We Don’t Sell Saddles Here
Patrick: [00:02:05] My guest today is Stewart Butterfield, founder, and CEO of Slack. Stewart's 2014 essay, Why We Don't Sell Saddles Here, had a massive impact on my own business journey, which made this discussion extra special. During our conversation, we discuss the concept of owner's delusion, how to frame the boundaries between product and market, and the challenge of changing people's mental models and behavior when introducing innovative products. I hope you enjoy this great conversation with Stewart Butterfield.
We Don't Sell Saddles Here
Patrick: [00:02:51] I found an interesting place to begin our discussion because of something that happened to me in 2017, so I took over our asset management business back then, and probably like any new CEO, I was a little bit floundering trying to figure out what I was going to do, what direction to take the business. There was a couple pieces of content that really influenced my direction. One was Positioning, a book about marketing. Another was this amazing video called Inventing on Principle by a programmer named Bret Victor, and a third was an essay that you wrote in 2014 called, We Don't Sell Saddles Here. I have to begin our discussion with that essay. Could you give us the background of why you wrote it and sort of what the main message was? And I'll have probably several questions about it because it was so influential on me personally.
Stewart: [00:03:33] Sure. Thank you. I guess there's two interconnected messages in there. One of them is to really focus on the kind of impact that we want and how we convey that to customers. We said, actually we still say it today, Slack is the kind of thing that you don't know you want, but once you have it, you can't live without it. That's great except that people don't know that they want it. Trying to explain to someone why they might want something that they have no idea that they want is tough. Not just because you have to create that understanding that's a prerequisite for desire, but also because people have a really finite amount of time and they have a set of concerns that are already occupying their mind, more or less all the time.
We talk about this on the product side of Slack a lot. Someone comes to our website and they come in with, if you imagine, like a threshold of intent necessary to do the Google search or to click the link or however it is that they got to the website, they come in with a bare amount to clear that threshold. No one's coming in with an enormous amount of excess intent, the way that they might if they were looking for a ticket for a specific Justin Bieber concert. But if they're looking for a ticket for a specific Justin Bieber concert, it doesn't really matter how bad Ticketmaster's website is.
It could be slow, it could barf on their credit card details, it could show error messages, because they know exactly what they want and they really want it a lot. Whereas, when they come to our website, they don't really know what they want exactly. They have enough of an idea that they thought this was worth checking out, and the non-specificity and the bare minimum threshold of intent means that we have to work extra hard and really put things in their terms.
I posted this on Twitter many years ago, asking for a name for this phenomenon whereby, when you go to a restaurant website, and this is a lot better now because Google local and all kinds of websites, service templates, and stuff like that. But traditionally, you go to a website for a restaurant and you want the address, the opening hours, the menu, or the telephone number. But what you get is like a slow loading Ken Burns effect for next food, and this music, which is really irritating, and then only 30 seconds later can you get it, and when you finally find the phone number, it's an image of the phone number, you can't text and you can't tap it. All this stuff.
I asked like, why is this? Someone came up with the name, owner's delusion, which I thought was perfect, because the people make the restaurant websites and the restaurateurs themselves inevitably have gone to other restaurant websites and the situation has been reversed and they know exactly what they want and they're probably frustrated as well. But when it's your own thing, you think it's so important, so significant that everyone is going to see it that way, and they don't. That was a much longer answer than I intended. Curious about the impact it had for you.
Patrick: [00:06:19] The subheader in the essay that says sell the innovation, not the product, that's sort of the jump off point of why it was so interesting to me. I think you said something like the real measure of innovation, is it some change of human behavior? Maybe you could explain that contrast between innovation and product.
Stewart: [00:06:36] Innovation is a weird word. It's kind of a collaboration where the meaning is so flexible that it can sometimes be difficult to hop on what people are saying. I got this from a extended essay or very short book called, Who Do You Want Your Customers To Become? That concept of the innovation being, in a very literal definitional sense, the change in human behavior was eye-opening for me. Because there is no innovation, no great innovation that had the result where people behaved exactly the same before and after.
Anytime people change their behavior in a really significant way, it's a result of some innovation. In the best case, and this is not true of every organization or every individual who try Slack, but in the best case, using Slack has a pretty profound change in how people operate. Because you go from typically a world where the principle means, your primary means of communication is email, to one that is channel-based. The difference is, when you join an organization, you start off with completely empty inbox despite the fact that there might've been tens of millions or hundreds of millions, or even billions of messages exchanged before you got there.
And you have no access to that history, and you also have a very partial, incomplete, fragmented view of what's going on across the organization. Maybe if you're lucky, one basis, point worth of access to the emails that are flowing through the board. Whereas in Slack, even for an organization that has tens of thousands of people, you could have 10%. So, you have three, sometimes four orders of magnitude, more access to information, and that results in pretty profound changes to behavior.
Patrick: [00:08:07] If you think about the concept of marketing from both ends, this zeros in on why this was so impactful on me, so everyone wants to innovate or create an innovation, but the builder, and they will come, theory just tends not to work out too well. Like you said, most people don't even know what they're missing. Talk to us about what you've learned about this term marketing from both ends, what that means and how you've gotten better at it as a business person and leader through your career.
Stewart: [00:08:31] I really struggle for the right way to convey this. Lately, this is not very effective so far, but lately I've been using the yin-yang image, where one side is market and the other side is product. Because people talk about product market fit. Because software is fun for many people to make or products are fun to think about, or people perhaps have romanticized an incorrect visions of what Steve Jobs does, or something like that. The emphasis really tends to be on product. Whereas marketing people often reduce to promotion or advertising. But I think you could argue about which side it is, but that boundary between the product and the market, where the product hits the human attention or the creation of utility is really the important thing.
Sometimes you can just adjust the product to fit exactly the market or people's mental models or behaviors, but much more often, you have to change those. Changing those is a whole different kind of work, and you have to take advantage of the concepts that people already have because no one has the time or attention to learn an enormous amount just to figure out what your product is. Going back to the website for a second, I would tell people to imagine someone comes through the website, they're annoyed with their boss because the conversation they just had and they have some health concerns, like they're worried they might have ulcerative colitis.
