of really interesting ideas. There's one I really like, which is this. What important truth do very few people agree with you on? I thought that was a really good way to describe a thesis. And so one way to talk about my story of joining YouTube is it was kind of crazy to join YouTube at the time, but once I had a view on it that not a lot of people agreed with, it was kind of inevitable. So my story actually has to do with the Super Bowl 2008. I'll come back to that in a second. So 2008, I decided I wanted to leave Microsoft. I Thought I wanted to go start a company. And so I started coming down to Silicon Valley, meeting different people and so on, and happened to get called by this guy, Jonathan Rosenberg, who's an old friend of mine, happened to be running product management at Google at the time. And he said, you got to come in and meet us at Google. I said, well, I don't know. I want to go start a company. I don't really want to go be part of a big one. He said, no, no, no, just come by and meet us. And said, all right, I'll come by. And so went in. And of course, if you think Google was exciting now, I mean, it was incredibly exciting in 2008, and everybody's pitch was really out there. Sergey had his set of things he was working on. Eric had a set of things he was working on. Vic had just joined, and he was doing a bunch of new things and so on, but all of them were kind of all over the place. And the last meeting of the day, I get together with Jonathan again. He said, what did you think? I said, I don't know. They all seem kind of crazy ideas, but I just think I still want to go do my startup. And he said, no, no, you don't understand. And for those of you that Jonathan's not speaking in this class, is he. He's actually a really good speaker for things like this. But Jonathan says things I'd say directly. Another way to put it, sometimes crassly, he said, you're totally missing the point. He said, google is, at its heart, not about Chrome and Android, all these other things. Google's an advertising company, and all the money in advertising seems to go to all these stupid TV ads, and nobody even watches those TV ads and used a bunch of words I didn't use. So I had never bought or sold an ad in my life. I knew nothing about advertising. So he said all that, and it's actually all surprising to me. I had no idea. All the. Like, in retrospect, it's kind of obvious. All the money in advertising goes to television ads. It's kind of obvious. But for me, I didn't know any of that. So I get on a plane, leave his office, get on a plane, and I start thinking about. Start thinking about what he said. And this happened to be just a couple weeks after the Super Bowl. And I don't know if people remember the super bowl, but it was actually a pretty iconic one. If any Patriots fans are here, I'm sorry. I started thinking about the super bowl, and we tend to throw a party for every Super Bowl. A bunch of our friends would come over, we'd all watch together. And I was thinking, he said, that nobody watches those stupid TV ads. And yet that one day, for those three hours, we watched every single ad. In fact, I remembered many times when people had said, hey, could you rewind it so we can see that ad again? And so I took out a sheet of paper and I wrote at the top. What would it take to make television feel like the super bowl every day? What's different about that day? And how could we recreate that experience every single day? So I wrote down a series of ideas, went home, went to sleep, woke up the next morning, looked at that set of ideas, looked at my list of startup ideas and said, ah, I don't know anything about advertising. Let somebody else do that. And so I write Jonathan and said, I don't think Google's for me. Thank you very much for hosting me, by the way. You mentioned something. I had a few ideas for you. Whenever you want, I'm happy to give them to you. I'm sure you already thought of them, happened to call me back, and I walked him through some of those ideas and he said, you may think they're obvious, but they're definitely not obvious. Here we have nobody working on anything like that, so if you really want to do that, you should come do it. And through a series of conversations, he convinced me to give up on starting this company. And I ended up through that, ended up working on YouTube. Now those set of ideas became something called TrueView. The basic idea was really simple. The idea was you should put a skip button on all the ads. That's what the user sees. That insight by itself, sort of not enough. The core of it was the idea that the way to change advertising, especially video advertising, was to change the incentives. And the reason I felt that the super bowl ads were so much better then the rest of the year was that the incentives were different and for a bunch of other reasons we can talk about later if people want to hear about them. But what we wanted to do was recreate that experience all year round. And so the idea was put a skip button on ads, but also don't charge advertisers if people skip them. And what that did was it created a completely different set of incentives. Now YouTube had no interest in showing an ad that people wanted to skip, and the advertisers had to build better and better ads if they actually wanted them to get watched. It was completely counterintuitive. I showed up not only. I mean, Jonathan wasn't underestimating it. Nobody was working on it, nobody believed in it. In fact, they were incredibly negatively predisposed to it. I got in front of a sales team in a room kind of like this and they kind of tore me to shreds. I said, are you crazy? You're going to put a skip button on ads? All you're going to do is erode all the revenue and while everybody skips their ads. Thankfully, it was the right insight and ended up working. And now Truview's a multi billion dollar business. And the metric we mostly focused on was revenue per hour for an hour of viewership of YouTube versus an hour of viewership of television. How did it compare? And by the time I left, it was about even between cable television and YouTube, which is a pretty amazing thing to think about with a set of skippable ads. But one of the lessons I learned from that is, is sort of back to the statement is, you know, Peter wrote this with a view towards starting a company. I think the same statement applies as you're scaling a company as well and that the most interesting innovations happen when you look hard at a problem and you come up generally with very simple insights. And those simple insights will often take a lot of time. And trufy took us almost three years to get out, much less to get to, much less to get to scale. But the core of really making progress is often finding those insights and really sticking by them. So as you think about scaling and what you're working on, a lot of the same principles for how you start apply as you scale as well. All right, number four metrics. Yeah. What were some of the things that Jonathan did to get. What did he do to. The question was what did Jonathan do to convince. You want to take a couple of questions now or you want me to wait? Okay, your class, you can decide. What are the things Jonathan did to convince me? I think mostly he fed my interest in solving this problem. There was some piece of it was confidence in myself. I walked into it saying I don't know anything about the space. And he gave me all the examples of all the people that have been successful in the space without knowing it and why he would rather have someone that doesn't know it but has an interesting idea for it than otherwise. And the other half of it was probably confidence in being able to do it in a big organization. And so he ended up handing me a team of people that were sort of cordoned off and then actually turning it into YouTube happened later. We started in a slightly different space, actually. That project ended up becoming a thing called Chromecast, those little dongles that sit in the back of the TVs. But he sort of addressed both those things. All right, so let's talk about metrics a bit now and why metrics matter. And so for this story, it's going to sort of fast forward time a bit to the 2011, 2012 period. Now, I call this the wandering period for YouTube. If you think about like 2008 to 2010, we kind of had a fire under our ass. And it was very, very clear that not only did the world think we were going to fail, that Google thought we were going to fail, it was not. You know, it was incredibly motivating to get through that process. But now we hit 2011, 2012, we hit a lot of our goals, you know, done a bunch of celebrations. We were well past the point of sustainability. Nobody was talking about shutting us down anymore. And all of a sudden, pace slowed and it wasn't very clear how to get it back. And that sort of manifests itself in a lot of ways. I think one observation was we had no real competitors. And I think I've managed to work on businesses of lots of different types. Ones in spaces that are incredibly competitive and ones in spaces that are not. Ones where you're the leader, ones where you're the follower. I think everybody thinks that it's much better to be the leader and be number one or clear number one in a market. And I tell you, there's a lot of downsides to that as well. One of the biggest is you sort of lose your sense of direction, you lose your sense of pace, of motivation. It's actually incredibly motivating to be number two to a fast moving number one. And we were feeling that. And so we were discussing what to do about it, and two stories came up that I thought I'd share before I tell you what we did about it. So number one, I'll come back to that guy in a second. So number one was Coca Cola. So there's this whole story about Coca Cola that they were having a strategy meeting and somebody asked, how long are we going to keep going back and forth on we're 51% share versus Pepsi and 49% or whatever that is. This doesn't seem like this is ever going to go anywhere. And somebody said, maybe we're looking at it wrong. Maybe we should measure Coke as percentage of the stomach. And. And by that measure, it's like very small percentage Share. And that