Let’s talk about Slack
Stewart Butterfield has done it again. At just 41, the Canadian seems to have stumbled upon his second and latest accidental success. After co-founding Flickr in 2004, which was sold to Yahoo in 2005 and changed photo-sharing forever, he is now running Slack Technologies, a company he co-founded in 2013 and which is aiming to revolutionize how people are communicating at work.
Slack and Flickr both began as side projects at companies where Butterfield and his colleagues were developing online games. While he proved unsuccessful with games (one never even launched, the other one had to shut down ultimately), Butterfield was more fortunate with the projects he started while working on them. After Flickr, which was bought by Yahoo for over $20 million, he is now in charge of Slack Technologies, selling his communication app to customers like Adobe, Paypal, Buzzfeed, Airbnb and The New York Times. Slack is growing at a tremendous pace. 45,000 teams with 365,000 daily active users are paying to use the service. The business software maker is not even a year old, but the company is already worth more than $1 billion.
While at DLD in Munich this year, I talked to Slack’s CEO about those numbers, about developers, a European office and the possibility of an IPO.
Some have written that Slack is killing email. Do you agree?
I think that there is a good chance that we will see a massive reduction in email use inside of organizations because of Slack and tools like Slack over the next five years. The advantages of using one centralized tool for internal communication are so big that most companies over the next ten years and maybe even faster than that will switch from primarily using email to primarily using one system.
When we ask our customers today, what they were using before Slack, most of them will say ‘nothing’. But of course they were using something. You have to press them on individual questions: If you were going to be late to a meeting and you wanted to tell someone and have them apologize to everyone else, how would you do it? Then they’ll say ‘Oh, well, SMS’. Or Hangouts; or Skype Chat.
There are methods. But it’s very heterogeneous across the organization. Some groups will use one method, some groups will use another. In contrast: When a whole team or company is using Slack, all of the information goes into one place. All the messages between people but also all of the messages from computers to people.
In our case pumping into our Slack instance every day are tens of thousands of messages. Literally. Every time someone signs up for Slack, every time we file a bug, every time someone checks in new code, every time someone tweets at us, every time a customer creates a new ticket in Zendesk, every time there is a database alert or some other monitoring situation… We have about 30 different services that are all integrated with Slack – in our instance. And there are close to a hundred across the board. Different companies will use different kinds of services.
Those advantages are so big that it will be difficult for people to ignore them. On the other hand email is the lowest common denominator form of communication. That isn’t meant to be insulting to email. But it’s the thing that will cross organizational boundaries. It’s how we organized this interview and it is how people like us will be coordinating this kind of meeting five years from now.
Slack wasn’t the first collaboration program, and it won’t be the last. Still it has been growing tremendously over the past year. Why do you think it has been such a success?
To be honest: I don’t know. My PR team hates it when I say that.
But there are also many factors. I think we did a good job. We made a lot of good design decisions and things like that. It was also just the environment that we came into. It’s the right time for it.
It was only ten years ago when half of the people in the U.S. had Internet access at home. Even though it was mostly dial-up. People forget how long these things have been around. It’s much shorter than you think. Although the iPhone was first announced in 2007, it wasn’t until almost 2008 that most people had them and there was only a small population. There wasn’t an App Store or any of the features that you’ve come to expect.
Facebook taught hundreds of millions of additional people how to use the Internet for social interaction. For me this is very native. I got my first Unix account 22 years ago in 1992. I’ve been using the Internet since then. But for many people the Internet was something that, if they used it at all, was used for stock quotes, sports scores, news and maybe some light email usage. But now almost everyone in the developed world uses some kind of messaging app – whether it’s just SMS or iMessage or WhatsApp or Skype or anyone of the billion of other things. Facebook Messaging, Twitter DM…
The ability to use something like Slack is new. I don’t mean that people are too dumb to use it before. They didn’t have a context or why they would use a messaging application in the context of work, and suddenly they did. So it is really the right time.
If we had launched five years earlier, it wouldn’t have been this successful.