And they got in arguing with their spouse that morning and they're worried their kid is messing up at school, and then they come to the website. It can be really tough to break through that. I think the genius of Steve Jobs was certainly on the product side, but probably even more on the marketing side, having an intuitive understanding of what people would accept or how people would interpret something, and really trying to focus on where those two meet, as opposed to the intrinsic properties of the product, which is where a lot of us get hung up.
So, you have marketing doing blasted awareness for your in comprehensible thing, and you end up tweaking and iterating on features that people could only possibly understand or get value from once they were already using the product. Then there's just this big, giant gap between the product and the market, or between what you're making and what people understand they might be buying.
Patrick: [00:10:45] It seems also like one thing to do here or consider is that you want to just create the market in the first place. So, if the product is sort of bottom up and it's good and useful, the top down is sell the dream, so to speak. In your essay, again, the idea is sell horseback riding, not saddles. I'm curious how what you've done in Slack's history or just across your career, that you think of as the most successful strategy for that market messaging creation.
Stewart: [00:11:12] I feel like I and the team more broadly have tried hundreds of things. It's rarely the case that something just works. It's more that each action kind of chips away at the problem. Maybe I'm just especially cynical about it . I'm having trouble thinking of the ones that actually have been effective.
Patrick: [00:11:30] One way to think about this too, for Slack specifically, is if the idea is changing, so literally, the first thing I did when I took over the business after reading the essay was to install Slack. The standard switch, like an email based culture, a lot of loss-ness and lack of context building for even a small firm, we were 40 people or something. By switching, all of a sudden, of course, we get much more context in the way you describe, but there was organizational. For me, it became clear as it happened and people were skeptical and slowly adopted it, but it was a change to how the business communicated. Another way to think about the question is the most interesting ways in which you saw companies change as a result of adopting this, that then you could communicate to other companies so that they could participate in that same benefit.
Stewart: [00:12:13] Thinking about that way, I think the challenge has often been trying to point out a problem that people don't perceive as a problem. Not because it's not a problem, but because they take for granted or are so used to it, that they assume that nothing can be done about it. In talking to leaders, it's often, whether they use a term or not, it's a question of alignment. That just means that everyone on the team has the same understanding of what's happening, where we're at, what the objectives are, what their respective roles are.
when there's a big deviation in people's thinking about this, you end up with a lot of ineffective work in the physics sense, being exerted. The difference between the best and the worst performing teams can be, not just like a very large multiple, but often, I think the lowest performing teams have a negative impact on the rest of the organization. It's a real challenge. People think about it all the time, but they think about it solely in terms of the kinds of words they can use or what illustrations are in the presentation that they're creating or something like that, as opposed to they have to work against the tools that they're using.
Really highlighting the ways in which email has made it difficult for people to achieve the kind of alignment that they want has been one of the most effective ways. But again, it's hard to point out a problem to someone if it's so taken for granted or has such a deep assumption, that this thing can't be changed, that people are unaware that they actually have a challenge.
Products & Design
Patrick: [00:13:43] One of the ideas I really liked that I, I'm stealing the term from Sarah Tavel at Benchmark, is this idea of a core action in software products. So, something that you're going to ask the customer to do over and over again, it's kind of obvious what that is in Slack, which is post messages and have conversations in these various channels, and sometimes directly with other coworkers. I'm really curious how the concept of design, product design, and just design in general, has evolved for you, going all the way back, maybe to the 5K contest. You could describe what that was and sort of the power of constraint or limitations in design.
Stewart: [00:14:15] I'll try not to get too deep into a philosophical rat hole here, but-
Patrick: [00:14:18] No, let's do it. I like that. I like the rabbit holes.
Stewart: [00:14:21] Okay. In 1999, I started a competition called the 5K contest and it was the best website or page that you could create in five kilobytes or less. That's, from today's perspective, an absurd, really small data, but in the world of 1999, where, I don't even know, maybe 35% and 40% of US households had internet access of any kind and it was 95% plus dial-up, you really had to respect the constraint of bandwidth consumption. I think because of the excesses of the era and the fact that the competition actually began in early 2000 when the dotcom crash started, there was a real much greater degree of enthusiasm for this thing than I would expected.
Because I thought like my community of internet people, some bloggers and stuff like that would enjoy it, but instead, it was like, I was interviewed for Playboy and in Brazilian newspapers, and Pravda, like the Russian news service. It just became a worldwide phenomenon. Part of that has to do with just with the depth of the relationship between constraints and creativity. Some of people listening here will play music, and you think about the key or mode kind of signature, the style of music.
In poetry, you have the rhyme scheme and the overall form, the neater, and there's all these conventions in visual art. The history of the 20th century was often, what happens if we remove constraints? And you get all the way from music that pushes the boundaries about some off notes and dissonance to pre-jazz and atonal music, and same thing with poetry. You get blank verse and free verse and abstract expressionism and visual arts. But even in, if I had the century, I don't have an off the top of my head, but before the common era, there were Latin forms of prose that were just, what is called a lipogram.
What I can write using only this vowel, no words that contain this vowel. There's a great book, it's by Christian Bok, B-O-K, and it's five chapters. In each chapter, you're only allowed words that start with one of the vowels. The E chapter is definitely the best. The U chapter can be very difficult. Anyway, that relationship between constraints and creativity, I think is like an ancient one. It's really so primitive and an embedded in creative activity that people fail to notice it. Again until the 20th century, when playing with the constraints themselves became a thing, and I think the five kilobyte constraint was a really useful one in trying to come up with a very pure concept.
But I think the reaction wasn't just about then with consumption, the reaction was about dotcom party scenes and the excesses. For people who had been online for longer, like had got online in '92 just before the web started to take off, there was definitely the techno, utopian, hippie ethos of the pioneers of early personal computing technology and the internet, and all the commercialism seemed really in conflict with that.
Patrick: [00:17:44] Maybe the second part of the question is how you suggest entrepreneurs think about using constraints or limitations to build better products. You've already pointed out like limitations create creative output in some interesting way. By telling people they can do less, they're able to more creatively do more in the strange way, whether or not you think that can be harnessed in business as well.