led to Coke investing in all these other businesses. And now they sell water and lemonade and all these other things. And so that was one kind of interesting thing sitting in our head is maybe YouTube is the clear number one, but maybe we're framing the market wrong. The other one was a story of this guy, this guy named Ben Hunt Davis. And so he was on the British rowing team. And in the late 90s, this team got together. This is a pretty famous story as well. You can look them up. Got together and decided that they were going to fix things. So this rowing team in Britain, and rowing is pretty popular there, hadn't won a title in many, many years, the Olympics or any of the worldwide championships and so on. And, and so they got together and they made this decision that they would make every decision with a single question. And that question was, will it make the boat go faster? And they're going to sit down and say, should person A or person B be in the back of the boat? They would say, well, all that matters is will it make the boat go faster? They'd say, where should we go to dinner at this restaurant or that restaurant? They'd say, well, it doesn't matter for any reason other than will it make the boat go faster? And they try to apply this theory and it has a happy ending. They ended up winning. But these two thoughts, these two stories really stuck in our heads as we thought about what to do. And what we ended up doing had something to do with a billion, which is why that guy's up there. We got together at our YouTube leadership off site and we announced two things. So number one, we announced a single unifying metric. We said, from now on, our version of will it make the boat go faster is going to be watch time. And we're going to make every decision based on whether or not it's going to increase watch time. I'll come back to why that's important in a second. And secondly, we set a goal and we said our goal as a team is to get to a billion hours a day of watch time. We put a time frame on the goal, but we didn't announce that publicly, so I won't talk about that here. We said we were going to get to a billion hours a day of watch time. Now just, I know that probably is a hard number to understand, so I'll give a little bit of context. At the time, we were doing about 100 million hours a day of watch time. So it's sort of 10x more than where we were at some of the comparables at the time, Google was in roughly the same ballpark. But it's kind of a dumb metric for Google because the whole point of Google is get on and off the site as fast as possible. There was no other Google property even close. So we're by far the largest at Google. In fact, in the rest of the Internet, the only property that had more than us was Facebook. And Facebook had about double that amount of watch time at the time. And so we set this goal and said, these are some of the comparables. But the way to think about it is that television at the time was about five and a half billion hours a day of watch time. So YouTube was about 100 million. Television is about five and a half billion hours a day. And so we said, if we're successful at this goal and getting to a billion hours a day seemed crazy, we would still only be 20% of television. So it's sort of our version of 1% of the stomach. So a few things about the goal. So first off, it was crazy. My head of data science got up before we announced it, and then after we announced it, told me I was crazy, said, it's never going to happen, certainly not in the time frame I had suggested, and maybe not ever. There was a bunch of very interesting implications for us to hit that goal. It would mean every team doing something completely differently. Just to give one example, at the time, YouTube streams a lot of. Obviously streams a lot of video, uses up a lot of the Internet's bandwidth. At the time, we were using about 20% of the Internet's bandwidth. It's a little bit hard to measure the Internet's bandwidth, but that was our sort of, sort of closest approximation in order to hit this goal. Based on the pace that the Internet was growing, the pace that YouTube would have to grow to hit this goal, the net impact would be that we'd be streaming at a billion hours a day. We'd be streaming roughly double the amount of capacity of the total Internet at the time. So the networking team had to get to work. And we kind of went through every team and described what this goal would mean for everyone now. So we made these two announcements. Watch time is going to be our single clarifying metric, and we're going to get to a billion hours a day of watch time. A couple things happened out of that announcement. Number one, it greatly clarified decision making. All of a sudden we would sit down in reviews and we'd argue about, and I think you'll see this as you scale Your businesses. One of the characteristics of every business these days is that you have tons and tons of data. And so you sit down in a review and you'd say, should we do A or B? And you'd say, well, this is what it does to views and this is what it does to revenue. This is what it does to profit. This is what it does to content creators. This is what this content creator versus that one or this group of users versus that one. And you could argue about it forever. And what it meant was lots of lack of clarity in decision making. One good example that I like to ask people is why do you think YouTube had a reputation for being a place for short form video? The restrictions on uploading only 10 minutes of videos, 10 minute long videos on YouTube were long gone at that point. And yet YouTube still had that reputation. And the answer was very simple. We had trained every algorithm to optimize for views, not watch time. And so we would go to creators and they'd say, oh, I built this great one hour documentary and I cut it up into five minute sections because I wanted 20 views on it, not I wanted 12 views on it instead of just one. And so we had to retrain ourselves, we had to retrain the marketplace and, and change how we thought about those incentives. The other thing that happened was it completely changed pace and that mission got embedded in every person of the team. Actually, if you walk around Google and ask people what YouTube was working on, they say, oh, they got that billion hour thing they're working on and you could sort of, you get immediate recognition from everybody. And what that meant was not only did pace increase, people started suggesting bolder initiatives and all sorts of innovation started happening again. And you got that feeling of a startup again. You got that feeling of what does it feel like when you have a goal, when you have a competitor. So the reason I talk about this one is I think as you grow, you'll see lots and lots of different metrics. I think picking a single clarifying metric is really, really hard. And I would argue most of the business I've been involved in before or after actually have struggled to do this. So I don't mean to imply that it's easy, but the implications of having that are, are great both for clarifying decision making and for helping people understand the frame of success. You know, what is the shape and size of success that you're aiming for? All right, number five, let's talk about making decisions. So one fun question I like to ask the team is what makes a good decision you're going to be in these roles, you're scaling all these, scaling these businesses. What's a good decision? And typical answers I would hear back are, you know, a good decision is one that's made quickly. That's probably the most common one. Another would be a good decision is one that's made with all the right people's input. My view of it is actually neither of those questions is a very good one for whether it's good decisions. I think there's a single question for whether you've made good decisions, and that is, did that decision stick? That's what the duct tape is for. It's the best metaphor I could come up with. So decisions that stick. So I wanted to tell a story. This is from the early days of my time at YouTube. So this is in the 2008-2009 period. One of the product managers had a suggestion, said, I think we should link out to other video websites. The typical example is you come to YouTube this back in that time frame, you come to YouTube and you search for Modern Family, which is a popular show at the time, still is. And on YouTube you get a bunch of garbage. And if you went to Google, you'd get links to watch Modern Family on other properties. You get it on abc, you get it on Hulu, wherever else. And this person's argument was really simple and compelling. Says, hey, we're YouTube. We're owned by this company called Google. Our job is to give people answers. We should give them the answer they're asking for. So they come to us and say, where do I watch Modern Family? We should tell them and we should direct them to those places. Pretty compelling argument. Of course, we didn't end up doing it, which I think everybody knows. Let me explain why. So there were a bunch of. There were a bunch of business people in the room who said, oh, if you do that, it's going to confuse our content partnerships in all these different ways. And those people were mostly ignored. That's not how we made those kind of decisions. But then somebody brought up this different argument. They said, hey, let's use the analogy of Amazon and Google Product Search. By any sort of logical conclusion, Google Product Search should have kicked Amazon's butt. Google Product Search is a full superset of Amazon. So why would you ever go to Amazon and look for a product when you could come to Google Product Search and find everything on Amazon and everything everywhere else on the Internet? Seems like a really sound thesis was completely wrong. Now, why is that? And in fact, you could apply the same argument to Google Video and YouTube. And the fact that Google had to go purchase YouTube when they had a great Google Video property is telling as well. Same pattern. Google Video search indexed all of YouTube, so it had everything on YouTube plus the rest of the Internet. But the net impact was that the narrower property actually won. And