— René Seifert (@reneseifert) January 19, 2015
Facebook recently introduced „Facebook for Work“. Any chance to have a look at it yet?
No, because it’s not available for everyone. It’s for preapproved companies only. From the description it sounds like it’s just a private instance of Facebook. It has the same functionality as Facebook – but it’s shut off.
So you are not afraid of it?
No. It may or may not be a good product. On the product side I’d be skeptical about the basic side of Facebook functionality being something that people could really use to completely replace or at least to consolidate other forms of information. This is basically a necessity for Slack. You have probably seen this with the editorial side use of Slack.
It’s all or nothing. It’s binary. If it doesn’t take off, then people won’t use it. If it’s just another thing that you have to use in addition to what you normally use, the value just isn’t there. This was my experience of using Yammer. More products became additive.
In addition it’s a difficult brand to make a working product. The odds of us being worried about it are the same as Snapchat being worried about SAP coming up with a messaging app. It’s just the wrong thing.
Lots of companies are using Slack. Wouldn’t it be convenient to connect those networks together?
That’s something we think about and struggle with.
There’s a real benefit to the fact that when I go to my phone and I see the Slack icon on the home screen and I tap it, I know that it’s all the people that I work with – but it’s also only the people I work with.
The cognitive cost of separating and isolating all these different signals in a really mixed environment is very high. When you look at your emails, it’s friends, it’s family, it’s former colleagues, it’s people you work with today, it’s people you work with outside. You probably get a crapload of emails from PR people, you get the pitches, but you also get jokes from people that you went to high school with. It’s a crazy mix. And 80 per cent of those messages are not from humans but from computers. ‘Someone followed you on Twitter’, ‘These shoes are for sale’… That’s a big mess.
We don’t want to bring that big mess to Slack. That would decrease its value. But it’s something that we very carefully think about. For organizations that work very closely together some kind of cross connect between multiple or different Slack instances might make sense. But the idea that everyone has a publicly accessible Slack address like people have a publicly accessible Skype handle? I don’t think that will happen.
Who knows what will happen five or ten years from now. But we have no plans to do that. It would take away one of the big advantages that teams have using Slack today.
Slack is now worth more than $1 billion. Compared to Flickr – which you sold in 2005 for roughly $22 million – what do you think is worse: that you sold too early and too cheap, or that we are seeing these high valuations of startups all over again?
I don’t really know if either of them is worse. We definitely would have made a lot more money it we had waited even six month to sell Flickr. But people forget what the context was at that time.
Flickr was the first thing that was successful in the Internet since the crash. It was a long gap. There was the Dotcom crash; there was Worldcom and Enron and all the accounting scandals; then there was 9/11 – the terrorist attacks. Such a deep pessimism that lasted a very long time. There was no investor interest in Internet stuff generally, but especially not in consumer facing Internet products. So Flickr was at that time and in that context an enormous success and it was a very large amount of money.
The advice that we got at that time was: You sell now and it will be a life-changing event. You’ll have money to make the kind of choices that you’d like to have after and all that’s true. So I don’t regret it at all.
At the same time: valuations for technology companies are high now. I personally don’t think there is a bubble in the sense that it can pop and everything is gone. I do think that the other shoe has yet to drop from 2008. You can’t have negative interest rates everywhere. It just doesn’t make sense.
I’m 41. I was born in a stagflation era. I graduated from high school in a very bad recession – at least in Canada in the early 90s. I worked my way through the dotcom crash in this industry. I personally lost a lot of money in 2008. My father went bankrupt in 1982 because he was a real-estate developer and interest rates were over 18 per cent.
I’m very conscious of the fact that there is a cycle. If interest rates go up and money pulls out of the public markets, there’ll be less enthusiasm for private technology companies’ stock and prices fall across the board. Even in a bad scenario – not the one where a meteor hits earth and life is extinguished but the kind of a worst-case scenario for us – maybe our value would be halved or something like that. We are growing incredibly quickly and people are paying for this service. Revenue has grown as quickly as we’ve grown expenses. It’s a good business.
