Twitter Ex-CEO Dick Costolo: The Operator's Case Against Founder Mode

I just felt like Twitter was such a drama. Twitter was like the drama queen of hyper-growth companies.

You know, it was like when I took over as CEO that year, the Twitter bird had hanged itself and exploded on two different magazine covers. So, I went up to our head of communications, Gabriel Stricker, and I was like, Gabriel, great news. Next year's goal, no dead birds on magazine covers. You know, low bar.

Thanks for having me. Appreciate it.

Was it the one I was in?

Oh, thank you. Great. You're great.

Elon and don't forget, you know, Elon hired hand Linda Yaccarino.

What was the pitch to me?

I mean yeah pitched me. The pitch was the initial pitch was I'm taking paternity leave for a couple weeks. I have Williams and taking paternity leave. I need someone to come in and like just hold down the fort for a couple weeks while I'm, you know, out.

But then while we were having that discussion, he's like, "Well, what if it was permanent? You know, what if you were COO?" And I was going to go start another company. So, rewinding the clock a little bit further, I'd sold my previous company, FeedBurner, to Google. It's where I rebumped into the Twitter guys because they were all there from Blogger and getting ready to leave and start Odeo, which became Twitter. Anyway, I'm leaving now. I'm leaving Google a couple years later. Ev texts me that. I was like, I'm going to go start another company again with one of my FeedBurner co-founders. And then Ev said, what if it was permanent? What if you were COO? And Twitter was already so it's summer 2009 it was already like had hit the, you know, Ashton Kutcher and Oprah Winfrey are, you know, fighting for who's going to get a million followers. I can see why you wanted to do it but zero revenue model, zero business model, all the articles were sure Twitter's interesting but it'll never they don't have a business model, never make money. And I just kind of had the, man, you don't get an opportunity to go into these to get on these rocket ships and when you get the opportunity you should do it. So I did it.

It's always both, you know. Yeah. It was the things that were different were it was very much a collective by which I mean like decisions were kind of made as a group. It wasn't okay well you're responsible for engineering I don't know make it figure it out and you know figure out what the right answer is or should we, you know, have should we make retweet a native thing or should it be, you know, just you write you had to write RT at the time and like rewrite the tweet. I remember that. And there was a discussion about well should we just have retweet be a thing you can do in Twitter and it would be kind of sit around the table and talk about it and decide the group would decide. Okay. You know so things moved methodically is a diplomatic way of putting it and decisions were reached over a period of time. It was chaotic and then it was also just it just didn't move very quickly. It didn't move quickly at all.

It was awful. I don't really have a side of it other than it was an awful like it's, you know, these things in hindsight are it happened very very quickly it happened super slowly and painfully and, you know, I was informed like in the driving rain on the road home by, you know, Evan, Jason that the board has decided that you should be the CEO. Yeah. But I had been told and then Bill Campbell called me and like you should just quit, you know, and then like why am I quitting, you know, and like anyway went and then like the board decided actually no Ev is going to remain CEO but you've been great and we're going to invest a bunch of your options and so long. Okay. And that I mean, no kidding. Over many, many days and weeks. And finally, I was like, listen, just write up the separation agreement and, you know, I'm good. I'm good. Like, whatever. I'll go find something. I'll go do something else. And then, you know, to make a long story short, the board got back together. I was like, oh, no, no, no, no. Remember Bijan Sabet from Spark texted me, don't quit. Okay. I was like, I'm not quitting. Like, okay. You guys told me you wanted me to leave cuz I was going to stay. Anyway, it was super crazy. And there was obviously a ton of stuff going on and anyway, long story short, they were finally like, "Look, you got to do it." Anyway, and, you know, really long story short, I was like, all right, I'll stay and do it.