we came to this conclusion that in certain markets, consistency beats comprehensiveness. And in Amazon's case, that's easy to describe. Why do you go to Amazon instead of elsewhere on the Internet? Well, you get a consistent experience for buying a product. You understand how the reviews work, you know how shipping works, you know how returns work. And all of those things actually trump the fact that they might be missing something. And basically the same thing was happening in the video space. And we believed that was going to continue. Probably a separate discussion about whether this applies to all markets or not. But we were quite convinced that it applied to video. So we made the decision and we didn't add external links to YouTube. Now, the importance of that story is, of course, the decision stuck because there's still not those links on YouTube. But the real importance of that story is, is the implications it had on future decisions. So as an example, the next thing that happened was we went and at the time we had a bunch of embedded video players on YouTube. We went and pulled all of those off. The next thing after that was the team came to us and said, we'd like to remove an option. At the time, we were giving video uploaders the option to decide whether videos were shown on mobile devices or not. We said, we'd like to remove that option if we're going to. If we're willing to sacrifice the size of our corpus in order to have a more consistent experience, then shouldn't the same principle apply? We should be willing to take the set of people that don't want their videos on every device and say, you can go host it somewhere else. Similarly, we did the same thing on interface at the time YouTube. The only property we built for YouTube was actually the webpage. Every other YouTube experience was built by that manufacturer. So the the Xbox team built their experience and each of the TV manufacturers built their experience, and Apple built their experience and so on. The Apple one is probably the one that's been talked about the most. We had to go through a very tough negotiation. Since Apple had been building the YouTube experience since the very first iPhone and pulling it off. That meant a lot of implications, including not being default installed on the iPhone, which seemed crazy at the time. But we went and made that decision as well. So we'd rather have a consistent experience across all devices and we'd sacrifice. We'd sacrifice distribution in order to do it. So the reason I mention is you'll find yourself in these roles and you'll make thousands of different decisions. You make them all the time. And it's often hard to figure out what are good decisions. And my main lesson was the best decision is not only the one that get made quickly and so on, but rather the one that sticks that makes it easier to make the next 10 decisions. Or in the very best case, it actually obviates the need for the next 10 decisions because everybody already understands the principle of what you're trying to get done. All right, number six, ecosystems. Anyone know who this guy is? All right, little hint, his name is Bauer. Now, does anybody know who this guy is? All right, somebody new. All right, so he's one of those guys. The actually the song behind one of those guys, I'm going to turn that off because that's disturbing. So I want to talk a little bit about ecosystems. In particular, I'm going to talk about the music ecosystem. So the music ecosystem and YouTube have a very symbiotic relationship. So the music industry really needs YouTube. If you go look at a bunch of different reports, Nielsen will tell you that that people under a certain age tend to use YouTube for music more than the terrestrial radio. So the music industry has very much aligned with figuring out how to make this YouTube thing work. The same was true in reverse for YouTube. Music is a very popular genre, but even more importantly, if you upload a video to YouTube that has music in it, we would license that on your behalf and share the revenue back with the music labels. The so that meant not only did we have to deal with Taylor Swift's music video, we had to deal with the birthday party video that had Taylor Swift in the background. So this was a really important symbiotic relationship, which meant that the negotiations were always really tough and they would happen about every three or four years. So 2011, we were going through our set of negotiations with the music labels, spent an all day session talking through lots of different things, and then afterwards we went out for drinks. And I happened to be sitting next to one of the other label executives and I asked him this question. I said, I just want to know that we just came out of a day long, very adversarial meeting, said when is that tone going to change? And the way I phrased the question was, when is it going to be the case that your relationship with YouTube and your ability to be successful on YouTube actually helps you win clients. The way a label works is they have to convince artists to join their label. Generally the way they. Historically, the way I used to convince them was, hey, we have relationships with all the record stores. And then over time, it became, we have a relationship with all the radio DJs. And I wanted to know, when's it going to be the case that you say, you should be signed by me because I know how to make you successful on YouTube? And I threw out some hypotheses. I said, is it going to be like, a certain amount of revenue? Is it like a certain percentage of an artist's revenue has to come from YouTube? Is it going to be a certain amount of viewership like that it has to reach a certain scale? And he kind of laughed and said, no, no, no, it's not any of those things. He said, you probably don't want to know the answer. I said, no, I want to know the answer. And he said, I'll tell you, it's one thing the artists don't really. The revenue piece is not really that important because one of the secrets of the music industry is that most of the money is actually not made on any of these distribution channels. It's mostly made in touring. So most of these artists actually are not that motivated by the revenue coming out of these channels and viewership sort of similarly. They just don't think that way. But there's one thing that they are obsessed with. Every artist you can think of is obsessed with where they stand on the Billboard charts and the revenue goes up or down. This guy was telling me, I never get a call from anybody, but if an artist drops from number two to number four on the charts, he's on the phone immediately, what are you doing to fix this? And said, oh, boy, that's kind of interesting. So we go home, go to bed, wake up, say, get together with the team, and said, so, what are we going to do with this negotiation? So we proposed this idea, said, what would it take for Billboard to actually use YouTube data as a signal in its rankings? And so we asked, does anybody call Billboard? Nope. Nobody thought to call Billboard. So we said, all right, well, maybe we should call Billboard. Maybe they'll negotiate a deal with us and let us do it. So we called Billboard. And not surprising in retrospect, although surprising to us at the time, they were super excited about it. And they said, of course we'd like to do it. We like to report on whatever's popular. YouTube seems to be where people are listening to music. We'd be happy to do it. And they said, what do you want to charge us for your data? And we said, are you kidding? We thought we'd have to pay you to give you the data. And they said, no, no, no. Okay, fine. So we decided to do it for free and we're going to send them all the data. So we start this project to send them all the data. Actually, their biggest question was, how the hell are we going to deal with all that data? So they wanted us to synthesize it in a whole bunch of different ways to make it manageable for them. And so we start this project and then we get a call from them just a few weeks before launch. And they said, hey, we'd like to launch a little bit early. And this is like four weeks before the launch. And we said, my history of all these partnerships, I'd never heard of a partner wanting to launch early. I said, okay, why? And they said, well, we really can't tell you, but we'd like to launch tomorrow. And we said, all right, well, you have the data, it's up to you. Go ahead and launch tomorrow. And we found out the next day why. And the reason was that this guy Bauer ended up at number one on the charts. And this is particularly interesting because at the time, Billboard had been around for, I think 57 years and they'd never had an unknown unsigned artist debut at number one. So he had no label and he'd never been on the chart before and he started at, at number one. And for us it was particularly interesting because if you actually went and looked at his views on YouTube, none of them were on anything he uploaded. This is the types of stuff you would see, I guess too fuzzy here to see it because people were taking his song and making all these kind of videos. That was really interesting to us. The long term impacts of that were enormous. It's still the case that, that YouTube has now become a primary vehicle for musicians and artists now invest a lot of energy into their music viewership, often at the cost of other things, partially because of the decision we made to share our data with Billboard. I think the lesson here is every business operates in an ecosystem. Doesn't matter how big you are. There's. There's sets of people that you need to rely on. And it's very easy to presume something about their incentives if you don't ask. So what I'd suggest is don't presume, go out Collect the data, ask the real hard questions, and you might find something completely counterintuitive like we did here. All right, let's talk about values. Switch to my. All right, so we're going to talk a bit about value. So as you scale, one of the things that'll matter a lot is your sense of values. The story I want to tell here is from September 11, 2012. So September 11, 2012, we get a call saying, hey, there's been something terrible