Stewart: [00:18:05] Yeah, I think you can. I'm going to introduce a visual metaphor, and no one can see me gesturing with my hands and no one can see the illustration, but probably everyone can imagine two dimensional graph, and on the horizontal axis is the effort that you put into some product design creation. On the vertical axis is the value. The curve is almost always very strongly S shaped. In other words, it's kind of flat at the beginning and then it has a very steep rise and then it tapers off again. It's if you imagine, if you're making a software product that has users, early on in the process, you have to design the schema for the users' table and then you have to implement that. Then there's a bunch of stuff around authentication and salting and hashing passwords, and the forgot your password flow.
If you just did all of that, you would be very low on the value scale. You would have exerted some effort, and there's no value for anything or anyone, because it doesn't do anything yet. Then there's this part that's really steep. Then there's a point beyond which you're expending all this energy and not getting a lot of result. That can be really hard to recognize. I think it takes a lot of experience and practice to understand when you've got to this deep part and understanding the difference between the problem of, we're exerting effort and it's because we're past the point of diminishing returns, that we're not seeing any value or we're observing effort and we haven't got to that steep part of the curve yet.
Because oftentimes, the reaction will be we added this feature or we made this change and we didn't see much of a result. Therefore, we're not going to pursue it anymore when they're just shy of the steep part. The steep part of those utility curves shows up in an inverse way in people's behavior. I lived in San Francisco two different periods, and one of them was pre ride hailing apps and the other one was post ride hailing apps. In the pre-ride hailing apps era, you would call a cab and tell them your address, and then mostly they just wouldn't come at all, even if they said they were coming, and other times it would take 45 minutes or even an hour.
And the cabs stunk and the drivers were rude and it was just a terrible system. Now, like most people, I will never call a taxi company on the phone and then tell them my location verbally, and then wait an indeterminate number of minutes where I have no feedback loop, then get in the cab and then have to pay with cash, if I can help it. If there is as a ride hailing service, and from a functional requirements point of view, select specify location or wait arrival facilitate payment. There's kind of check mark, check mark, check mark. They're, in some sense, functionally equivalent, but the experience is so different that people's behavior radically changes.
I don't know the number of the increase of hired rides in the US, but I know it's a large multiple, and I also know that people's preferences, when you cross this threshold of it being easier enough, are such that they will never change their behavior backwards. The challenge, I think, is to find those steep parts of the curve, and the constraints are really useful in winnowing down what is more or less an infinite space of possibilities of a future starter, things to try or design approaches, to something that's a lot more manageable and that helps you determine whether it's worth proceeding down a different path.
Patrick: [00:21:21] Both of the companies for which you're famous, Slack and Flickr before it, emerged from, don't want to use too harsh a word, but failed attempt at creating video games that would have huge adoption online. Why do you think that was? Why have you continued to kind of go back to that concept video games? Obviously an interesting design or constraint challenge in and of themselves. What's behind that happening twice to you? When I doubt that's really happened, except the only one other person I can think of is Jason Citron at Discord was also trying to create a game and emerged with a tool, kind of a similar story. Why has this happened twice to you? It's so interesting to me.
Stewart: [00:21:54] I mentioned that I first got online in 1992. I should say I was living in my hometown, Victoria, British Columbia. So, it's on an island off the west coast of North America, so it feels very provincial. It feels very removed from the rest of society. If it's after 9:00 PM, you can't even get a ferry to go to the mainland. The web hadn't taken off yet, but the internet had an enormous amount of activity, not relative today, but still hundreds of thousands of people who were connected. And they were using Usenet, and email, and IRC, and a Unix program called talk, which was like messaging, where you could see each letter as it was typed.
So, it's kind of like a nerve-wracking and very irritating thing to use. You could seek backspacing and to fix their typos. For someone who grew up feeling utter removed to the greater world, the idea that you could transcend geography and you could transcend time and you could find these communities with interests. It was really just magical. From that moment, the idea of the use of computing technology to facilitate human interaction was just the most interesting practical idea in the world. There might be philosophical ideas that are more interesting. The first game was called Game Neverending, and came from the realization that you could make dynamic webpages.
This was 2001, 2002, it was called remote scripting. A friend of mine, [Erika Sola] , was one of three or four people who were really invented this. Later became Ajax and then Single-Page apps, and now it's just like a webpage. But the tech was so cool. Think of it, again, that's entirely a browser and make it something that's really dynamic. I think it failed ... Actually, it probably would have failed anyway, because the idea is not especially commercially viable. It was a no combat absurd, surreal, Monty Python meets Dr. Seuss world, has a pretty narrow appeal.
But it also failed because it was 2002 when we started it. So, there was the dotcom crash and then there was WorldCom and Enron accounting scandals, and then there was 9/11. It was just the all time worst market conditions, and certainly the worst time to try to raise money for something that was online. With that field, we created Flickr. The second failure was very different. We created Flickr, it was acquired by Yahoo. We worked at Yahoo for several years and then we thought, okay, now it's 2009 versus 2002, and there's all of this amazing open source software that didn't exist before, and we've gone from a world where the Netherlands was the only country to have greater than 50% penetration of internet connections at home, to almost every country.
There's an order or maybe two orders of magnitude, more people online, and they had faster connections, and they had better computers, and hardware was cheaper, and we were more experienced. It's just all of these 10X factors that led us to believe that, well, we'll never possibly fail this time. We did again, and this is the long story, and maybe the idea wasn't commercially viable, but I think that the real reason was we went all in on flash as a client technology and kind of a content pipeline management technology at a time when flash was about to die and people's discretionary computing time shifted from PCs to mobile.
We had been using, going all the way back to 1992, internet relay chatter, IRC as the backbone of our communications. But because IRC is a 1989 era protocol, it didn't have really basic things that people associate with messaging services, so it didn't have the ability to store and forward messages. If you and I, Patrick, were online at the same time, I just couldn't send you a message. It just wouldn't reach you. So, we built a bot to log all the messages that people sent so you could catch up again when you came back up online. Just incrementally, over the course of many years, we made these improvements that were based on something being so exasperating that we couldn't stand it anymore or something being such an obvious opportunity that we couldn't help, but take advantage of it.
When I say we, I mean it's 90% Cal Henderson co-founder and CTO. Once we had the messages logged in database, you want to be able to search them. And once we realized that everyone's attention was focused or aggregated here, we started doing things like database alerts would flow into the channels, or when you uploaded a file to the file server, that would be announced with the URL, so you didn't have to tell people, and the daily stats around where you posted into channels. All the stuff that became precursors to Slack. Then, when we decided to shut down the game, we realized that we would never work with an assistant like this again, and so probably other people would like it too.