Well, Jack was the chairman of the board still. Ev had made Jack chairman of the board when he made himself CEO. Ev's still on the board. So two of the founders, Ev and Jack are still on the board and Biz is in the company and Jason Goldman's in the company running product and, you know, and a bunch of other people. Yeah. The first thing that I felt and I sort of foreshadowed this. The first thing that I just felt we had to do was like velocity, velocity, velocity. And that remained a concern of mine throughout my time there, like faster, faster, faster, cadence, cadence, cadence. Like how can we go faster? Why do things take so long? If this takes six weeks, what would have to be true for it to be two weeks? Yeah. We just I did stuff like that constantly because I was always like, man, this is like dental surgery. But it was starting from a position of we're not going to sit around the table and vote on things anymore. I remember vividly the status line changed from like what are you what's going on or what's happening to something else or maybe it changed to what's happening. I don't even remember. I wasn't even involved in it. And one of the front-end engineers, Brit Seatville, came into my office. He's like, "Hey man, what's going on?" I was like, "What are you talking about?" He's like, "The status bar changed to what's happening." I was like, "Okay." He's like, "I didn't no one asked me, you know, what I thought about that." I was like, "Oh, man. Yeah, we're not doing that anymore." Like it's going to start being there's going to start being a lot more of that. So, that was thing number one.

Thing number two was there were a set of operating principles that were written on a whiteboard that Biz had been involved in with either on his own or with other people but they were like, you know, like sometimes you'll get on a United Airlines flight or whatever and they'll say here at United Airlines we believe in integrity and honesty. I'm like, "Okay, who cares?" Like, that doesn't help me. That tells me nothing about every like whatever. And also, it doesn't help me make any decisions and these things are meaningless and they're usually weaponized against management by like, you know, well, it says right here anyway. So, it's like, okay, we got to get rid of these things. They're like not useful.

Uh, no. They didn't really. And so, good for them. We got rid of those. I like there were just I remember there were just like everything was just kind of like messy, you know? I remember there was this kitchen and it was just kind of a mess. There were like dirty dishes in the sink and everything and I think I sent out an email. I was like this is why the site crashes. Like you may not get it, but like this is why the site crashes because everything's just kind of like oh well someone else will clean it up. And I was like this is ridiculous. And it was just I just tried to get this sort of like we're not going to do this like oh well we'll get them next year stuff anymore. There was a poster on one of the floors that was like we'll make better mistakes next time or something like that you know and I was like get that poster out of here. It's not I don't really care about it but it's just sort of indicative of oh well you know and then the site would crash again and people would go like oh well we'll figure it out you know we'll hopefully won't happen again next year whatever. So it was a lot of cleaning that up and a lot of velocity velocity velocity.

Yeah. Because you it's like you get bigger and get less done.

Yeah. You know, so a couple things initially when we were smaller, this is sub 200 people. What I did to speed things up was one, we're not going to make decisions as a group. We're going to push decisions down the stack. By the way, I don't have to be like, don't come to me and ask me what you should do about this unless you have no idea and want my opinion. Like, figure it out, you know, and don't tell your team what to do. Tell them to figure it out and push as many decisions down the stack as you can. And I don't want to hear, you know, I remember I would talk to Jack Dorsey at these dinners we would have from week to week when he was chairman and was CEO of Square down the street from us. We would get together on Tuesday nights and have dinner. And I remember telling him, "God, I hate hearing Dick said in the company as a reason for things." Because what would happen was I'll give you an example really well. The Android app started taking longer and longer and longer to boot. And so we would have a metrics meeting. I was like, "Hey man, the you can see the usage numbers going down almost linearly with the Android boot time going up. People are firing up Twitter and it's taking too long and they're moving on to something else." Like over here on iOS, that doesn't happen. It's got to be the same boot. It's got to take just as long to boot Android as it does iOS. Okay. So, two, three nights later, week later, I'm walking around like one of the engineering floors after hours and this Android engineer comes up to me and goes, "Hey, can I ask you a question?" I'm like, "Yeah." He goes, "Hey, did you say remove the JSON file from the Android boot system?" I was like, "I didn't I know what a JSON file is, but I couldn't tell you what's in the Android boot JSON." And he's like, "Okay, because I was told Dick said remove that." And I was like, "Aha, I think I know what's going on here." Right? And this would happen throughout the company. Some manager director decides here's a way to solve this problem and the engineer pushes back against them and they just go, you know what, Dick said, do it. And it became like this reason for doing things. So part of the speed was no, we're not going to have someone say that. It's push the decisions down the stack. And by the way, if your team makes a decision and you can't defend it to your manager, like that's don't go, well, so and so said like have the engineer talk to the director, whatever.