happened. There was an attack on the US Embassy in Libya and the US Ambassador, Olivia Chris Stevens, had been killed. And the reason we got the call was the world was convinced that it had to do with this video, that this was a series of protesters that were protesting a video called Innocence Muslims. And everyone wanted to know what we were going to do about it. Every government, every journalist, everybody wanted to know what we were going to do about it. And so. And there was sort of compelling arguments on both sides. There was a series of people who, you know, really, really wanted to take the video down. It's just stupid video. Just take it down if it can help save even one person's. One person's life. And there was, you know, a lot of growing, you know, every hour we were getting called saying, hey, there's been another protest in a different country. And they got the same thing happening now in Yemen, happening in Egypt, and it was kind of spreading through the region. And so there was a set of arguments that way. On the other hand, our policy team could find nothing wrong with the video. They couldn't find any reason for us to pull it down. And so they were sort of staunchly arguing that we should leave it alone. And the next 24 hours are really tough. And I got probably, I think, one of the best short term education possible on how to think about the laws across the world for freedom of speech, hate speech and incitement of violence, which is really the intersection of those laws, is what really had to be figured out in this case. Now, the net impact is the video actually stayed up except for a few select countries. But thankfully it was also determined that the video actually didn't have anything to do with the. With the assassination of the ambassador. Now Hillary Clinton gets to go deal with the rest of that. I got to, at the time, yeah. So let's see. But the reason I think this is an important lesson for this group is these days, as these properties and businesses get bigger and bigger, we're sort of in an interesting spot where the level of impact these companies and products can have is kind of unprecedented. I mean, if you think about that situation we were in, we were the last decider, like, whether this thing stayed up or down. There was no government that could force us to do anything. There was no group to go to, There was no judge to make a real call on it. We were going to decide and we had to be ready to do it. And I think one of the things to think about as you scale, hopefully if you get to the point of working in a business that has that level of impact, it also comes with an enormous amount of responsibility. And you have to think about your values in advance so that by the time you get hit by this, you know you have some frame in which to make those decisions. All right, now switch gears. Talking about people. Go ahead. In terms of, like, how values apply to scaling. I mean, how do you deal, though, with, like, situations where, for example, going into China, such a huge market opportunity, but they're very tough about how they want to pull it off, or other countries that have very different understandings of freedom of speech. I mean, how can you leave such huge businesses, such a huge part of business on the table? Yeah. So the question was roughly how could you leave China on the table? So my perspective on that was, well, there's two parts. One, it wasn't really a choice. There wasn't really for YouTube in particular, and I think someone else will probably speak to the Google piece of that. But for YouTube in particular, there was no choice. There was no real opportunity to go into China. I tried many different tactics of ways to do it, but there just wasn't a practical way to do it. In my opinion, I would have done it. And in fact, we did launch in a number of. Actually through experiences like that, we launched YouTube in all sorts of countries where we didn't really agree with quite their way of doing business or the form of government, because, you know, the trade. And the way I would recommend it was I believe that allowing people to express themselves, even small amounts, was better than not being in the region at all. But for China in particular, it ended up being kind of a moot question. All right, let's switch to talking about people. So pick the right table to sit at. You figured out the right values, you figured out the right thesis you want to invest in. Now you need a crew, you need a set of people that are actually going to help you get this done. And so I picked off a few things to talk about in terms of how to judge talent, how to work with them yourself, and how to organize them. So let's talk first off on judging talent. And so I'll walk through this framework in a second just to introduce it some context. This is a framework that was given to me by an old boss for how to think about comparing people. Generally, framed in the notion of how do you tell the difference between someone who's senior and someone who's junior? We applied this at Google across a bunch of different functions. It's being used for how the product management promotion committees, for example, compare different levels of product managers. And generally the question there was phrased as, what's the difference between a level three and a level six product manager, or level three and level six engineer? And those terms, you don't have to worry about the numbers. You can reframe that as what does it mean to be a junior product manager versus a senior product manager? So on. So we thought about it in these two axes. So one axis that would generally get referred to is scope. And we'd say, okay, the way you should compare people is junior. People would start off and they'd work on features, and then they'd gradually work on groups of features, and eventually they work on areas of a product. And then they'd cover a whole product and then they'd cover multiple different products. And that's how you should judge junior versus senior is the scope of the thing that they've got ownership for. We proposed a slightly different axis for looking at this with an acronym that does not make a word at all. It's pshe. So it stands for Problem Solution, How Execution. So I'll walk it from the bottom up. So you start off as a junior employee. Someone hands you a problem, someone hands you the rough solution and says, this is how we're going to solve the problem. They give you a list of how, a list of instructions. Well, what you should do should go meet this person, you should assemble this group, you should write this document, and these are the things you have to do in order to implement the solution. And the job of the person is to go execute, go follow these instructions, and that's the job. And then gradually that person becomes a little more senior and gets handed a problem and gets handed a solution. But now that person has to figure out the how, figure out who to talk to, how to set up the team, what the cadence should be, whatever that might mean. And then at the next stage, the person becomes a little more senior, gets handed a problem with no solution. The job of the person is to go figure out the solution. And at some point in that person's career, they're handed a space, and they have to come back and tell you what the problems are and decide where to focus on. So this is a progression. We use to think about how to evaluate people a lot. You could put different labels on this axis. People like to call it leadership. I like to use the term training wheels. In fact, the way I like to phrase this is if you were to put the two axes together, is to say, given this person, what's the biggest scope you could hand this person with no training wheels, what amount of support do they need to get started? Now, there's a couple counterintuitive things about this. So one was the path that people generally took in evolving through these stages generally wasn't linear. In fact, what would generally happen is early in people's career, they would be sort of growing by growing scope, and at some point, the curve would turn almost completely the other direction, and it would be less about the scope and more about some of the mechanisms, how they did the job. And then gradually it would turn back at the top and this period in the middle, I called this the trough of disillusionment. And to put it in terms of like a promotion committee. So you're sitting there, you're trying to decide this level three person, this level six person, how do you tell the difference? And the way I would phrase it is the difference between a level three person and level six person had nothing to do with scope. They could be doing the same job. It was much more about how they did the job than the job they did, which makes it much, much harder to evaluate how good a job they're doing. But this frame of thinking about how people grow was very helpful in trying to figure out, and we use this across disciplines, trying to figure out what does success really look like and how do people grow through it. A second tool I want to talk about is something we use called dream Teams. And this tool, the basic idea here is really simple. We would take groups of people and we'd ask a single question. The question was, if you were, who would you hire first if you were starting this team from scratch? We used to call it the Dream Team question. We used to ask in a whole bunch of different ways. We'd ask managers for their teams, we'd ask peers for their people on their collective teams and so on, and try to form a view on what the dream teams were. Then we'd take that list, and we would generally draw three lines. So the first line is, I call it the awesome Bar. So those Are the set of people you sort of build a team around. At the other end of the spectrum was the we probably wouldn't have hired this person again bar. And then there was a middle line that I called the hiring bar. And just to explain each one. So the top one. Well, actually, let's start on the other end. They wouldn't hire again bar. First off, there was no set place to draw that bar. It was entirely qualitative. You could draw that bar and say, there's nobody underneath that. You could