Usually when we talk to CEOs as one of our Conversations from the Corner Office, it's actually from their corner office or at least the conference room next door maybe. Today, though, something a little different. Evan Williams and Biz Stone, the Twitter guys, are here in Aspen this week for the Ideas Festival. They did a couple of panel discussions on innovation. Small things, like, I don't know, 140 character messages turning into big ideas.
Evan Williams says that nine months after Twitter launched, it was still mainly used by web types, until the South by Southwest conference in 2008.
Kai Ryssdal: Evan Williams, Biz Stone, good to have you guys with us.
Biz Stone: Thank you, Kai. It's great to be here.
Evan Williams: Thanks, Kai.
Ryssdal: So Twitter was, in the beginning anyway, this thing that was leftover after this other project you worked on, this other company you started, and this is kind of the remnants. Did you know when you figured out that it was the remnants what exactly you had, Evan?
Williams: I don't think we can claim to have seen the future when we first started Twitter. I mean, Jack Dorsey whose idea it was and who worked on it at the time had been thinking about it for a long time, but we thought it was a neat little idea and was worth playing with so we went for it.
Stone: Yeah. Jack and I worked on it together for two weeks to get the prototype built.
Ryssdal: Two weeks? That's it?
Stone: That's how long it took. And I would say for about nine months after that, if you were looking at it like an investor, you would've considered it a failure, but neither Jack nor I saw it that way. We were having a ton of fun working on it.
Ryssdal: So you guys were just smashing buttons, Biz, for a couple of weeks?
Stone: Yeah. Well, it was Ev's idea. He wanted to shake things up a little bit at this other start-up that we were doing together and he said, "Everybody just pair up and try something totally different for two weeks. You have two weeks." So everyone paired up, and Jack and I had become friends so we paired up and we took the two weeks to do Twitter.
Ryssdal: And then Ev, was it "We're gonna do this social networking thing with it" or did you know what you were gonna do with it?
Williams: Well, both Biz and I had spent a lot of years working on blogging and the sort of end systems that help people express themselves and put thoughts out there, so we were kind of familiar with the area and thought... We didn't think of it as a social network. We thought of it as a way to express yourself and find out what's happening, but it was very similar to stuff we had built before in a lighter, more fun form.
Ryssdal: Biz, this thing you mentioned a moment ago about if you were looking at it as an investor, what was the VC pitch, the pitch you guys eventually made when you had a company and you realized you had to go out and do something?
Stone: Well we were in a different situation by that point, because like I said, for the first several months not much was going on, but then March of 2007 something happened that sort of changed everything. It was the first time that Twitter was released into the wild, so to speak, and the metaphor that I like to use is that of a flock of birds moving around an object in flight. It looks so choreographed. It looks beautiful, it looks complicated, but it's not.
The mechanics of it are very simple, and we saw it's basically rudimentary communication among individuals in real time allowing them to move as one organism, and that's what we saw people doing at this festival we went to, and that's when we realized we had something big. So it started to grow and that's a big difference when you go to investors.
Ryssdal: This was SXSW that he's talking about.
Stone: That's right.
Ryssdal: The first time that you guys saw this flock of birds that he talks about.
Ryssdal: Describe it for me.
Williams: So we had - Twitter was still pretty small and this was about nine months after it first launched publicly, but this was in the web development in a sort of Internet geek community that we were a part of, where we had a critical mass of people who used it. And at this event they all converged, and Twitter became the way to communicate at this event.
Ryssdal: And what were they saying? SXSW, big music conference, a lot of people there -- what was going on on Twitter?
Williams: Yeah. They also have an interactive or web part of SXSW, and so this is where it was happening. It's a very social conference, so there's panels and there's content and people would talk about what panel they were at, but they'd also talk about what parties they were going to, and that was a critical piece of information. You could see we put some screens up in the hallways that showed who was there and what people were saying, and it just energized this event, which was a great event to begin with. It brought this whole new layer of information and richness to it.
Ryssdal: And so you get out of SXSW in 2008, Biz, and then what do you do? You come home and you say, "You know what? We got something."
Stone: Yeah. We were certainly renewed and excited about what we were working on, and it was starting to grow faster because we got a lot of attention out of that festival. So we started to grow pretty good in 2008. Other things started happening around the world that again were just affirmations of our work. There was an incident in Egypt where a student from UC Berkeley traveled to Egypt to photograph activism and he kept missing the protests, and he asked his friends in Egypt, "How do you know where to protest?" and they said, "We're using Twitter. Get on it." So he gets on it and he makes the next protest, but he gets arrested by Egyptian police. He gets thrown in the back of a car panicking.