I don’t have any special knowledge. But I think the same is true in the case of many of the highly valued Internet companies like Airbnb and Uber. They actually are bringing in a significant amount of revenue. I don’t know if their exact valuations are justified. But I don’t think that it’s like tulips. It’s not fake. It’s not something that will just disappear. It might be hit hard. Just like energy companies got hit hard by falling oil prices over the last couple of months. But I don’t think it’s just going to disappear.
From the idea to the product: How long does it take you to implement a new feature?
It’s highly variable.
Last year we had this challenge that I made for the company, that if we got to 25.000 paying customers by May 1st then there would be some big reward. I wrote this long thing, trying to motivate everyone, saying that it’s not just about PR. It has to be efficient customer support. People have to feel that they are well being taken care of. The site has to be very quick and responsive. We shouldn’t ship any bugs and we can’t afford any downtime. We have to keep on releasing new features at a steady pace so people feel that we’re investing in the product. All these things.
Three of the four things I listed a year ago in that memo as our highest priorities and major new features are still not done. So sometimes it can take a very long time. Sometimes it can be very quick. It really depends on the complexity. And it’s not the surface complexity of the feature.
Some things we can implement entirely on the server side. But on some things every client has to be updated at the same time in a simultaneous way and thoroughly tested. That means for us right now: desktop, web, Android and iOS. Those ones can be significantly more complicated because each one has to be designed, developed and tested individually.
Let’s see if I can come up with an example…
Mentioning someone’s name is a special occurrence on Slack because you get an alert only when someone mentions your name. We want to make that easier for people. That might seem like an obvious thing.
Many people are used to using Facebook where when you start typing someone’s name it will be automatically completed. So there is the technical implementation, the UI design, how it feels… We don’t want to interrupt people. If there’s someone named Andrew and the English word “and” might cause the trigger there, we insert some special stop-words. The same thing is true in a bunch of different languages with a bunch of different words.
Then there is the fact that often people will put their name into the system as “Michael” for example, but they like to be called Mike. So we are adding a new nickname field. We have to think about what kind of feedback we give to people that they are doing it properly. On Facebook they put a blue background and a little border around it so you know that this is an official mention of someone’s name. That kind of stuff.
There is frequent feedback from engineers about the usability. There is a handful of people who will be involved in the development of any new feature. Myself, two of the co-founders, the head of costumer support – there is always at least one of those people who have been around. We are growing so quickly. There is only a handful of people who have knowledge of how everything fits together. We have six or seven designers now. So it’s a long process of back and forth.
Zendesk, Twitter, Dropbox: You have talked about third-party apps and services that work with Slack. What’s your strategy regarding developers?
I don’t think we have done a very good job there yet. Do you remember the Steve-Ballmer-video where he is screaming ‘developers, developers, developers’? I think we are actually almost exactly at the same position.
We have always made it very easy for people to export all their data from Slack. So there’s not a huge switching cost in that respect. However, there is a switching cost when people integrate their own internal systems – and also for every third-party integration.
It does two things for us. One: it increases the value of Slack for people’s daily use. So they like us better and are more inclined to pay and recommend us. And two: the more they build on it, the less likely they are to ever leave. Especially the things that people build internally.
I don’t think we have done a great job on evangelizing or supporting people’s use of the API yet. Just again: we have a lot going on and we have grown so quickly. But it’s something that in the future will be critically important to us.
Where do you see Slack in two years – in terms of valuation?
It makes a difference what’s going on in the macro environment. Assuming that everything is exactly the same in two years as it is now – which is super unlikely – but in exactly the same environment our valuation should be significantly higher.
If we are valuing this as a business in the classic sense – based on future cash flow or future profitability of the company – we have a very long way to go to justify our valuation. But we are being valued on the basis of the trajectory of our growth and if our trajectory remains the same the valuation should stay fairly constant.
We will double in revenue and in terms of the number of businesses using Slack and in terms of daily active users and everything else at least once this year – but probably twice and maybe a little bit more than twice. That will happen again the following year. So two years from now we should be eight times bigger, ten times bigger, twelve times bigger than we are now. Something like that.