Although, when you said other people at that point, we thought other eight person software development teams. Here's the thing, all four of those, the two games, Flickr and Slack are to me, kind of at their essence, the same. It's all the use of computing technology to facilitate human interaction. So, you have a massively multiplayer game that didn't work and massively multiplayer photo-sharing that did work, because that was what separated Flickr from other things. It was, you could put a photo and make it public and other people can comment on it. You could tag it and other people could tag photos, and then they would show up together and you could create groups.
Then another massively multiplayer game that failed and then massively multiplayer workplace software. But there is a real deep, common thread, different in the work context than it is in the play context, but there is the establishment of identity and participation or interaction and the gradual accretion of the residue of those interactions that increases in value over time. I just like making software that connects people, because here's a specific set of challenges there, and the context is less important for me than the fact that, that's what it's for.
It turns out that, that's, excluding Google, I guess, almost all of the value of the consumer internet at this point has different ways of facilitating interaction with other people. Google obviously takes advantage of those interactions and discreet human actions to build the index and provide search. It's really big and grateful that Slack is the one that was really successful because it is such a deep and interesting problem, a thread that is very common through all of those, even if the veneer is different.
Communication Tools
Patrick: [00:27:55] It's a fascinating idea that most of the value is just finding ways to help people communicate with each other. How much white space do you think is still there? I think even when Slack started, if you had pooled an audience like saying, do we need more communication tools? And maybe if you pooled them today, everyone would say no, like there's 40 different ways that I communicate with other people in every medium voice video, text, Slack, WhatsApp, there's a million places. So, it seems like settled ground. Do you think that, that's true or that we, as people, will just continue to find interesting, innovative new ways to communicate with each other?
Stewart: [00:28:27] A little bit of both. It's important to understand why I think Slack works or why it's been successful. It's not because it's a better way to enter text into a box and hit enter to send it, or to read other people's texts, although we know, obviously we focus on the design, but channel-based messaging is just very different than individual messaging. Slack is a communication platform that puts the team or the organization first, as opposed to putting the individual first. An email, I have a unique inbox that is different from everyone else's, as does everyone else. The difficulty of that, I mentioned, the first day on the job you have a completely emptied inbox, despite the fact that there's all these messages that already existed.
But the other problem is I just don't see things the same way you do. Unless I manually select emails to forward them to you, you'll just forever miss this context, whereas you can join a channel six months after it started, and it didn't matter that you weren't there. The history was being created and indexed, and you can catch up on it. If you end up in a situation where there's a channel for everything that the organization is paying attention to, so for us, that means every customer, every team, every office location, every function within the company, every incident, every event that we're planning, everything, then everyone knows where to go, at least theoretically, to ask their question or to give their update or to get caught up in something, which is very, very different than email.
Channel-based messaging is a new category and there will definitely be continued iteration and competition in terms of products, but I want to make clear that it wasn't just, this is better than iMessage because of style of the app or anything like that. It's better than iMessage because you can have channels.
Patrick: [00:30:15] If you think about the future of communication, another angle is how we can get better at it as people, especially within organizations. Because when we talked last time, you made this good point that, especially executives, literally all they do, their entire job is different forms of communication, whether that's message building, message delivery, meetings, planning, talking to other executives, talking to clients, whatever. It's just all communication. We don't ever receive training on this topic in our lives.
We're not taught how to communicate well in school. We're really not taught how to communicate well by our organizations, by our jobs. What are your thoughts on the future of this skill? Because if all this value has been created, facilitating communication, but there's been nothing invested in making people better at that, it seems like a very odd dynamic. What do you make of that?
Stewart: [00:31:03] It's kind of hard to explain. I think the only way of really understanding the delta between how much should be invested in helping people become more effective communicators or investigating tools for better communication and the amount that actually, is that people just, they're so familiar with the problem that they don't see it as a problem. They don't see it as something that can be solved. People hate email, people hate meetings, people hate Slack, people hate everything that has to do with communication. Don't get me wrong.
Always there's a population for each of these things that hates it because it can be really taxing, because you have to overcome different ways of thinking in order to get your idea across, and you have to kind of parse out what people are trying to say. There's subtexts and there's power dynamics, and there's all these other things at play that make it hard. But yeah, for executives, it's more or less 100% of what they do. For most managers, it's 95%. Where there are real attempts to innovate or to do things differently, they stand out in a pretty profound way. Everyone knows about this six page Amazon memo and how many of those things can you count?
How may times have people actually taken their frustration with the ineffectiveness of certain communications and said, okay, here's what we're going to try. We're going to write a six page memo, we're not going to have any presentations. Everyone's going to sit in silence and read it all together from the beginning of the meeting, so they really understand we're in the same place, and then we're going to discuss this. That format forces people to come up with a lot more clarity in how they're communicating and all the rationale that goes along with it.
That's one. Another one I've heard, and I don't know if this is apocryphal or not, but Sheryl Sandberg saying no more PowerPoint presentations, but she meant in meetings with her, and the sales team, thousands and thousands of people took that as no more PowerPoint presentations period, and then they had an all hands and there was a big complaint from the salespeople saying it's really hard to sell to people without giving them presentations. It's slightly worse than that, because most of that communication is just about really basic acts of coordination and communicating state.
Or, I guess if you include changes and objectives as state, and it's more or less all communicating state. That's the daily standup when the quarterly business review and all the approval medians and the catch-ups, I would say 50% plus of the communication that's popping or 50% puffing of the effort of knowledge workers is just that basic communication and coordination. It's not the interesting creative strategic parts. Although, obviously those are really important. When you look at it that way, you have a company with 10,000 employees and let's say a billion dollars in payroll costs and 50% of the effort, 50% of the time is going into this basic communication coordination.
There's $500 million, maybe zero effort towards trying to improve the efficacy of that. If you're able to make any difference at all, it's disproportionate to more or less any change you can make across the company. That is the biggest difference between the best performing teams and the worst performing ones and also the best performing organizations.