They didn't take away his mobile, so he managed to tap out a one-word tweet, which was "arrested", and when that information got back to his friends in Berkeley, they took it seriously, went to the dean, the dean went to the lawyer, lawyer went to the consulate. A few hours later another one word tweet, "freed," and that got back to us and we actually invited this student James Buck into our office and just chatted with him. Things like that started happening in 2008, 2009. Every time our eyes opened wider, and we just realized we had to really double down.
Ryssdal: And then last year the big news was the Iranian protest and that was all over Twitter. It was real time news.
Williams: Yeah. That was a very exciting event, and we had trouble knowing exactly what was going on, but we were getting reports. We were observing this just like everybody else from inside Twitter HQ, and the interesting thing that happened is we had some maintenance planned, which we had to do with our hosting provider, and we had been putting it off already. We announced on the site that we were gonna do this maintenance, and we started getting lots of calls and e-mails from people saying that time window, of course, we chose the time window that would be off time for us in the U.S., and it happened to be during a very critical moment with the election, so we were asked to put it off. There was some consternation about that on our team 'cause we really needed to do it, but we managed to put it off and work with our hosting provider to put it off, and that felt like the right thing to do to keep that information flowing.
Ryssdal: All right, so let me ask you this though. For all the real tangible news and information and good that comes out of this system that you guys created, what do you do when 99 percent of what's on there is, "My cat just coughed up a hairball," right? I mean, what do you do?
Williams: Sure. Well the answer to that, and this goes back to I was working on this type of stuff for a long time, is it doesn't really matter. There's two parts of that: One, "My cat coughed up a hairball" was important to someone, the person who wrote it, and maybe they'll occasionally write things that are important to more people, but the real answer is if you listen to 100 percent of the phone calls that went on, you'd find out that people talk about what seems like pretty trivial stuff amongst themselves, but no one says the telephone isn't a profound and important communication technology.
Stone: The key thing, I think, is that people use Twitter every day for whatever reasons -- social, fun, organizing parties, whatever it is -- so they're just trained to use it. So when something does occur, something significant, the first thing you do is you reach for your phone and you make a tweet. That means that Twitter gets involved in geopolitical events all around the world, it gets involved in disasters, in emergencies. Sort of everywhere something's happening, Twitter's people and Twitter are there kind of reporting on it from the scene.
Williams: Just to add another component to that, unlike the telephones, because the vast majority of the tweets are public, you can derive really interesting meaning out of these what individually look like trivial mutterings. You can see what people are talking about this movie and what the buzz is about this or this product or general sentiment about almost any topic that's going on in the world by just aggregating the tweets.
Stone: That's another answer to your original question of what do you do? How do you get to this good information that's certainly in there? That falls on us to do, to create discovery mechanisms, filtering mechanisms, way of surfacing information that's relevant to you where you are when you need it.
Ryssdal: And it also helps not coincidentally to play into the business aspect of this thing, which is how you guys make money, because for all the fun and good you're having you're not in this for nothing, right, Biz?
Stone: That's right. Yeah. We certainly believe that the open exchange of information can have a positive global impact, and sometimes we're looked at as hippies, but we...
Ryssdal: Which is funny, 'cause you don't look like it.
Stone: We believe that's true, and we also believe that you can make a great business out of it and that you can have a wonderful time at work working on such a meaningful thing. We've just recently begun launching our monetization platform, promoted tweets and promoted trends, which allow businesses to surface their tweets in front of more people.
Ryssdal: Promoted... I'm sorry, go ahead.
Stone: Promoted tweets and promoted trends.
Ryssdal: So promoted tweets and promoted trends, is that another way of saying "search" on Twitter, right, Evan?
Williams: Yeah. It's part of search and that's where they surface, so promoted tweets is simply a tweet that would surface in search and we take it out of the chronological order and put it on top. And trends is a feature of the site where we surface what a lot of people are talking about, so something that's got a lot of buzz recently. What a company can do now is they can promote something that already people are talking about, but maybe it didn't hit the top 10, and we'll put it up there in that list.
The first one we did was for "Toy Story 3," obviously a big event. A lot of people are talking about the movie. Disney/Pixar wanted to make sure people really saw that it was up there, so they promoted "Toy Story 3" as a promoted trend. People would click on it and get search results and see what everybody was saying about this movie.
Stone: By the way, not a bad box office opening.
Ryssdal: I would say. You guys picked a good thing.
Stone: Or they did!
Ryssdal: Oh yes. It was Twitter. I mean that brings up another way this thing gets used, right? I mean not only is it news and not only is it social events and stuff like that, but you guys can play a hand in the fortune of movies, right? Somebody sees a movie, it's a dud, you pull out your phone, you tweet that this thing stinks, and then it eventually gets out there and you guys have killed Hollywood.