draw the bar and put the whole team underneath that. But the exercise was to be intellectually honest with how you feel about the. About the team. And there's a set of steps we would take when people were under that bar. I'll come back in just a second. At the other end of the spectrum was the awesome bar. And that's really where you want to invest your time. You take those set of people, you spend your time talking about them. You spend your time making sure they feel like they're valued, make sure that they are working on challenging things. We would rotate them through teams regularly to try and make sure that they were getting exposed to new areas of product, getting challenged in different ways. The one in the middle, the hiring bar was sort of interesting too. It's very natural in an organization to set your hiring bar at the same place as the wouldn't hire again bar. We need somebody to replace Joe. So just get me somebody that's better than Joe. And it's a very natural organizational instinct to say. Or the other way gets phrased is, well, we hired that person, and this other person seems just about as good, so we should just do the same thing. And so we would set the hiring bar and the YouTube hiring committees actually had a list of people. And so it wasn't a list of. Just a list of quality. So it wasn't. They can write code this fast or they can design at the speed or whatever synthetic metrics you might come up with. We would say, hey, we've decided our hiring bar is way up here. And we'd like. As you evaluate candidates, we'd like you to think about these four or five frames of reference of people we think are emblematic of of those capabilities. There's lots of tactics there to talk about if people are interested in. But the main lessons I wanted to express is you can pick whichever tools you like. If you like those two tools, then go ahead. You can pick different ones. But thinking about the ways you're going to judge people. For example, this Way of judging has perspective. There's many leaders who would pick a very different set of things that they would order. This is my perspective on how I think about how people grow. And it was reflected in the team. And there were leaders that had a different set. Like, you could almost express it as my set of values. And then thinking about how you want to manage that across your team and make sure that you're spending your energy on the top of your team, not on your bottom is a really hard thing to do, but a very important thing to do. Go ahead. What was your question? Yeah, I just wanted to make sure I was on the same. Who would you hire first? The three lines thing, Are they separate exercises? No, no, same exercise. So you build your list. You say, who would I hire first? And you sort of end up with an ordered list of people, and then you draw lines within that list. And this is always, like, at the midpoint of the end of the project, presumably. So we would do it every quarter, and so we do it on a time cadence. But. Sorry, I should be repeating the question. So the question was, what's the cadence for doing it? We would do it every quarter, and I actually think that doing it more frequently is actually better. If you have the right tools for doing it, you can do it pretty quickly. Go ahead. Question there. Do you fire anyone based on it? The question was, do you fire anyone based on it? So I was in the luxurious position of being at a company that people fought like crazy to get into. So generally speaking, people that were below that bottom line, once they were told that they were below that line, they generally made the right choices. So this is not a rule I would apply everywhere I've been in situations where that's not true. But that was sort of the notion of how it worked for us. But in some cases, absolutely, we did. I think that's the. I think building that muscle into how you think about your top and bottom is a part of making sure. The only way to make sure you have a healthy team is to treat your top very well and make sure that your bottom doesn't stay long for very long. All right, let's talk about your roles. Now you're leading this team. There's lots of different things to talk about as your role changes. You become a manager, you become a manager manager. At some point, you're running a bunch of functions, many functions where you don't actually understand what they do. Lots of different things to talk about there. I picked one to talk about, which is how to be a Bad manager. So. And I'll come back to why that implies something about being a good manager. I think there's three ways to be a bad manager. So one is the one that's most typically talked about is being a micromanager. That's a sort of easy example to understand is somebody comes to you with a problem, and you take the problem from them and you solve it. And the downsides of that, I think, are well understood. I think another one is. Another pattern is what I call dictators. And you tend to see this a lot more in sales organizations. So the story I would tell is the sales leader is sitting down with their sales rep. Sales rep comes in and says, all right, I had a $10 million quota this quarter, but I just wanted to let you know I had two major customers in that quota, and one of them, the company went bankrupt, and the other one, the head of the company, passed away. I just wanted to let you know. And the sales manager sits there and says, you know, I kind of blanked out for a second, and I may not have heard everything, but are you still bringing me $10 million this quarter or not? And that's sort of the spirit of those types of leaders. There's a third one. I call these the over empathizers. And this person is slightly different than either of those. So that person, employee, brings a problem, and manager sits, listen very carefully, doesn't take the problem from them, but spends all their energy trying to debug the problem with them. Says, oh, tell me more about it. Maybe it's because of this, maybe it's because of that, and so on. That may seem like a really good thing to do, but one of the implications of that is now you've taken the responsibility off of that person and onto yourself. The main lesson of this is I think everybody has their natural tendency, and you probably all know where your natural tendency is. Mine, by the way, is at the top. And any team I work with, I make sure they know that that's where I end up in if pushed. And I think being a good manager is actually not about not being any one of these things. It's about being able to be these things when you need to be. And sometimes the right thing to do is to be a dictator. Sometimes the right thing to do is be a micromanager. Sometimes the right thing to do is empathize. But understanding which one you naturally are and stretching yourself in the other ones is really important. And one way to think about it is the goal of the manager set inspiring Goals, coach when needed, and hold people accountable. And those can often be at odds with each other. All right, one last one. I'm going to talk a little bit about how to organize teams. And again, lots of different things we could talk about there. How do you organize functions? How do you organize projects? Do you move people around? Do you not? So on this one principle that I found, we used all the time, and it's told with a story. So the story is this guy walks up to these three workers and asks them what they're working on. And the first one says, hey, I'm responsible for taking bricks. I take bricks from over there and I put them over there. And then he goes to the second guy and says, what are you working on? And he says, oh, I'm responsible for laying bricks. So I take those bricks and I put the mortar and I stick them on top of each other, and that's what I do. And. And he goes to the third guy and he says, what are you responsible for? And he says, well, I'm here to build a cathedral. And this difference, I think, is really critical as you think about how you set up projects and teams. And hopefully the lesson is clear. People that are working on a cathedral, sort of similar to the purpose stuff we talked about at the beginning, the people working on a cathedral have a higher sense of purpose, better set of perspective on what they're doing, and so on. So when we first talked about this at YouTube, we talked about it on, off site, and someone asked for an example. So to give you the example, we asked one of the managers in the team to stand up. And this manager stands up. And I said, so explain to everybody else what you're responsible for. That person said, oh, I run the related videos team. I said, okay, what's the related videos team? Well, our job is, you know, when you come to a watch a YouTube video, there's all those videos, like, on the right side of the watch page. Our job is to pick which videos go there. And I asked everybody around the room, what did that sound like? Did that sound like laying bricks? Did that sound like building a cathedral? And that led to a conversation where that team's goal was reframed and broadened. And we changed how we thought about that experience. But the main lesson I would give is, as you think about organizing teams and setting up goals, what. It's one of the hardest things to do. But the question to be asked is, when you look at each person and look at their team and you kind of work your way through the Organization is how connected do people feel to a cathedral? And do they describe it as a cathedral or do they describe it as laying bricks? So those are 10, hopefully interesting stories, some about how to pick things that have high tailwinds. Pick things where you have a high sense of purpose as you execute. Look for, look for those theses. Look for that thing that you believe that not everybody else believes. Try to find that unifying metric. Try to make decisions that stick. Look hard at your ecosystem to find non counterintuitive patterns. Think about your values before they're tested. And as you're building up your team, think about the ways you want to think about talent and measure them, what type of manager you are, and how to structure your team around cathedrals, not just projects.