So very common. So here's how it surfaced. We ran an experiment and it was like, oh, this is actually really super interesting. What are the hypotheticals about what this experiment would do in production? The engineer runs through it in an experimentation meeting. Like, great. Let's go ahead and launch this to, you know, x% of users and see what happens. Yeah. One week, two week, three weeks go by. I'm like, hey, where's I'm like looking at the app. I'm like, where's this engagement experiment we're running? And I find the engineer. He's like, "Oh, I'm waiting for, you know, Trust and Safety to sign off on." I'm like, "What?" He's like, "Oh, yeah. Well, I asked so and so if we could launch this. It was ready to launch." And they didn't say no, but they said, "You have to go get approval from your security." And I started realizing that when people would try to go do something, they wouldn't be told, "No, you can't do that." It was something much much worse and slower than that.

You got to get security. You got to go ask these 14 people. And if they all say yes, okay, everyone's got this problem. How do you fix it?

So what I did was it's very similar to what Bezos used to do at Amazon with biased action. I had it from my improv background. I was like, we're going to have a bias to yes. And biased to yes is only your direct the person you report to is allowed to tell you you're not allowed to do that or legal can tell you if you do that you'll be fired because it's illegal and a violation of the privacy policy. But no other organization period is allowed to tell you you can't do that or you need to go ask these people for permission. Just your direct manager. And that means if you have an idea for an experiment and your manager signs off on it and you have time to do it because you're done with whatever else you're working on, no one can tell you you can't do that or the marketing team's not allowed to tell you you can't use the Mailchimp software we license because that's the marketing team's, right? And so bias to yes was I don't want to hear about something being stalled because organization X said you have to go get approval from organization Y. And that was great. Experiments started flying out. In fact, engineers started going like, wait, I don't have to necessarily get design to build all these wireframes for something. I can just do it myself since we're only launching it to 1% and see if it works. And so that was just great. It worked throughout the organization and I would repeatedly get on stage at all-hands meetings and talk about like bias to yes and what it means and what it doesn't mean. It doesn't mean you can go stop doing what you're supposed to do because it's a priority and do whatever you want to do because your manager is going to go like if you do that I'm going to fire you because you're being measured on.

Yeah. Don't solve problems with processes. So, it happens all the time in organizations like some customer problem happens, the customer churns out or the advertiser quits. Yeah. And the CEO goes that can never happen again. Yeah. You know, and so what happens? Some step gets added to some process that's like before you do I remember at Google famously the launch checklist got to be like 17 pages long or something and Larry was like this is insane I was like yeah it's cuz you know it's like the FDA never another thalidomide or the TSA like make sure there's no water bottle on the plane and pretty soon you're not managing to outcomes you're managing to the processes and you got to go through nine steps before you, you know, and instead of just managing to the outcome of make sure there's not a bomb on the plane. So the way we did that, the way you do that is you solve problems with DRIs, operating control and not processes. You should probably only really have a couple organizational processes and everything else should be span of control and decision-making.

Yeah. You want them to take risks. You want them to be bold. You want them to have this bias to action and doing things and not going around and tiptoeing around asking for permission. Yeah. But one of the implications is you can't punish people for making mistakes. And I got to the point where I would tell the team it's not the job of leadership to prevent mistakes from happening. It's the job of leadership to correct mistakes quickly when they happen. So, if you screw something up, like, by all means, tell us so we can help you fix it. But no one should get in trouble for screwing something up or making a mistake. And I didn't come to that immediately, but I came to realize that that was an important artifact, if you will, of pushing decisions down the stack and bias to yes. And we're not going to have lots of processes.