I don’t think the valuation should be eight or ten or twelve times higher than it is today. But it should be twice or for times as high.
How many people are working for Slack today?
We just crossed a hundred. Which is crazy. A year ago we had twelve or fourteen.
Most of them are working in software development. Engineering, product and design – that’s almost the whole company. 15 to 20 people are working at the customer experience team, which is functional testing but mostly customer support. There are six people in accounts, which is essentially sales. Three people in marketing – all of those people are new and joined us in the last six weeks or so.
How important is Germany for Slack?
30 per cent of our daily active users overall are in Europe. I would have thought that the UK would be our biggest country after the US. But Japan is our biggest country after the US. We don’t have any Japanese language version. We have paying customers in Poland and in Portugal and in Italy. And it’s all English language. I mean, Dutch people: no problem. They all speak English. It’s easy for them. But I would have thought no Italian would ever use Slack. But they do!
We are thinking about how to make that better for people. There is the language of the software, which is the hardest part for us to deal with. It’s just very complicated and a big pain in the butt. Then there are some things that are easier – like offering support in local languages. Japanese, German, Spanish. Maybe Chinese. But I’m not sure what to do about China. I’m not sure if we are ever going to invest because we are not sure, if we will just get shut down at some point. So right now we are just ignoring China.
But offering local language support, local language documentation and local currency billing is really important and we are working on that. It’s doubly hard for Europeans. First of all because now the US-Dollar is going crazy against everything so it’s more expensive. But also we are set up in a very American-centric way where it’s monthly payments by credit card, which is the least convenient way for people in Japan and the least convenient way for people in Germany to pay. Not just because of the lack of credit cards generally or the decreased use, but especially because of the decreased use in a business context.
So yes. Germany is important to us. The rate at which we are growing right now means that we have our hands full just keeping up with what we are already doing. But we are talking about all those things and we are likely to open a European office this year. We had talked about eventually doing it. But it occurred to us last month that we have to do it in the next couple of months.
And where will it be?
Probably in Ireland. It’s cheap. And they speak English as well.
In 2005 you sold Flickr to Yahoo – only one year after officially launching it. Do you have similar plans with Slack? How about an IPO?
I used to say that I would never be the CEO of a public company. I like the part of this that I get to do, which is more on the product side; my background is in design. I really enjoy that. I enjoy business strategy as well. And I guess I’m starting to enjoy management in general, although it’s very slow. It’s like when a child drinks a beer for the first time and hates it but eventually comes to like it.
Being the CEO of a public company you have to deal with activists and shareholders. There is a huge amount of regulation and there are all those obligations and the disclosure. Not that disclosure itself is such a bad thing, but it’s a lot of work that’s not needed.
The dream scenario, for me and for a lot of the entrepreneurs I know, is something that is not realistic: you grow the company; it generates a lot of cash; you are able to go to a bank and say ‘look, we have predictable cash flow and we’d like to borrow $20 billion’; then you buyout your investors. Maybe that kind of thing will be possible someday, but I’m not sure if it ever will be.
Going public for Slack is a bigger unlikelihood than selling it. Again – never say never.
We did a good job but we also had the perfect timing. This was the time that this product should have launched. And there are a lot of things that we had no idea what the right answer was. Slack is a very, very strange name for a product. It was very risky to do that. It turned out well. But it could have turned out poorly.
A lot of decisions we made very early on, with very little information, went in our favor. That could have gone either way. When we started Slack, we thought it would be useful and that there’d be customers for it. We didn’t realize how strategically important this was.
In retrospect it makes sense. Communication is the most fundamental activity that an organization does. But communication is also probably the biggest problem at most organizations. If you ask most managers, executives and teams what their challenges are, the one that’s universal is communication.
Messaging itself is turning out to be the most fundamental application of the Internet in general. All of those things kind of converge. It’s impossible that I’ll ever have an opportunity this big again. So there is no incentive to sell.
Do you remember who your first customer was?
Yes. Not paid customer. I don’t remember who that was, but I remember the very first teams that we convinced to try it. We said ‘Please try this product!’ They were friends of ours.