So can you say a little bit more about what that you said mentioned? There were two main periods of YouTube. So there was the period of, there was a fire lit under you and everyone thought you were going to fail. And then there was the sort of period of wandering that sort of followed after that. Yeah, describe what things were like at the beginning of the period you were going to fail. Do you have. How big was the team, what was the area of focus, and how close were you still to the founding essence of what YouTube was?

So just to be clear, I think I would probably describe YouTube as. I'd probably say there's probably four phases. So there was the founding phase that was mainly before YouTube was purchased and it's before I joined the team. So I know a lot of it only secondhand. After the purchase, the dynamics changed and then there was what I call the wandering period. And then I think the phase that YouTube is currently in, which I would call acceleration. And I think, you know, actually I think that seems to be going incredibly well. So I can talk about sort of the last three. The first one's a little bit different, you know, and YouTube is a very unique case. I don't know that a lot of businesses go through this kind of dip, but YouTube got, you know, on one side was an enormous success. You know, bought. Bought for $1.6 billion at the time. Seemed like a lot of money before a whole bunch of other acquisitions got done and, you know, was a very popular property and so on. But we were, you know, our backs were to the wall. I mean, it was a very common conversation of could this be shut down, could it be spun out? It wasn't a, it wasn't a foregone conclusion that you two would be allowed to continue going through this, but just I'll tell one story to help people understand what that. What that felt like. So we were. Let's see, this was. Must have been like October 18th or 19th. It was three or four days after my second daughter was born. I go into work, which my wife was not happy about, and meet with Patrick Bichette, who's our CFO at the time. This is sort of my first meeting with him. You know, I joined YouTube about a month earlier and he had three charts in front of him and he was. He was totally ready for this meeting. And one was, here's how. Here's how much money YouTube is losing per year. It was a lot of money. Second one was, here's how much money YouTube is losing per view. And that was not. That was enough money. And then there was a third one that said, here's what views are doing. And views weren't doing this. They were doing this right. And he said, this is the worst business on the planet. This is. There's been a lot of talk recently about negative gross margin businesses and so on. We were like the worst. So I would say on one side, it was incredibly focusing. Everybody completely understood what needed to be done and anything. Which meant across all functions, it meant obviously the monetization team was under a lot of pressure to find some way to create an ad unit and so on, some of the stuff I talked about. But it also meant things like the networking team was shaving costs as fast as possible, and the content team was constantly renegotiating and rethinking, thinking how we dealt with content licenses. And the viewership team was thinking about creative ways to rebuild the experience around this goal. So it was very clarifying. So I would say at that time, the sentiment would be intense. It would be intense, but you'd make resolutions fast and you'd go do them. I'd say that wandering period was tough because all of a sudden it was kind of like we went from this totally nimble small company to feeling very big all of a sudden. How do you pick between doing Project A or Project B? Well, they're all going to move us forward. They all seem good. So everything gets funded. Everything kind of moves, which kind of mean nothing moves. It meant all the projects kind of intersect each other. All of a sudden people are working on things that. That don't have alignment. One of the ways I like to describe it is in a healthy dynamic, somebody brings up a new project and says, hey, I want to do. You're working on A, I'm working on B. I Want to do. Here's project C in healthy Dynamic, Both A and B. First off, they're too busy to even do it. So the healthy dynamic is we're already racing. Jonathan Rosenberg used to the analogy of like, I'm going to drop this ball in the middle of the table. In a healthy dynamic, the ball should drop on the floor. I should have to force people to go do the new thing. And then when somebody picks it up, everybody is super supportive because that's the nature of that business is going. And I would say temporarily that changed. And I think we managed to change it back, but it wasn't easy. Got it.