Yeah. Well, I did two things. First of all, I taught this management class at Twitter. It was actually based on Ben Horowitz teaching a management class at LoudCloud.

And Ben had in his outline from his management class a statement that I like highlighted and underlined and circled three times. Yeah. Was make sure everybody understands what you understand. So I would always tell my managers that like your job is to make sure everybody understands what the priorities are and what it means for them and their team, how they fit into those priorities and more importantly for your team, what it means for them personally and professionally if we achieve what we have to achieve. Those kinds of things are great. I remember somebody telling me a story of taking a factory tour at a Tesla shop and seeing some guy who's like screwing a widget into the car and asking the guy like, "What are you doing?" And the guy didn't just say screwing the widget into the car. He said, "Well, if I do this and we're able to build these kinds of cars and these kinds of cars can become more popular in America." I mean, the guy understood the story and I thought that was great.

Yeah. Like you want people to understand how much more motivating is it to work in a company where you know if I'm successful in my work? I understand what the implications are for both me personally and the company.

Priorities and context. Good news travels fast, bad news travels slow. And it's slow as the org gets bigger, it slows down and the signal gets distorted.

Yeah. I said earlier, it's not the job of leadership to prevent mistakes from happening. It's the job of leadership to help correct them quickly when they happen. So, I would always communicate, you got to tell leadership if something's screwed up. If your situation is screwed up or your team or some decision is a mess, you have to communicate that constantly. And then secondly, I would regularly communicate the priorities to the team over and over and over at all-hands meetings. Here's what the priorities are that we're working on. This is how we're going to measure them. This is why they're important and this is what they mean for us as a company if we're successful. I always wanted people to understand not just what we had to do but why it was important to do it. So, and I think that scaling communication architecture is like one of the most important things in any company. And part of it is never saying because so and so said that's not context. That doesn't help your team understand context for why some decision was made or what they should do or how they should go do it. So that combination of pushing decisions down the stack, making sure that everybody understands what you understand and that a lot of that is context and always communicating problems or issues up the stack as a means of helping get it resolved quickly were the ways we dealt with that. One last thing that I picked up from a Steve Jobs trick.

Yeah. So when Steve was managing both Pixar and Apple. So he's CEO of Pixar still, but now he's back in the seat at Apple. So now he's only at Pixar one day a week because there's all this stuff I got to go fix at Apple. So they're making like Toy Story 2. And what Steve would do, so he's got one day to come up to speed on what the heck's going on at Pixar. So he'd go into the office on Friday and he would say, "All right, today I want to sit down with the illustrators, you know, without Lasseter, the director of the movie, John, without Lasseter in the room. I just want to sit down, me and the illustrators." And he would go, there's like eight or nine or whatever of them, he'd say, "Janet, tell me something that's not working well at Pixar." Yeah. And then, "Bob, do you agree with that?" Taking notes like, "Okay, Phyllis, tell me something that's working well at Pixar. Not on the movie, not about your job." And at the end of the meeting, 45 minute, one hour meeting, he'd say, "Okay, here's what I heard. People are having these kinds of issues. This is going really well. I'm going to go talk to John about it, and he'll deal with the rest of this on Monday. You can follow up with him." So, he's not saying, "Oh, yeah, that sucks." Or, "No, you're wrong or whatever." He's taking notes, making sure he hears everything that was said, and then going to talk to Lasseter about it. I started doing that with teams. You quickly find out if that team knows what the priorities are is in sync with what you're hearing from the leadership of that team. And some of the leaders of the company hated it. And I was like, well, now I'm concerned about why you hate it because Yeah. But you're not in that meeting doing things like throwing the director under the bus. You're going, "Okay, here's what I heard. I'm going to go talk to my management team about it on Monday, and Chris or whoever will follow up with you guys."

