Adobe-full-color Adobe-white Adobe-black logo-white Adobe-full Adobe Behance arrow-down arrow-down 2 arrow-right arrow-right 2 Line Created with Sketch. close-tablet-03 close-tablet-05 comment dropdown-close dropdown-open facebook instagram linkedin rss search share twitter

Stewart Butterfield: How We Scaled the Fastest-Growing Business App Ever

About this talk

Slack is one of the fastest-growing B2B businesses of all time. So how can the company continue to innovate in the face of breakneck growth? In this sit-down interview with Fortune writer Erin Griffith, Slack founder Stewart Butterfield shares how he leads when his team (and bank account) are getting larger each week. His biggest struggle: finding the right people. “Every practice that we develop for how to manage becomes obsolete in 60 days,” he says. 

Stewart Butterfield, CEO & Co-Founder, Slack

Stewart Butterfield is the co-founder and CEO of Slack, the platform for team communication. Prior to Slack, Stewart co-founded and lead Flickr from its inception in late 2003 through its 2005 acquisition by Yahoo! and until 2008 by which point it was one of the largest web services in the world with over 50 million users and billions of photos.

In nearly two decades working on the web, Stewart has had a distinguished career as a designer, entrepreneur, and technologist. He has been named one of the “100 Most Influential People in the World” by Time Magazine, BusinessWeek‘s “Top 50 Leaders,” and been featured in interviews and articles by hundreds of publications and broadcasters, including the Wall Street Journal, the BBC, The New York Times, CNN, the Financial Times and has appeared on the cover of Newsweek magazine.

Butterfield graduated from the Universities of Victoria and Cambridge, with degrees in philosophy and retains academic interests in cognitive science, the history and philosophy of science, and economics.

Erin Griffith, Writer, Fortune

Erin Griffith reports on startups, venture capital and Fortune 500 tech companies for Fortune. Previously, she worked at PandoDaily, Adweek, peHUB and Mergermarket. Her work has appeared in The Atlantic, BBC,, Cosmopolitan and


Full Transcript

EG: I’m just going to jump right into it. We’re supposed to be talking about scaling. So the first question I want to ask you is– there are two ways to grow company. You can either just reinvest your profits and organically, brick-by-brick grow a company. Or you can take venture capital and use it to fuel crazy hyper growth. And that is, obviously, the path that Slack has chosen. How did you decide that was the right way right way to grow your company?

SB: I guess part would be laziness because– when we started the company, it was 2009, and was to do something totally different. Web based, massively multiplayer game. And it was easy to raise money, in contrast to any other previous venture I had because it was good times in the economy, and because we were.

EG: Wait, 2009?

SB: Yeah. Well, the times in the venture capital economy. The crazy little subset of a little subset.

And if you can get a million and a half bucks for very little effort and a relatively small percentage of the company, makes your life a lot easier. Starting off and having the initial focus be making money. So I guess everything else is an elaboration of that up to several hundred million dollars in the bank.

EG: There’s a big jump from that, in 2009, raising a million dollars, and what you did a couple weeks ago. Valued at over $2 million, how much money do you have the bank, you’ve hardly spent any of it. How did you make that decision? Oh well, I have a lot of responsibilities. But I’m responsible to our employees, to our customers, to shareholders, to my partners. And the analysis was are we better off to have 100% of what we have or in 94.3% of what we have, plus also $160 million. The latter seemed like the more prudent choice.

Look honestly, it’s not that we needed the money. It’s that it’s the cheapest time ever raise money. So if you never thought that you might want to raise money because bad shit might happen in the future, then this was a good time to do it.

And I’ll say this very quickly. But I was born I 1973. That was stagflation, which was–economists thought was impossible. My father was a real estate developer and went to bust in ’82 because interest rates went to 18%. I graduated from high school in ’91 into a bad recession. I worked through the crash. I lost a lot of money in 2008. So I’m very conscious of there being a cycle. And if money is very cheap then I’ll take it.

EG: And the business side of things, I mean why don’t you walk me through how that giant bank account that you now have is translating into how you’re growing the company.

SB: It actually doesn’t have that much of an impact. I guess in this environment

EG: It’s not burning a hole in your pocket? You’re not dying to spend it?