So in that period at the beginning when you were making the big push to get the cost down, get the revenue up and so forth and turn YouTube into an effective business, what were key org changes you had to make in order to make that successful?

I mean, I would say the organization actually didn't matter a ton. I think the people mattered a lot. And so I would say. And we were pretty active on that. I mean, the tolerance for a not stellar person in a role was incredibly low then. And it was, you know, we move people out of positions incredibly fast just because we couldn't. We didn't have time to wait to see if something was going to work out. It was clear that person wasn't working. We moved fast. And on the other side, we pulled in the best help from anywhere we could. And sometimes that was sometimes. Actually one interesting story is the guy who. So I told the story about TrueView, this ad format I was mentioning, it took a long time to actually get it shipped. Took almost 3 years to get it shipped. So a lot of actually what we did to write the ship was done without it. That project actually got shelved. I couldn't get it through the sales team. And so we didn't launch it for a while. And there was a guy on the sales team who really, really wanted to be a product manager. And he was really smart. And so we decided to give him a shot, but we couldn't move him over. And so I said, here's the deal. I'll let you do it in your 20% time as long as you keep doing your current job. And you have to take this project that's gotten shelved. And I really want to see it work. And so I think we got really creative about how to get the right people on the right problems as you're sort of motivated the right direction.

Got it. And what you mentioned earlier when you were talking about how valuable having that single target metric was for the entire organization. What's the process you put in place for company communications? I mean, one of the things which really changes as a company sort of grows is the way in which you talk to the staff as a whole. So that very clarifying meeting you had with the cfo, how did you communicate that effectively to the team and how was that different than in the future when you came up with that one metric?