Yeah, you can rejigger them once in a while. Once in a while, I don't know, like you know if you've got four huge priorities for the first four to five months of the year you can't get to February and go these two are stupid and we're changing once or twice a year you can go this one is let's actually change that one to this other thing. If you start doing it regularly people are like well we can get to work on the direct messaging architecture but Dick's probably going to 50/50 he's going to change you just can't do stuff like that by the way that's another way companies slow is because people are just like, "Look, I'm not going to bust my ass on this thing when there's a 50% chance Dick's going to throw it out the window in a month."

No. Focus is focus I don't think focus is ever overrated. However, the fact that you can do more quickly and since they're two-way doors, you're like great. No, I mean, this is the old Jeff Bezos only take a long time to make decisions that are existential. Like, we might die as a company if we make the wrong decision. There's fewer and fewer of those decisions that can be like if we screwed it up we can go back and do the other one. Do those right away. So focus is still important but it's also true that you should feel like you can do a lot more now with less because you can. There was a point in 2011 when we wanted to try to get Jeff Bezos to join the board of the company like we have an independent seat open who would be the best possible independent director Jeff Bezos. So Jack and I go meet with Jeff Bezos. I think it was like at the Allen & Company conference or something like that. And Jeff asks a question about strategy and Steve Jobs was still alive at the time. And Jack said, "Well, Steve says the most important part about being a CEO is focus." And you should say no to almost everything. And it's in fact it's the CEO's job to be the editor-in-chief. And you should say no to almost everything and do very very few things. And Bezos looks at Jack and goes, "Oh, yeah. Well, I like to do everything." And he does the big Bezos laugh and he goes, "My team has to talk me out of stuff." He goes, "You know, there's lots of ways to be successful." And I remember that like it was yesterday. And he's right. There's many ways to climb the mountain. Yeah. I think it's true that today he would say, "Great, I can do even more now. We can still be focused, but I can do even more things." I think those things can both be true.

Okay. And mostly late night. It started because Bill Campbell for people who don't know or listening to the podcast, Bill was on the board of Apple and famously coached Larry, Sergey, and Eric at Google. I wrote a book about him called Trillion Dollar Coach. The Benchmark folks had also brought him into Twitter before I got there. And Bill famously said when I took over as CEO like you could roll a grenade into Twitter at 5:30 p.m. and only take out the cleaning people. So, and he was right. So, I was like, "Okay, how do you you can't tell people you just can't tell people work harder, you know?" They're like, "Okay, you work harder." You know? So, what I did was I just I mean, it's such a stupid thing, but I was like, "All right, I'm going to go out get dinner. I'm gonna come back at like 9:00 PM, 9:30 p.m. I'm gonna walk around and whoever's here, I'm going to talk about the stuff they're doing at the next all-hands meeting and I'm gonna bump up the prioritization of that." Yep. And I remember like Marcel Molina at no radio on Twitter. That's a little promotion for Marcel there. Marcel was always working late on stuff and I would like what are you working on and I'm working on native retweet or whatever and I would get up in front of the company and go like what Marcel is doing and people were like how does he always know what Marcel is doing and other people would go like oh Dick comes back after dinner at night and walks around and people slowly slowly slowly started to change the culture of okay the people who are around when Dick's walking around are getting their stuff prioritized.

Yeah. And you Yeah. So I would just go, you know, up to engineers and go like, "What's going on? What are you working on?" Yeah.

Yeah. I was trying to help a director in the company be more strategic. They were super tactical.

And I was finally like your job as a director as you grow in this company is you can't just say like well I put out this forest fire and I put like your job is to map the territory and start to do forestry management. Like what are we going to do so that next year if there's a fire over here that highway is not going to be in trouble. Like you can only work so many hours a day and stamping out fires isn't going to work as we scale. You need to push those down the stack. Your team should be like fighting the fire that's over there because it might burn a house down or something. You need to be focused more on like thinking about this whole pie and what's going to be true, what has to be true 24 months from now so this whole territory is in better shape. And I just that was the metaphor I had for helping people try to think strategically versus like well I'm chopping down the trees look at all the trees but that's not like leadership and strategy that's tactics.