Patrick: [00:34:03] Do you think that means that we might start seeing really fast acceleration and productivity at firms that start placing more emphasis on this? I like Dustin Moskovitz's phrase work about work and how much of our time is, like you just said, spent on just coordinating stuff. Now that we have a glimpse into what a purely digital organization might look like, thanks to COVID, what do you think that means? What is digital first? What does that mean for you as you think about your company? What do you think it unlocks potentially? Does it start to address this more structured, well thought out communication problem for companies to move faster in the future?
Stewart: [00:34:38] There's an enormous opportunity in front us now. My hope is that we'll collectively really take advantage of it. I'm a little bit pessimistic about it though, because that's not where the conversation is at. The conversation is that how many days per week do people come into the office? I think that the term hybrid, while it may be useful right now, in 10 or 20 years, it's going to feel like paperless office does now. It's kind of a quaint and in some sense, bizarre way to think about it.
Because the amount of paper in an office is no kind of indication of how effective it is, or even the methods that it employs to operate, because paper is useful in many ways and people getting together physically is useful, but we obviously have the technology today to do what was so exciting for me and others in the nineties, which is transcend geography and transcend, to a large degree, time. The degree to which we take advantage of that, I think we'll have real productivity impact. Then maybe there's a footnote there that whether that shows up in the econometric sense of activity, we'll see, because the follow up, the productivity paradox, maybe, in 1987.
We've been struggling with why isn't there the kind of impact from information technology that we all know that there must be? But yeah, right now, I think that there's just this real opportunity to think about how we change the way that we work and never waste a crisis. We've been knocked so far off of equilibrium that people will accept changes that might've been more difficult to implant previously. I think Dustin and I are very aligned in how we think about this stuff, with maybe one exception and maybe work about work is just more kind of marketing speak.
But while I can list all these reviews and update meetings and catch ups and stuff like that, in a way, the tone of voice indicates, I don't think they're important, I actually think they're critically important. Without the context of the objectives and current state and roles, you just are adrift and people end up working on the wrong things. I think the most fundamental lament of CEOs that I've talked to, and we have a great weekly SaaS CEO call, and obviously spend a lot of time talking to other people, and there's all this support and commiseration, stuff like that.
But at heart, most of it is trying to figure out why someone ended up doing something when, from your perspective, that's obviously the wrong thing, but they're not dumb, so something's wrong somewhere in the system. We really should, we, again, collectively use this as an opportunity to get creative and to explore. I don't have all the answers here obviously, but to kind of explore and iterate and experiment on new ways of communication that give us the best of both worlds. Because the other thing about digital first, work from home, hybrid, all of this stuff, it came from a comment that, in a very heated staff meeting, my team at Slack, our head of people, Nadia Rawlinson, pointed out, like when we're talking about what's different, are you talking about what's different from February, 2020 or are you talking about what's different from today?
Because it's easier to imagine February, 2020. Everything's the same, except we only come into the office two days a week, but obviously, we move into the future from the present moment. That's quite different. It's not reducing our time in the office to two days a week. It's increasing our time in the office to two days a week. The other thing about thinking about starting from today is none of this work remotely stuff had to be in people's houses except that there was a deadly pandemic that forced us all to be effectively quarantined inside of our houses.
But you could just as easily go work at a coffee shop or a co-working space. That's not your office because you live in a place where your company doesn't have any physical offices. There's huge advantage to companies to be able to hire in a greater variety of places. There's a huge amount of optionality and value for employees to have the choice about how much time they spend in the office. Everything kind of worked out. If you think about what's different from today, life will be better for everyone because we'll be able to eat at restaurants and go to the theater and grandparents can visit their grandkids and all that kind of stuff that we were prevented from doing during the pandemic.
Everything has worked and life gets better for people at an individual level. And then, we'll also get back the office and in person collaboration as another tool for creating the kind of alignment which is necessary for the organization to achieve stuff. There is I think where the struggle is in most conversations, because the purpose of an office is many folds. San Francisco's our headquarters. Probably, I think about half of our employees are based there, although many have moved during the pandemic.
We have a 10 story glass tower and it has our logo on the side of it, and it's this projection of power into the world, and it's a place to bring recruits and impress them. It's a cultural touchstone for existing employees. It's a place to host customers, it's a place for these all hands and events, and training people and inspire collaborative conversations. Then it's like factory farm battery chicken housing for people to sit and use their laptops all by themselves with no interaction. That's the least valuable, but takes up the most square footage.
I've said almost that exact same thing verbatim for like nine months, maybe 12 months now, yet there are zero new plans for how Slack intends to lay out their office side, but I want to point out the potential hypocrisy in how hard it is to effect these changes, because I don't want to go back to an office. That's the greatest amount of square footage. I do want to think about all of those other uses of the office, other purposes and how we can get the most value out of that investment, but also just had the greatest impact in how well people work together.
Leadership
Patrick: [00:40:15] One of the things that sits next to this issue is how leadership changes in a digital first organization. I think, in some ways, it's easier. There's maybe more leverage like Slack provides, in some ways, it's a lot harder. There's no substitute for sitting in a room with people and having them see you and communicate with them. You've said elsewhere that there's this book called Leadership and Self-deception, which has, I think informed some of your thinking on leadership. I'm really interested in this idea of self-deception. Maybe you could describe what that means to you in the context of leadership and then we can talk about how that might apply to becoming a better leader in a digital first world.
Stewart: [00:40:50] Leadership and Self-Deception is a book that was given to me by a slack employee. If you ever read it, or maybe when this conversation is done, you'll think that if anyone ever recommends this book to you, then they think you're an asshole. It's a very intriguing title. This employee gave it to me and it was on the, literally on the coffee table in a stack of a hundred bucks for like a year, and then I finally brought it upstairs to the bedroom and put it on the bedside table and where had sat for another year, and then I finally read it.
There's a lot in it, but I think that there's a pretty easy way to synopsize it. That's using one of its stories where it's a man who's in bed, and it's 2:00 in the morning and his wife there, and the baby starts crying, and the man wakes up and then he starts thinking, oh, that's so annoying. I'm so tired. I hope my wife wakes up. I have a really important meeting tomorrow. I just really need to get some sleep. God, why isn't she waking up? What's the matter with her? She's so lazy. You can do this with a lot more nuance and color than what I just did.