SB: No. And we’re– frugal is maybe overstating it, but we’re definitely not spendy. But this is one aspect of capital being super cheap in this crazy world is that San Francisco is mental. We leased a new office about a year ago for $65 a square foot, or $62 a square foot. And we had to move because we’re going really quickly, and nine months later we sublet it for $75. So from $62 to $75 in 9 months. And salaries are like that. And everything is in this hyper inflationary 20%, 40% a year kind of environment. So to have a company in San Francisco you kind of gotta do that.

EG: So what’s your biggest chal– So aside from real estate, what’s your biggest challenge right now in scaling Slack?

SB: The biggest challenge is always people. The technology is– especially compared to when we started Flickr, hardware is now essentially free, like in comparison to 2003 prices, it’s 98% cheaper. There’s a lot more and better open-source software that gets us a lot further. There’s a lot more experienced developers who have done this before, so it’s easier. So the technological scaling component is not trivial, but it’s close to trivial. But when at the beginning to 2014 we were 14 people. At the end we’re 90 people. Now it’s 130 people, and by the end of year it’s going to be 250 to 280.

And that’s crazy because everything, every practice that we develop for how to manage becomes obsolete in 60 days. So by the time we get a rhythm and feel like, OK, well now we’re doing the right thing, it’s screwed. And we have to start all over again.

EG: So, you’re in a really unique situation to be observing this because not only are you a company that’s experiencing hyper growth, but you’re also in the business of helping companies communicate with each other. So maybe you can give us some examples of what you’ve learned in studying the communication within your own company, but also the way that all the companies that use your software communicate.

SB: Sure, so how many people here use Slack?

EG: Oh my God. Yeah, that’s crazy.

SB: But this is exactly the sweet spot for the audience, I mean for our audience. Because you have a lot of people in tech. We have a huge number of people in creative industries and in media. But those of you use it probably notice that it’s a big change in your behavior. And we’re really conscious of this in the early stages, that we were asking a lot of our customers.

If we average over all of our daily active users– so including people who are just getting started or don’t use it that much– they spend a little over 10 hours in a day connected and two hours and 15 minutes in active use. So that’s big a chunk of your time. And it’s going to be the software that you spend the most time with over the course today, other than perhaps whatever you do as your job function.

So that might be Photoshop, or that might be Excel, or that might be email, or it might be an IDE, because you’re writing code.

EG: But the idea that it’s not going to be email. That you’re going to kill email.

SB: Well, naw, I mean, that’s– we’ll kill email inside of most organizations that use Slack over the long run, but we’ll never kill email in the general sense because it’s the lowest common denominator. It’s how you cross organizational boundaries. It’s how we set up this talk right now and how people like us will set up talks like this five year from now and 10 years from now.

But, sometimes a small group of people will take to Slack, and it’s just perfect. It’s exactly what they wanted. Other times there’s a lot of resistance. And sometimes the resistance is we did a bad job explaining what the functionality is, we had bugs, or some people just don’t like to change, or there were really significant workflows that were built up around the way people had communicated at that company or that organization over the course of decades.

But sometimes there’s just bad habits. And one of those is information hoarding, particularly by people in the middle management positions that kind of protective their position by controlling who knows what and when. And one of the benefits of Slack is that it increases the sense of transparency.

And I don’t mean transparency like in a political sense. Like it’s not Edward Snowden transparency. But it’s that people can see into other parts of the organization and see what their peers are doing. And that can be threatening to some people.

And it’s usually only threatening when there is a bad situation. So I feel like it causes positive change.

EG: And it’s interesting, though, because not only are you– this is not just a productivity tool. You’re actually taking a philosophy of how companies should operate and forcing them a little bit into that.

SB: That’s very interesting because we deliberately try not to force the idea of a message inside of an organization being the lowest common denominator format. So it can be a person sending a message to another person. And they can be really long, elaborate, formatted, bulleted point messages. Or it could just be like question mark or just a single emoji, or something like that.

But we don’t want to be prescriptive in how people use it, so that’s a side effect of its usage is transparency. But it’s not like we’re trying to foist that on organizations who don’t want it. And another place where that shows up is in time and information management.

So here’s a really quick example. We designed it, initially, so that you would only get a notification if someone mentioned your name of if someone sent you a direct message, because the assumption was that they wanted your attentions. So you should get notification. And otherwise, we won’t give you a notification, if there’s just people talking in a channel. And we won’t show you a number. Because a number’s just going to stress you out.