Oh boy I can hardly wait. I'm sure most people did but anyway.

Yeah. It's just like I'm super impatient. I know this is a personal flaw, but I'm super impatient. Like people would come in and go, "Here's why I'm sad." You know, like, and they would start talking to me like, "Okay, stop." Like, "It seems like you don't like working here. You should quit." Okay. Like, "No, I want to tell you I'm sad." You know, like, "Okay, I don't care. Who cares? I'm sorry. I'm not trying to be a jerk, but like we have a million things to do." I remember. And then you know I just like this is I just direct and blunt with people. I'm not a jerk. You don't have to be a jerk. People think like I think a lot of times especially maybe engineering leaders who come up in the organization like coding and programming and doing stuff on my own and now I have to manage people. They think when you tell them you have to like look you have to manage performance in your team and you have to manage poor performers out. They think you have to be a jerk when you do that. You don't have to be a jerk about it. You just have to be direct about it and forthright and not, you know, no one's sadder about this than me. You know, don't say stupid stuff like that. You can be empathetic and still be straightforward.

Yep.

Yeah. My advice for that CEO, Bill Campbell, actually gave me this advice early on. He was like, write down what you want to say and then say that and don't say anything else. So the night before if you write down the honest feedback that you know I would give Brian especially if I have to give Brian a bad performance review the night before if you write it down I guarantee you 99 out of 100 people if they go into the meeting the next day they don't say that. Okay. They're like listen Brian instead they're like you know it's really hard time here for everybody. We're all working, you know, just shut the up. Like the person knows you're about to hit them over the head or

Something. And you also don't have to say like, "Hey, Brian, that's just the way it is." You know, too bad if you don't like it. You just write down what you want to say and be forthright and honest and you know, you can say like, "I understand. It's like if you don't want to deal with it anymore, I get it." And say those things and don't add a bunch of fluff around it. I thought that was great feedback and I would do that.

Yeah, sit with the silence.

Yeah. It's unnatural. The best person. So, I used to do a thing in my management class where it was a six-hour course. There was a ton of stuff in it, but one of the things I would do, it was about giving people feedback and like, okay, like Brian, you're going to give Dick feedback and the feedback is you've been working on this thing for six weeks and busting your ass and you've been here till midnight every night and we have to cancel the project because it's no longer a priority.

Yep. You know, this is made up. Of course, the only person who did the role play that was ever good at sitting in silence was Elad Gil, current venture capitalist. Elad sat in the chair and gave the person the feedback and I say to the employee, the IC before the role play, I kind of pull them aside and go, you're this is [__] and you're going to tell Dick and you don't want to hear this and it shouldn't be cancelled. They need to look at the data, whatever. And Elad was great at just sitting there. It's hard.

And Elad was great at that and the only person who ever did it.

It's a great question. I actually tried to promote from within as much as possible because you want people inside the company to feel like we're not going to get capped out at which point I can only get to this level because anything above that Dick just goes out and finds someone. So Kevin Weil for example was promoted from within from an IC when I got there in 2009. Well there were only about 40 people in the company when I got there. To eventually running all of product after running all of ads product. So he was promoted from within. Anthony was... I just felt like Twitter was such a drama. Twitter was like the drama queen of hypergrowth companies.

You know it was like constantly in the news. When I took over as CEO that year, the Twitter bird had hanged itself and exploded on two different magazine covers. So, I went up to our head of communications, Gabriel Stricker, and I was like, Gabriel, great news. Next year's goal, no dead birds on magazine covers, you know, low bar. Yes. Anyway, but it was just constant drama. And Anthony was so good at Goldman Sachs where he ran technology and media at dealing with investors and communicating with them etc that I had been trying for a while to bring him into Twitter and I was finally able to and he was of course he's fantastic.