But the point of the book is when one violates their own sense of what's right, the immediate response is to create a narrative where someone else is wrong. So, she's being lazy. He should've heard of that and taken care of the baby when he heard the baby crying. Despite the fact that he's already awake and not going to take care of the baby, there's an instinctual response to come up with a narrative for why his wife is at fault, and Leadership and Self Deception is mostly about how that shows up in the workplace. It shows up a lot.
Maybe there are opportunities in a digital first era to help people move beyond those, but the least effective relationships between managers, employees, to executives, and rest of the company, or CEOs, and everyone else, I think, arise out of this disconnect, because the buck stops here. It's our responsibility as leaders to figure out what needs to happen to create the kind of work and the culture and the results that you want. Instead of being a CEO of a company with a bunch of people, you were the physicist doing basic research and trying to develop a new material for something.
You might get frustrated and hurl something against the wall or something like that, but you don't blame physics for the difficulty that you're having for accomplishing this thing. Leaders blame people all the time. Sometimes people are in the wrong role, but if they're in the wrong role, who put them there? There's a lot of thinking or falling back on the idea that these people are too stupid or they have some other defacto or problem that prevents them doing the thing that your genius mind has envisioned them doing. That's never the case though. It's on you to figure out, just like the physicists or the material scientists that strongly degrades a new material, what it is that's causing the results that you don't want? And how to uncover that dynamic and repair it.
Patrick: [00:43:39] It could sort of be summarized as the classic, it's not you, it's me problem that we always fall in over and over. Is there a particular example from your own history of, I just think it's a really powerful concept, of some sort of self deception that you underwent and ultimately fixed or recognize that you could think could nail home the point?
Stewart: [00:44:01] Of course, that’s one way I ever do . It's very deeply ingrained. There's no one single dramatic story, but it is ... My instinctual reaction when something doesn't unfold the way that I hope or expected is to blame people, to blame other people. One time out of a hundred, or maybe even less than that, there is some one person who did the wrong thing and therefore like a blame worthy action, but it's almost, in every case, a failure on my part, sometimes other leaders, to recognize the incentives that have been created that lead people to do the thing that you don't want them to do.
Let me get a little bit more concrete. I have put enormous effort over the course of many years, with limited success, into getting product managers and designers to understand that it's fine to be wrong. Being wrong is just the process of getting to the right answer. If you imagine you're sitting down to write something, you have an important message, and you write the first sentence, and then you write a second sentence, and then you start the third sentence, but you realize, oh, the first two sentences are way too grammatically complicated, and I can simplify that and you change it, and you're editing, you make typos, and you go back and forth and back and forth, and it's this real iterative process.
And you don't, at every stage, berate yourself or chastise yourself from making some mistakes. That's just how you get to a good piece of writing in the end. That's the only way to make a good software, in exactly the same way, is to take a best guess and make your attempt, and then realize what's wrong with it. There are a number of things that lead up to this position where people feel like their job is to be right on the first try and being right on the first try is more or less impossible unless you do something that really doesn't matter.
People end up in this position where, rather than thinking about the food that you're cooking at the end, they're thinking about the recipe. Before they ever try cooking it, they're arguing about the recipe and preparing slide presentations that have static images of the recipe, and then debate about whether it's a good recipe or not, static images of user interfaces, for example, without having to ever built anything. If you're really cooking something, you taste it along the way, and you adjust as necessary, and you thicken it, or reduce it, or you add more water, or make these adjustments in response to what's happening.
But if you feel like you can't proceed until there's a high degree of certainty, both in yourself and, I think critically in others, then you use that all of the time that was available on sitting in conference rooms, looking at photos of recipes and then asking people if they think it's okay, because if enough people say, I think it's a good idea, then if it doesn't work, there's a diffusion of responsibility. My job, if I don't think that's what should happen, my job is to figure out what incentives to put in place, what to communicate, what techniques, what kind of training we need, how I need to articulate the goals so that, that doesn't happen.
But the self-perception is the fault is external to me, an error that these people are making and these people are causing all the problems. If you're the CEO, you're the CEO.
Patrick: [00:46:59] Does it follow then that the highest calling of a leader would be to create the conditions for others to experience or achieve mastery in what they do? What do you think about that concept of mastery and it's importance in business?
Stewart: [00:47:14] People like exercising their talent and improving and learning. I try to take a lot of inspiration on the software product development side, from the way video games work, because if you think about first, you can move and then you can move and jump, and then you can move and jump and shoot. Each level kind of unfolds, introducing a new dynamic for you to master until basically the game ends because the game is essentially one very long tutorial, and once you figure it out, how to play the game, it's over. I think that's a good approach to making any kind of software, but the satisfaction that people get from that learning is something that feels universal and it cuts across different forms of creativity and different hobbies and different everything.
There's nothing more frustrating than having that thwarted. People end up very unhappy at organizations where nothing can get done. That's probably, this is kind of a separate topic, but I was having a long discussion with a CIO of one of the big banks, and we're talking about how to measure software developer productivity. She had been around for a long time too, and all of the ... There's all these famous failures, like drop in the number of lines of code that people write and then people just change their coding style, or count the number of bugs that are fixed and people, before they make any change, will file a bug and then close it afterwards.
Introducing all this overhead, which obviously slows things down. Maybe the best way to measure productivity is employee engagement or happiness because people can be unhappy for different reasons, but they can't be happy without being effective. It's a necessary condition. So, if the engagement scores are low, if your software engineers are happy, then they're definitely not effective.
Philosophy, the Future, and More
Patrick: [00:48:59] For context for these conversations, I write down like these weird little notes to myself. One of the ones that I wrote for this one is hippies and APIs. I think you've talked before about how the hippie mentality may shed some light on this software concept of APIs and how powerful they are. Can you explore why those two things might belong in the same sentence?
Stewart: [00:49:17] The idea of open source software, open protocols of APIs that are available to everyone, I think this is one that's ingrained in the approaches and kind of antecedents of where we ended up today. Often, there's a Trinitarian approach that has real value for the people who engage in it. Basically every open source software company has followed that same pattern of make this freely available, and then there's still an opportunity to profit off it rather than keeping a black box and proprietary. But I will say that in the early days of Flickr, one of the things that made us successful was we had APIs for everything.