But people would install the app, and they’d sit next to each other, and someone would send a message to a channel, and the person would say, fuck, I didn’t get a notification. This app is crappy. It’s broken because what they expected was that ever message would cause a notification. And so we had to set up so that, at first, every message will cause a notification. And then once you get to, I think it’s like 100 messages and three people on the team, you get a little message from Slack Bot that says, hey, that’s a pretty stressful set of preferences you have right now. How about we change it so they only get notifications when someone mentions your name?

And almost everyone will do that. Similarly, when you install the mobile up and enable push notifications we say, hey, maybe you don’t want to get email notifications anymore. We’ve just turned that off for you. If you want to turn it back on, go here.

There’s a lot of that because there’s a lot– and we feel this. Because when we first started making it, it was eight people on the team. And we had designed, literally, the perfect product for teams of eight people exactly. If you were more than eight people, sorry. It was terrible.

And one of the very first external customers we had was Artio, the music service. And they had about 100 employees. And, God, we got so many complaints from them about, this is impossible to manage. And there’s too many channels, and there’s all these problems.

So it took us a little while to figure that out. But inside of our company because we’re growing so fast, is the same problem that I just referenced about management– When it’s eight people, everyone weighs in on every decision. It doesn’t matter whether it’s your domain or not, or if it’s their expertise. Because only five out of the eight people care about any particular decision, and it’s fine for everyone to speak up. Because the other seven people can read the messages generated by that person. But it gets exponentially harder to do that as you get bigger.

So we’ve got to 40 people, and now there’s like 36 people giving their opinion about the color a button, or something like that. It’s the equivalent of the email chain from hell. Yeah. And also every single message that someone sends has to be– or it doesn’t have to be– but often is read by many other people. So we got to the point where we had an internal memo, the title of which was, Slack is a Distracting Menace. But it just meant that we had to evolve how we used it, what the expectations were, and the social protocols were. And we definitely feel that– because we get about 15,000 tweets a month from people. 15,000. 15,000. And another 15,000 help tickets a month through Zendesk.

And if you work at Slack then every single person you know in the industry also uses Slack. So you go to dinner with some friends, and it’s like there’s a bug, there’s a feature request, bug, bug, feature request. So we get a huge amount of feedback, and a lot of it is about that. It’s like how to adjust the way we work, and how to adjust the way we communicate, to a system where anyone can send you a message at any time, (including at 2:00 AM).

EG: I love the idea that Slack Bot is kind of like the exact opposite of Clippy.

SB: Right. It should be. Trying to help you instead of stress you out.

EG: So I want to talk about design because I know that’s a big feature of this conference. And you have a design background. So I’d love to hear you talk about how the skills of a designer are applicable to someone like you on the business side.

SB: If any of you have ever heard my name before Flickr, but if you were in this industry in like in ’99 2000, 2001, then maybe you heard of the 5K contest, which is something that I started in reaction to a designer who I managed who said– we were making templates. This is in like 1998 or 1999– Templates for some client of ours. And I said they had to be under 10 kilobytes, for the naked template for the page. He said, that’s impossible. You could never do anything good.

So I made a web design competition that was 5K. And I didn’t really know this at the time, but it’s part of this really big tradition of highly constrained design competitions, and creative competition in general. So there was the 4K demo seen for graphics programmers. There is two color print design competitions for students in graphic design schools.

Often there’s like the 48 hour filmmaking competitions and that one month novel. And if you back up a little bit and think about creative practice in general, often it’s about the selection of constraints. I mean, in music you have tempo. You have a key that you’re playing in. Poetry you have the meter and you have a rhyme scheme, often. And people will arbitrarily select constraints in order to highlight some aspect of it. There’s a– God, I’m trying to remember his name now, but in Don Norman’s book, “The Design of Everyday Things,” he quotes someone else in a letter who has a definition design as, “Design is a successive application of constraints until only a unique solution remains.”

So part of the process of design is to identify the constraints and then apply them in the right order. That’s absolutely true in business, all the time.

EG: And how are you doing at Slack as the CEO?

SB: Well, its organizational design is a lot of it. So who do we need? How many people do we need? How do they work together? At what point does there need to be another level of management in the organization? Who is missing from this team that’s trying to get something done? And who do they absolutely need? Because there’s a lot of organizational drift is the natural state—

EG: Organizational drift?

SB: Drift, yeah. Like whatever is happening on Tuesday will be the thing that happens on Wednesday except some random change that someone made.

EG: So you’re talking about staying focused also.