Williams: Well, we haven't killed Hollywood. This is the thing that people have dubbed the "Twitter effect," and all it really is doing is speeding up what's been happening forever, which is word of mouth. So what this does is a studio can no longer put out a film that's bad and hope to trick people into paying for movie tickets over the weekend. If it's bad, then after the first screening people will start talking about how it's bad. Now if it's good, the buzz will build and the momentum will build and way more people will go see it. So the net result here is that we're gonna positively reinforce good content.
Ryssdal: Okay. Let me ask you this question though, picking up on that thing you said about it's speeding up natural word of mouth and what actually happens: Isn't that the last thing we need today? Life is busy enough and fragmented enough that really we could all just do a step back.
Williams: Yeah. I agree with you, and we built exactly that kind of sentiment into Twitter from the beginning. We actually have something called sleep mode. Twitter was designed to be this system that you just scan for information that's important or useful to you and then walk away, and if you wanna take a break you take a break. If you wanna jump in during the Aspen Ideas Festival and see what everyone is saying and listen to everybody, that's great. When you wanna walk away, you can. With e-mail, it just keeps building and building.
Ryssdal: But here's the thing though, and maybe this is just me, Ev, but when I step away from Twitter for a little while I feel like I've missed something, and I need to go back and scroll down and read all that stuff and it turns into this... There's three hours gone. Boom.
Williams: Well that's a good sign that there's compelling stuff on there.
Ryssdal: Or I'm just into cats with hairballs. That's another thing.
Williams: We do wanna address that. We really don't... We have an explicit goal to make Twitter an antidote to information overload, not a source of it, and it's not there yet, but that's one of our big long-term technology and product challenges is how to make it so you feel like Twitter is helpful to you, not missing the information you care about in the world. That should allow you to step away for a few hours and come back and find out what the most important stuff is.
Ryssdal: Sounds a little bit like the mantra of a company you guys both worked for, Google, and don't be evil. That's kind of the way it rings.
Stone: More like "Be helpful."
Ryssdal: All right. Fair enough. You mentioned Twitter HQ a moment ago. Do you guys have like big plasma screens on the walls where you scroll through and all the workers in the cubes can see the tweets that are going by?
Stone: The only big plasma screens we have are in the ops group, where we can monitor the overall health of the network so that we can see if anything is going wrong. But we don't have screens with tweets on them.
Williams: One of the ops screens is only... It's a search for "Twitter is broken" or...
Ryssdal: The fail whale, right? That thing that comes up. Yeah.
William: Yeah, so specifically what they're looking for is signs of something going wrong.
Stone: So the screens aren't... They're not happy time screens.
Ryssdal: Oh, so they're the bad time screens. So what I did as a good Twitter user was when I discovered I was gonna be interviewing you guys I went out and I said, "Hey, my Twitter followers, I'm gonna be interviewing @Ev and @Biz and send me your notes and thoughts about what you think I should ask," and by far the overwhelming majority of them were scalability. How come this thing is not shaky, but there are times when you want it to be there and it's not?
Williams: Yeah. We particularly had some trouble lately with the World Cup and a confluence of events that have hit the limits again with the system, and we're working very hard to fix that. We had some issues early on after that first SXSW, and the first couple years building an infrastructure that could handle the load, and lately we've run into those limits again. Between then and now, we've built much more solid infrastructure handling orders of magnitude and more traffic, but we certainly don't have the stability and performance that we need and plan to have.
Stone: I'll add a little bit to that in that certainly the World Cup'ss going on, we're seeing record amounts of usage of the system, and on top of that our engineering team is also addressing, at the same time, technical debt and just improving the system while this is all going on. So they're kind of poking at it and fixing it in what is historically the biggest usage we've ever seen. So it's a bold move, but it'll pay off in the end.
Ryssdal: You have to hope, right? Because you can't monetize this stuff if the system is down and that ain't no way to run a business, right?
Stone and Williams: No.
Ryssdal: Obviously, did a lot of research getting ready to talk to you guys. One of the things I read was in the New York Times, they had that Sunday Styles thing, they were at the dinner with you guys or whatever it was. At the end of the piece one of you says to the other one, and I can't remember who it was, "Man, you write really boring tweets."
Stone: Yeah. That was Ev to me. Yeah. He's been busting my chops for years.
Ryssdal: Which brings up this question: Writing good tweets, and this sounds stupid, but writing good tweets is hard work.
Stone: It is.
Ryssdal: I mean you've got 140 characters and you can't really say a lot.
Stone: Yeah. If you had more time you would've written a shorter letter, right?
Stone: Yeah. It's an art form. Some people have really taken to it and other people not so much, but you can tell when a tweet has been expertly crafted.
Ryssdal: You can. It is actually a work of art. It's true. Evan Williams and Biz Stone from Twitter. Thanks a lot, guys.
Stone: Thank you so much for having us.
Williams: Thanks, Kai.