You could just do anything you wanted and that excited people and they experimented. A handful of apps that people created out of the APIs, I think had genuine popularity and were useful, but mostly, it just created a community of enthusiasts. In the case of Slack, it created 2,000 apps in the app directory and something like 900,000 custom integrations that are actively used by a customer. These are integrations that customers created themselves. That's like a mind boggling number. Had we not done that, had we kept it in a black box, I think there would've been a lot less utility.
Patrick: [00:50:29] So, it's almost like if you adopt this open mindset to give tools away, it's a long-term advantage even if it comes at maybe a perceived short-term cost. If you just kept everything proprietary, maybe you capture more value in the near term, but it sounds like your experiences in the long-term kind of a hippie idea, that open source or just letting people have access is a net benefit to companies even if it might not seem like that in the short-term, does that kind of sum up the idea?
Stewart: [00:50:56] Yeah. I think people talk about abundance mindset or growth mindset. If there's one part of this interview that's most interesting or relevant to people who are more on the investor side, it might be this. Software is so far from zero sum as an industry. There might be a handful of software categories that are more or less saturated. In other words, every large organization has chosen an email platform. So, to get your new platform adopted, you're going to have to take a customer from someone else, but the whole thing is growing at a mind-boggling rate. I haven't done it, but someone could easily look back over the software companies that have gone public in the last 24 or 30 months and sum up their market cap or sum up their revenue today. All of that is stuff that didn't exist five years or seven years prior.
If you think about the number of software vendors that any given company has, and Slack, a company of 2,700 to 2,800 people has buy software from 450 different vendors, that's still the ... Times the average for large enterprises is a thousand different cloud services in use. That number has just gone up monotonically for decades and will continue to go up decades, where you look at the number of dollars per employee per year, that any company spends on software, and that just goes up. The number of minutes per day that people spend using software goes up inexorably as well. All of these things are going up, so I figure you're much better off in the world of software if you have that abundance mindset or that growth mindset and not think about zero sum dynamics.
Because we're really, I think, today, maybe not quite as bad, but things are going to disappear or ways of working are going to disappear because of what computers did right before they were disrupted in that way. I'm old enough that I had a college job that was a file clerk that was like correspondence comes into the office and you make photocopy of it and then put it in a file, and that's like the database of correspondence of the company and sending more stuff that was outgoing. Maybe there's people who still do that at that margin, but there's definitely not very many of those people left.
But the weirder ones for me to imagine are like, if went to a bank and in 1932, how did they know how much money you have? Is there an index card that says Patrick O'Shaughnessy and they just like flip through the Os until they find it and then they pull it out, and it says you have 300. I mean, genuinely, I have no idea. Computing completely revolutionized the stuff that was easy. In other words, doing arithmetic perfectly at scale and remembering things perfectly forever and a bunch of other stuff that computers do better than people, but we're getting into finer and finer territory and automating all the things that are automatable with almost definitionally, more tractable problems for software to approach.
But that means that the work that's left over for humans to do is going to be more demanding of their intelligence and creativity, because that's the hardest stuff automate. This comes from that openness APIs and taking a generous approach. I think that has worked really well for many, many companies. I think it gives you a better and more accurate picture of the future that you're working towards.
Patrick: [00:54:02] Can you talk about this idea of the backhoe and the shovel, and whether or not, you just talked about how much leverage software technology is giving us, whether or not you think technology is fundamentally amoral? I know that's kind of a weird philosophy question, but you're a philosophy guy like me, so I figure, why not? What do you think about maybe the downsides or the amorality of tech?
Stewart: [00:54:21] I mean, I think it is largely amoral, and there's a bunch of famous examples of that. But the research that went towards both fertilizers that allowed the green revolution and Nazi death chambers, stuff like that, not to be too grand, but the shovel and the backhoe is something that I use to explain the value of Slack, but also some of the challenges that come along with it. I apologize, it becomes a grim analogy. We want to do for organizations what backhoes do for people whose job it is to dig ditches. Because digging a ditch with backhoe is a lot easier than digging it with a shovel, a hundred times easier or something like that. You can take a lot more ditches. You also can accidentally do damage with a backhoe if not wielded properly, that you couldn't possibly accidentally do with a shovel.
You could kill someone with backhoe accidentally, but you can't accidentally kill someone with a shovel. You have to really be determined to do it. I think a lot of technology ends up in that way, where it's simultaneously much more powerful, but introduces new dangers or new risks or at least a greater requirement for mastery or understanding it, and how you wield that technology so that you have the positive impacts without the negative ones. If you look back at mid 19th century, the Thames and the Charles River in Boston would catch on fire. Eventually, we learned how to have the benefits of the industrial revolution without as many people dying of black, long and dangerous factory conditions and all that.
I don't think workplace software has the same danger profile as the first coal fire plants and stuff like that, but I do think there's a similar phenomenon where it takes us a while to figure out how to incorporate this new power into our existing habits and practices in a way that gives us the benefit without some of the side effects.
Patrick: [00:56:10] What do you think are the most interesting open questions about the future?
Stewart: [00:56:16] One of the big ones I think is how more effective we're able to get in materials and energy. I listened to a podcasts with Marc Andreessen, I don't know if it's new or if-
Patrick: [00:56:27] Yeah, I can see that.
Stewart: [00:56:28] The first half, I was like, just ... It was a remarkable summary, both of the power of software, but also the long arc of technology. At an intuitive level, I feel like everything will work out, but looking at the world in a more practical way with my puff and naive understanding, it can be hard to figure how it's going to work out okay, given climate change and the impact of bringing the other several billion people up to the standard of life and is enjoyed by the people in the richest countries. That's a big one. Maybe less existential, but really interesting to me is, do we ever get to the 15 hour work week? Does that ever happen? The jobs that disappear, as Marc points out, are almost, well, they're usually the worst jobs.
Either the most dangerous, so there was mind numbing, or repetitive. Humans are amazing at inventing new needs because there's just a lot more life coaches and massage therapists and aromatherapy people than there were a hundred years ago. But John Adams has this quote, I'm going to brutally mangle the paraphrasing, but it's something like I must study politics and more so that my sons may study commerce and navigation, very important, naval architecture, so that their sons may study porcelain and tapestry and painting. The idea is this brutal period we're going through right now, it's just laying the groundwork for productive capacity and economic interactions and growth and prosperity.