SB: Yeah, staying focused and making sure that you’re continually optimizing. And there’s a bunch of constraints in business, obviously. So we talked about this backstage—

EG: Money. Obviously that’s not a constraint for you.

SB: Yes. But you should be making money, and people should feel that they’re doing work that’s interesting to them, and that they’re being recognized for it, and that they’re motivated to do a good job, and all of that kind of stuff. But there’s a very long term design challenge where you don’t get any feedback on whether the design of your organization is good or the design of your business model is good for like a year or five years or in some cases 10 years.

Because everything’s easy in retrospect, but you have so little information at the time to make a decision. So how do you keep yourself in check that way? Lots of smart people that I work with and really challenging one another. I get a lot of– I feel incredibly fortunate, in many ways, but incredibly fortunate that we got very good and supportive investors. And there’s a lot of actual wisdom there.

So people complain about VCs in the same way people complain about lawyers. And there’s lawyer jokes, there are VC jokes. Some of my best friends are lawyers. I don’t mean that as like a euphemism or anything like that. It’s literally true. One of my best friends is a law professor here at Fordham. And the same thing can be true of VCs. They can give really good advice. But also we have a lot of—

EG: You heard it here.

SB: Yeah. We have a lot of employees who have a lot of different backgrounds.

EG: So one employee that you’ve told me that you specifically are not hiring is growth hackers. Can you tell me about why you hate that position?

SB: It’s so easy to fall into the trap of like maximizing for some local position or juking the stats in the– Juking the stats. Yeah, in “The Wire” terminology. So a quick example of that is an email that Twitter would send that would give you the top five interesting tweets from your network today. And they would show you the tweet, but they didn’t link directly to the tweet. They linked just to the timeline of the person who tweeted it. And the inference, and I’m almost positive this is why, they did it is because you would go click on the timeline, and then you would try to find that tweet. And you would end up frustrated and angry at Twitter, but you’d spend a few more moments on the site. And so when they did A/B test, if some people got this email and some people got this email, they’re like oh, people got this email spent a little bit longer on the site, and that’s the result that I want.

So that when my performance review comes up in April, I can say, we increase this number by 17%. Yeah. That’s very tricky, Twitter.

EG: All right, I think I have time for one more question. I want to refer to something you just showed me which is an article that you wrote for “Design Magazine,” and one of the pieces of advice you included was related to the mission. Everybody needs to know why they’re doing something. So can you tell everyone a little bit more about what that means?

SB: Yeah. We two and a half– or not even quite—two and a quarter years ago, we started development on Slack. And we ended up making, again, huge decisions that were based on very little information, so they could’ve gone either way.

But the roadmap that we set out was is exactly what we’ve done. And the mission that we chose then, which was to make people’s working lives simpler, more pleasant, and more productive, is something that we can get behind. It’s not like we’re changing the world, or saving people’s lives, or anything like that, but it is nice—

EG: Are you allowed to say that in Silicon Valley? You have to say we’re changing the world. I thought that was required.

SB: We’re changing the world and making the world a better place through office communication software. [LAUGHTER] But there’s two parts that come with that. So one part is the internal side. One part is the external expression of that. So the number one quality of people we look for in hiring is empathy.

Because if you can’t empathize with people, you’re just not going to be able to do a good job in many things. Not every job requires empathy, but it’s a huge bonus, even if you’re a back end programmer. The outward expression that is courtesy. And courtesy, not in a fake, polite way, but in genuinely trying to anticipate what people need. And it turns out that that mission for the product is a great mission to turn back into the company.

So that if all of us are continually trying to make each other’s working lives simpler, more pleasant, and more productive, then you take a little bit more care in how you communicate something. If you’re going to ask someone a question, you give them a little bit more context. If you’re going to call a meeting, you have some result you want out of that meeting, and you communicate it in advance, so people don’t come and feel like they’re wasting their time. And that trickles down into everything. That people should feel like they’re being respected, and they’re being taken care of. It that manifests itself in the software, then people are going to love it. And they’re going to recommend it, and then the people to whom they recommend it will also love it. And they’ll also recommend it, which is the best case scenario.

But also it’ll be a good and healthy place to work. It sounds like a great place to work. All right, well, that wasn’t really insightful. We have like 14 seconds left.

EG: So everybody, please join me in thanking Stewart.

blog comments powered by Disqus

More talks like this

Visit the 99U Conference site