The first big one was Rdio, a music service. It was an eye-opener for us. They had over a hundred people working at the company. We were eight people at the time. We had perfectly designed Slack for a team of eight people. Everything was optimized for exactly that number. Suddenly we realized how many things didn’t work with a hundred people.
Not only were we just eight people, but also eight experts in using Slack. We had designed it ourselves. There were all those things they found confusing. For example the number of notifications they got. It was incredibly helpful to us. Every team that we got at the beginning was a huge boost.
We started development in January 2013. Rdio started using it in May or June, just a few months after launching it. It was a very young product. We released it to the public in February 2014.
In the six months, from when we got our very first teams to when we launched publicly, a hundred per cent of my attention was on looking at what made teams successful or not. It’s a very big thing to ask people to change very fundamental behaviors and to use a different product. There is a different dynamic to convince a group of people to do something and to convince one person.
Every time a team adopts Slack, there will be one person or maybe a small group of people who first suggested it, but you really need everyone’s buy in. Especially early in the process, one person who doesn’t like the idea, can be a problem.
We looked closely at that dynamic. We wanted to make sure that the people, who were suggesting it in the first place, had the resources they needed; that the initial user experience was good enough and that people understood the value. I think we did a fantastic job on that. I would still give us a ‘D’ letter grade. We did a better job than most people do, but it’s so hard.
When someone has never used a system like this before and they are the first persons in and they haven’t invited anyone yet, there is literally no value in Slack. There is no one they can send a message to. There is no archive for them to browse. There is nothing for them to search. Nothing exists. Trying to convince them to take the next step and start inviting people is a challenge. It’s a miracle that anyone uses Slack at all.
Everyone knows about the first Tweet. Can you remember what the first message ever sent on Slack was?
I’m sure the very first messages were like ‘test’ or ‘asdf’ and stuff like that.
You started Flickr while working on a game. Slack was launched while working on another game. Any more game plans?
No. I’m done. That’s it.
In both cases it was not even so much a game. It was trying to use a framework of play to create a venue for social interaction.
My father plays Bridge. He wouldn’t invite those same people over to his house, just to hang out, but he also doesn’t like playing Bridge against the computer. So it’s not the people, and it’s not the game. It’s the combination. You get competitive, you’re trash-talking them – I wanted to be able to create that online. It’s very idealistic and I still think it’s a very nice and beautiful idea, but it’s forever doomed to fail. People don’t want that.
Ten years ago it was Flickr. Now Slack. What will it be in 2025?
I hope still Slack. I’ve got another few decades left of doing this kind of stuff, and I like this one the best. So I like to stick with it.
At Slack – are you using Slack only or email as well?
It’s just Slack.
Personally, I probably use more email than anyone else at Slack – except for the customer support team. But inside of Slack, we would never think of sending someone a message any other way. That’s not ideological. It’s convenient for both parties. They already have Slack open. They use Slack search all the time. There is a big value to having everything at one place and don’t have to look for it all over the place. No one would ever send Google Hangout messages, or Skype message, or SMS – even Twitter DM and Facebook Message would be rare.
Did you celebrate any special customers joining Slack? Do you have a list with companies you’d like to have on board?
There were so many, to be honest. Some of them were just cool customers that we thought would be influential and could help bring more people online. Airbnb was like that, and Stripe, the payments company. Hot startups.
For different people at the company there are different ones. There are two teams at NASA that use Slack. The New York Times has a huge team. That is cool. I didn’t care about it so much, but a lot of people were super excited to find out that the media company Marvel has two teams using Slack.
Any plans for a deeper integration of Slack into iOS? Like posting links from Safari directly to a specific channel?
It’s surprisingly difficult to do that. Apple made the sharing extension possible in iOS 8. But there is a lot you have to do yourself. But it’s coming. The iOS team is very under water. For the last five or six months they have been doing a complete rewrite to support offline mode and to make it faster. Same way with Android. So they’ve kind of hobbled on mobile. It’s almost done though.
A German version of this interview was published on t3n.