But the reason that you do that is so that people can express themselves creatively, which is certainly something that I relate to. I mean, there's no right or wrong to this, but my feeling is that the purpose of life, to some degree is that creative expression. Whether that comes out in sculpture or management or technical innovation or whatever it is, do we ever get there or do we invent new work challenges that reduce our capacity for creative activity?
Patrick: [00:58:33] What is your creative outlet? What is the creative output that gets your juices flowing the most that you can do repetitively?
Stewart: [00:58:40] In the work context, it's writing. I don't think I take enough time to leverage that, because if I really work hard in thinking about something carefully and get to that degree of clarity, the impact of something I've written has been certainly disproportionate to the impact of the time I spend in meetings by hour basis. Outside of that though, I recently started playing jazz guitar again. Basically, when I was in high school, I played in the jazz band, mostly bass and something like trumpet. I'm less good than I would like to be, so there's a whole element of mastery, but it's also just like a whole fascinating theoretical component that interacts with the creative component in a really interesting way.
Patrick: [00:59:25] You wrote this fabulous resignation memo leaving Yahoo, I think of it as the tin resignation memo. What was so interesting about it to me was just like an obvious mental clarity that you had reached about the state of things for you personally in your career. Can you just describe why you wrote, and we'll link to it so people can read it, this highly unusual and entertaining and unique memo?
Stewart: [00:59:46] The approximate cause of that is pretty prosaic. I told my then boss that I'm quitting, and he said, "Okay, that's fine. Can you please send it to me in an email so I can forward it to HR." People have parsed it for all of this meaning because Yahoo was in a really weird state at that point in 2008. When I started there, we had significantly more revenue than Google. By the time I left, Google was, I don't know, probably 5X or 8X Yahoo's revenue. But I was just ripping off the style of Thomas Zweibel, the fictional publisher of The Onion, in the paper versions of The Onion. Maybe they do it online too, but he had a column that was written from this crotchety 19th century titan of business perspective, and so that's how that came to be.
Patrick: [01:00:31] If we go back to where we started this concept of, We Don't Sell Saddles Here, and obviously a lot of people out there are either investors looking into businesses that they want to be exceptional or people that want to start businesses, what closing advice would you have for that batch of people through the context of what you've learned building Slack? What is deserving of someone's care and attention and creativity and potential mastery?
Stewart: [01:00:55] Anything that is worth doing. Anything that's worth doing is worth doing well. Gosh, I'm trying to think of which part of this is more important, because it's called We Don't Sell Saddles Here, and there's this whole history that I wasn't aware of when I wrote it, but I use the examples of layman selling athletic wear and trying to juice the market for yoga by hosting free yoga classes, to see if I grow that, or can't remember now if I mentioned Harley-Davidson or Up. They sell motorcycles, but they also sound like independence and freedom and stuff like that. But there's another part to this as well, which is just the degree of crash that you put in. Because of the owner solution, it is far too easy to assume that people care or understand something in a way that's similar to the way that you care or understand the product that you're making when they are coming in from a totally different angle.
I remember, God, six months ago or something like that, I won't say what brand of television interrupted the show to say that in an upcoming software update, some app that I'd never heard of for news would be unavailable. There's definitely a disconnect between the degree of importance that had in the minds of the people who caused that notification to appear and the end users like me. That's why the degree of consideration really matters, the degree of effort, the attempt to get into the shoes of the people who are going to be using the product and making it good.
I worry that am I conveying that well because it just sounds like do a good job, and everyone who ... Presumably, everyone wants to do a good job, but there is this concept of the steep part of utility curve on the one hand or the point at which something is easier enough or more convenient enough that people won't go back. I think that any product, any service that ends up massively successful found that, found something that it makes so much easier than the alternatives that everyone changes their behavior in a pretty profound way.
That would even include older generations of enterprise software and that people love to paint and complain about. That was easier enough than the alternatives of the paper files or trying to build the software yourself or whatever that people adopted it on mass and continue to use it today.
Patrick: [01:03:07] I think that's such a cool closing point, your idea of innovation being the sum of behavior change, like the aspiration to make something so good that people change how they live their lives is a cool closing motivation. I think you might know my closing question for everybody, which is to ask, what is the kindest thing that anyone's ever done for you?
Stewart: [01:03:23] You know what's funny, Patrick is, I don't know how many episodes I've listened to, but many. I'm always lie, if I was ever on the podcast, I would think of that question, and I actually completely forgot until we got to this point. It's hard to know the kindest, but I think somebody that came from a place of great kindness and had a really positive impact on my life was actually one of my undergraduate professors. That's the one going to grad school for a long season. It didn't work exactly, but telling me, "You don't want to do this. If you love it, you can subscribe to the journals. You can go to the conferences, and you're good at it, but I wish that I wasn't a professor and I started in the '60s, and have graduated hundreds or even thousands of PhDs, and they're all looking for jobs."
In '98, I made the choice to give up on the PhD program and go work on the internet, which was like new and exciting and paid twice as much as any philosophy related job you could get. I think if he hadn't done that, I very likely would have ended up in academic, which is not a horrible life, but I'm very grateful for the life that I ended up with.
Patrick: [01:04:27] If you had to assign one philosophy book for everyone in the world to read, what would it be?
Stewart: [01:04:32] It depends how much support they had, because I think like Spinoza's tractatus would be a great thing for anyone to read, like it's-
Patrick: [01:04:40] Advanced material, yeah.
Stewart: [01:04:41] It would actually just be a philosophy 101 primer, she gave an overview of it. People will say, and I definitely believe, that the most profound question is why is there something instead of nothing? By profound, it might just be the most unanswerable. But the reason, maybe this ties it all up, that we don't put the effort into helping people become more effective communicators that they know that they should is because you can't go through life questioning everything, unless you are a professional philosopher because it's just exhausting and paralyzing, and yet the effort that humans have put into digging deep into fundamental assumptions has created most of the progress.
The most political and economic theories that have led us to this stage in the world for better or worse, I guess, that all good, all advancement, all travel down the upward arc comes from questioning things that we have assumed answers to and don't question enough.
Patrick: [01:05:40] Well, Stewart, I'm so appreciative of your time and for your writing, and I hope you continue to do more of it. As someone that was directly affected by that essay and some of the other things you put out there, in particular, it's a pleasure to get to connect with you in this format today. Thank you so much.
Stewart: [01:05:53] Yeah, thank you.