So I'm a big believer in there's a lot of different communication forums and we used to do a lot of the ones I think these days are pretty standard across companies. We do the TGIFs and you do Q&As and so on. I was a pretty big believer in writing a weekly mail and actually as a general sentiment, I think I didn't talk about here, but I'd say leaders that write things down. I think there's a lot of reasons to write or not write things down, but I think leaders that write things down tend to deal with less communication issues. And I would say to say it differently, writing it down generally meant you had to clarify your thought process in a completely different way. It's very easy to leave a meeting with, okay, we decided this thing. Everyone go tell your teams and they'll tell their teams and they'll tell their teams. And first off, you don't actually. Everybody doesn't quite know what was decided. And so if you didn't write it down and then secondly, you're going to play this game of telephone of I wasn't there, but the person who was there said that this thing happened and so on. So I used to Send a weekly YouTube all mail that would go to everyone. Robert would do something very similar and we would lay out. Some of it was updates. You know, here's things that happened. And generally we try to pick something like, we made this hard decision, we want people to understand why. And I think that was pretty effective at communicating things like that.

Yeah. So let's actually go ahead, please.

So you mentioned earlier in the talk that Sal Khan couldn't have been able to create his business three or four years earlier. How do you think about timing in terms of product and business space? Yeah.

So just repeating it so Sal Khan wouldn't have been able to make his business happen three or four years earlier. How do you think about timing when. When it comes to the space?

Yeah. Timing was actually one of the stories I ended up cutting. I think timing is critically important and unfortunately, it's really hard to. It's really Hard to predict. It's kind of related to the tailwind point I talked about. I mean, I think YouTube was far from the first at trying to make online video work. I've had many arguments with Mark Cuban about whether he was the first or not.

I bet I know what he thinks.

Yeah. So timing is kind of everything. And unfortunately, you don't actually know if you're at the right time until afterwards. I don't know that there's a prescription there, but certainly I think once you feel it, maybe that's related to the insight of what do you believe that other people don't believe? Sometimes that is entirely about timing. That you believe the right set of things are coming together to make it happen. I mean, why was YouTube timing at the right time? There's a lot of things that happened at that moment. Networks were finally big enough to be able to stream video. We were finally putting a phone, sorry, a camera in everyone's pocket. So the amount of creation that was happening was completely different. And then there was a bunch of things in the funding and investment environment that made it easier to imagine taking something that was actually a pretty capital intensive bet that actually made that work. And a whole bunch of things that had to come together at the right moment for that to happen. And a lot of luck comes into it, too. I think timing really matters.

Please.

You mentioned at the beginning of your talk a transition from broadband to cable to Internet video, which has totally shaped the industry. How do you think virtual reality will change?

So how do you think the virtual reality is going to play its role as the next step of the transformation of entertainment?

How do I think that virtual reality would change? See, John's in the back, and John and I have talked about this before. I think virtual reality right now is a fantastic promise. I think the timing of it is unclear. I think it's. Anybody who's put on an Oculus and gone through one of those demos, you can't help but being wowed. It is really amazing to see it. And yet I think one of my lessons from YouTube was the shift I talked about with going from broadband cable to Internet video. You can go back to the late 90s and you'll find the same. You can go back earlier and you'll find very similar ideas. But the set of things that have to come together for virtual AI to get there are sort of similar in terms of the network effects of. Of devices and content and applications and all of that coming together at the same time. So I don't have a guess on when that'll be? I don't think it's quite yet, but I do think it'll be great when it happens.

One more question.

You talked about wanting to start a company a couple of times. What made you decide the last time or times you dropped it to do it? Did you have the idea team, all those things?

You've had the opportunity a couple of times to start a company and you've finally gone and done it. So what's changed in the decision making process from when you decided not to do it in the past?

That's a good question. So essentially when I was at YouTube and people would come to me and say, I want to start a company, I used to ask them two questions generally as a way to discourage them actually. And my two questions were, do you have an idea you can't imagine not working on? And you have a person or a team you can't imagine not working with. And my view, the way I would phrase the question is if you don't have both, then you shouldn't do it. And generally that was enough to dissuade a lot of people. As it turned out, most people had one or not the other, or sometimes had neither, which I think was particularly interesting. I think the. By the way, I don't have a. I'm not so sure those questions have led to the right startups or not, but that's like as a life choice, that's how I think about it. For me, the simple answer is I found myself answering both those questions as yes. And once you're at that point, I think anyone who started a company will kind of describe it the same way. It's a feeling of like, damn, I'm trapped. Like I have to do it. You can't imagine not doing it. And that's when you have to do it. And that was the feeling I had. And so I did it.

So we're going to break now because we're right out of time.

I can stick around if people want to.

That'd be awesome. I was going to ask, so thank you very much for joining us.